text stringlengths 12 4.76M | timestamp stringlengths 26 26 | url stringlengths 32 32 |
|---|---|---|
Turkish Airlines (THY) announced on Jan. 17 that it had cancelled a total of 246 of its domestic and international flights over forecasted snowfall in Istanbul on Jan. 18.
“Our flights on the chart have been cancelled due to forecasted meteorological events and heavy snowfall in Istanbul’s Ataturk and Sabiha Gokcen airports. We urge our passengers to get informed and get updates about their flights by checking our website or calling out call centre,” the carrier said in a statement on its website. THY had also cancelled some 95 flights over strong Lodos winds in Istanbul on Jan. 16.
“Due to storm-strength ‘Lodos’ winds, flights on Jan. 16 and 17 have been cancelled to reduce potential capacity,” it said. | 2023-09-07T01:26:57.367551 | https://example.com/article/4031 |
Sewage sludge is usually discarded at the bottom of settling ponds. Old sludges are of little value in terms of the useful organic content, as they have been biodegraded and may contain high levels of heavy metals. They are usually the anaerobically digested or aerated end products of waste water treatment. These sludges often have a useful nutrient content of phosphate and nitrates. New sludges still have a high organic content and generate sulfur gases in particular hydrogen sulfide.
U.S. Pat. Nos. 3,942,970 and 5,387,257 propose the use of sewage sludges in fertilizer.
However most governments set maximum acceptable levels for metal content in fertilizers and soil enhancers particularly for arsenic, cadmium, chromium, copper, mercury, molybdenum, nickel, lead, selenium and zinc because of the potential that these metals will enter the food chain.
U.S. Pat. No. 5,009,793 discloses the removal of heavy metals from sludges and contaminated soils.
U.S. Pat. No. 5,178,776 discloses the removal of metals from sludges by acidification followed by a heating step to sterilize the treated solids.
U.S. Pat. No. 5,051,191 treats non oxidized sewage sludge at a low solids content to remove heavy metals by acid solubilization and subsequent precipitation from the liquid phase to leave a decontaminated sludge that is neutralized and then used as a soil conditioner and fertilizer.
USA application 2002/0153307 improves on this process by operating at a higher acid pH to reduce the solubilization of fertilizing elements in the sludge.
A problem with prior art processes is that a waste stream is still produced. Also because the value of the decontaminated sludge is not high the cost of carrying out the process must be kept to a minimum. Some prior art processes use expensive chemicals and require heating of the sludge all of which adds to processing costs.
It is an object of this invention to provide a process that overcomes the problems of the prior art. | 2023-09-26T01:26:57.367551 | https://example.com/article/4243 |
Meta
SALISBURY SAMMY
This is a design I was asked to do to “toughen up” the mascot for use with Salisbury’s athletics programs. Ultimately, the university did not approve having different versions and they’ve stuck with the original. | 2024-03-13T01:26:57.367551 | https://example.com/article/5735 |
---
abstract: 'To the best of our knowledge, there are no general well-founded robust methods for statistical unsupervised learning. Most of the unsupervised methods explicitly or implicitly depend on the kernel covariance operator (kernel CO) or kernel cross-covariance operator (kernel CCO). They are sensitive to contaminated data, even when using bounded positive definite kernels. First, we propose robust kernel covariance operator (robust kernel CO) and robust kernel cross-covariance operator (robust kernel CCO) based on a generalized loss function instead of the quadratic loss function. Second, we propose influence function of classical kernel canonical correlation analysis ([*classical kernel CCA*]{}). Third, using this influence function, we propose a visualization method to detect influential observations from two sets of data. Finally, we propose a method based on robust kernel CO and robust kernel CCO, called [*robust kernel CCA*]{}, which is designed for contaminated data and less sensitive to noise than [*classical kernel CCA*]{}. The principles we describe also apply to many kernel methods which must deal with the issue of kernel CO or kernel CCO. Experiments on synthesized and imaging genetics analysis demonstrate that the proposed visualization and [*robust kernel CCA*]{} can be applied effectively to both ideal data and contaminated data. The robust methods show the superior performance over the state-of-the-art methods.'
author:
- |
**Md. Ashad Alam$^{1,2}$, Kenji Fukumizu$^3$ and Yu-Ping Wang$^1$**\
$^{1}$Biomedical Engineering Department, Tulane University\
New Orlenas, LA 70118,USA\
$^2$Department of Statistics, Hajee Mohammad Danesh Science and Technology\
University Dinajpur 5200, Bangladesh\
$^3$The Institute of Statistical Mathematics, Tachikawa, Tokyo 190-8562, Japan
bibliography:
- 'Ref-UKIF.bib'
title: 'Robust Kernel (Cross-) Covariance Operators in Reproducing Kernel Hilbert Space toward Kernel Methods'
---
Introduction {#sec:Intro}
============
\[sec:Intro\] The incorporation of various unsupervised learning methods for multiple data sources into genomic analysis is a rather recent topic. Using the dual representations, the task of learning with multiple data sources is related to the kernel-based data fusion, which has been actively studied in the last decade [@Back-08; @Steinwart-08; @Hofmann-08]. Kernel fusion in unsupervised learning has a close connection with unsupervised kernel methods. As unsupervised kernel methods, kernel principal component analysis [@Schlkof-kpca; @Ashad-14 kernel PCA], kernel canonical correlation analysis [@Akaho; @Back-02 classical kernel CCA], weighted multiple kernel CCA and others have been extensively studied in unsupervised kernel fusion for decades [@Yu-11]. But these methods are not robust; these are sensitive to contaminated data. Even though a number of researches has been done on robustness issue for supervised learning, especially support vector machine for classification and regression [@Christmann-04; @Christmann-07; @Debruyne-08], there are no general well-founded robust methods for unsupervised learning.
Robustness is an essential and challenging issue in statistical machine learning for multiple sources data analysis. Because [*outliers*]{}, data that cause surprise in relation to the majority of the data, are often occur in the real data. Outliers may be right, but we need to examine for transcription errors. They can play havoc with classical statistical methods or statistical machine learning methods. To overcome this problem, since 1960 many robust methods have been developed, which are less sensitive to outliers. The goals of robust statistics are to use the methods from the bulk of the data and indicate the points deviating from the original pattern for further investment [@Huber-09; @Hampel-11]. In recent years, a robust kernel density estimation (robust kernel DE) has been proposed [@Kim-12], which is less sensitive than the kernel density estimation. To the best of our knowledge, two spacial robust kernel PCA methods have been proposed based on weighted eigenvalues decomposition [@Huang-KPCA] and spherical kernel PCA [@Debruyne-10]. They show that the influence function (IF), a well-known measure of robustness, of kernel PCA can be arbitrary large for unbounded kernels.
During the last ten years, a number of papers have been about the properties of kernel CCA, CCA using positive definite kernels, called [*classical kernel CCA*]{} and its variants have been proposed [@Fukumizu-SCKCCA; @Hardoon2009; @Otopal-12; @Ashad-15]. Due to the properties of eigen decomposition it is still a well applied methods for multiple souses data analysis. In recent years, two canonical correlation analysis (CCA) methods based on Hilbert-Schmidt independence criterion (hsicCCA) and centered kernel target alignment (ktaCCA) have been proposed by [@Chang-13]. These methods are able to extract nonlinear structure of the data as well. Due to the gradient based optimization, these methods are not able to extract all canonical variates using the same initial value and do not work for high dimensional datasets. For more details, see Section \[sec:classification\]. An empirical comparison and sensitivity analysis for robust linear CCA and classical kernel CCA is also discussed, and gives similar interpretation as kernel PCA for kernel CCA without any theoretical results [@Ashad-10].
Most of the kernel methods explicitly or implicitly depend on kernel covariance operator (kernel CO) or kernel cross-covariance operator (kernel CCO). Among others, these are most useful tools of unsupervised kernel methods but have not been robust yet. They can be formulated as an empirical optimization problem to achieve robustness by combining empirical optimization problem with ideas of Huber or Hampel’s M-estimation model [@Huber-09; @Hampel-11]. The robust kernel CO and robust kernel CCO can be computed efficiently via a kernelized iteratively re-weighted least square (KIRWLS) problem. In robust kernel DE based on robust kernel mean elements (robust kernel ME) is used KIRWLS in reproducing kernel Hilbert space (RKHS) [@Kim-12]. [@Debruyne-10] have proposed a visualization methods for detecting influential observations from one set of the data using IF of kernel PCA. In addition, [@Romanazii-92] has proposed the IF of canonical correlation and canonical vectors of linear CCA but the IF of classical kernel CCA and any robust kernel CCA have not been proposed, yet. All of these considerations motivate us to conduct studies on robust kernel CCO toward kernel unsupervised methods. Contribution of this paper is fourfold. First, we propose robust kernel CO and robust kernel CCO based on generalized loss function instead of the quadratic loss function. Second, we propose IF of classical kernel CCA: kernel canonical correlation (kernel CC) and kernel canonical variates (kernel CV). Third, to detect influential observations from multiple sets of data, we propose a visualization method using the inflection function of kernel CCA. Finally, we propose a method based on robust kernel CO and robust kernel CCO, called [*robust kernel CCA*]{}, which is less sensitive than classical kernel CCA. Experiments on synthesized and imaging genetics analysis demonstrate that the proposed visualization and robust [*kernel CCA*]{} can be applied effectively to both ideal data (ID) and contaminated data (CD).
The remainder of this paper is organized as follows. In the next Section, we provide a brief review of kernel ME, kernel CCO, robust kernel ME, robust kernel CO, robust kernel CCO and robust Gram matrices with algorithms. In Section $3$, we discuss in brief the IF, IF of kernel ME and IF of kernel CO and kernel CCO. After a brief review of classical kernel CCA in Section \[sec:CKCCA\], we propose the IF of classical kernel CCA: kernel CC and kernel CV in Section \[sec:IFKCCA\]. The [*robust kernel CCA*]{} is proposed in Section \[sec:RKCCA\]. In Section $5$, we describe experiments conducted on both synthesized data and the imaging genetics analysis with a visualizing method. In Appendix, we discuss the results in detail.
Classical and robust kernel (cross-) covariance operator in RKHS
================================================================
Kernel ME, kernel CO and kernel CCO with positive definite kernel have been extensively applied to nonparametric statistical inference through representing distribution in the form of means and covariance in RKHS [@Gretton-08; @Fukumizu-08; @Song-08; @Kim-12; @Gretton-12]. Basic notion of kernel MEs, kernel CO and kernel CCO with its robustness through IF are briefly discussed below.
Classical kernel (cross-) covariance operator {#sec:kernel CCO}
---------------------------------------------
Let $F_X$, $F_Y$ and $F_{XY}$ be the probability measure on ${\mathcal}{X}$, ${\mathcal}{Y}$ and $ {\mathcal}{X}\times{\mathcal}{Y}$, respectively. Also let $X_1, X_2, \ldots, X_n$,; $ Y_1, Y_2, \ldots, Y_n$ and $(X_1, Y_1),(X_2, Y_2), \ldots, (X_2, Y_2)$ be the random sample from the distribution $F_X$, $F_Y$ and $F_{XY}$, respectively. A symmetric kernel $k(\cdot,\cdot)$ defined on a space is called [*positive definite kernel*]{} if the Gram matrix $(k(X_i, X_j))_{ij}$ is positive semi-definite [@Aron-RKHS]. By the reproduction properties and kernel trick, the kernel can evaluate the inner product of any two feature vectors efficiently without knowing an explicit form of either the [*feature map*]{} ($\Phi(\cdot) = k(\cdot, X), \forall X\in {\mathcal}{X}$) or [*feature spaces*]{} (${\mathcal}{H}$). In addition, the computational cost does not depend on dimension of the original space after computing the Gram matrices [@Fukumizu-14; @Ashad-14].
A mapping ${\mathcal}{M}_X:= {\mathbb}{E}_X[\Phi(X)] = {\mathbb}{E}_X[k(\cdot, X)]$ with ${\mathbb}{E}_X[k(X, X)] < \infty$ is an element of the RKHS ${\mathcal}{H}_X$. By the reproducing property with $X\in {\mathcal}{X}$, [*kernel mean elements*]{} is defined as $$\langle {\mathcal}{M}_X, f \rangle_{{\mathcal}{H}_X} = \langle {\mathbb}{E}_X[k(\cdot, X)], f \rangle = {\mathbb}{E}_X\langle k(\cdot, X), f \rangle_{{\mathcal}{H}_X}={\mathbb}{E}_X[f(X)],$
for all $f\in {\mathcal}{H}_X$$. Given an independent and identically distributed sample, the mapping $m_X=\frac{1}{n}\sum_{i=1}^n\Phi(X_i)= \frac{1}{n} \sum_{i=1}^n k(\cdot, X_i)$ is an empirical element of the RKHS, ${\mathcal}{H}_X$, $\langle m_X, f \rangle_{{\mathcal}{H}_X}= \langle \frac{1}{n}\sum_{i=1}^n k(\cdot, X_i), f\rangle = \frac{1}{n}\sum_{i=1}^n f(X_i).
$ The sample kernel ME of the feature vectors ${\mathbf}{\Phi}(X_i)$ can be regraded as a solution to the empirical risk optimization problem [@Kim-12] $$\begin{aligned}
\label{EROP1}
{\operatornamewithlimits{argmin}}_{f\in {\mathcal}{H}_X} \frac{1}{n}\sum_{i=1}^n\| {\mathbf}{\Phi}(X_i)- f\|^2_{{\mathcal}{H}_X}.\end{aligned}$$ Similarly, we can define kernel CCO as an empirical risk optimization problem. An operator, $\Sigma_{YX} := {\mathcal}{H}_X \to {\mathcal}{Y}_Y$ with ${\mathbb}{E}_X[k_X(X, X)] < \infty$, and ${\mathbb}{E}_Y[k_Y(Y, Y)] < \infty$, by the reproducing property which is defined as $$\begin{aligned}
\langle f_X, \Sigma_{YX}f_X\rangle_{{\mathcal}{H}_Y} &=& {\mathbb}{E}_{XY}\left[\langle f_X, k_X(\cdot, X) - {\mathcal}{M}_X \rangle_{{\mathcal}{H}_X} \langle f_Y, k_Y(\cdot, Y) - {\mathcal}{M}_Y \rangle_{{\mathcal}{H}_Y} \right]\nonumber\\&=& {\mathbb}{E}_{XY}\left[(f_X(X) - E_X[f(X)]) (f_Y(Y) - E_Y[f(Y)])\right] \nonumber\end{aligned}$$ and called [*kernel CCO*]{}. Given the pair of independent and identically distributed sample, $(X_i, Y_i)_{i=1}^n$, the kernel CCO is an operator of the RKHS, ${\mathcal}{H}_X\otimes{\mathcal}{H}_Y$, Eq. (\[EROP1\]) becomes $$\begin{aligned}
\label{EROP2}
{\operatornamewithlimits{argmin}}_{\Sigma_{XY}\in {\mathcal}{H}_X\otimes {\mathcal}{H}_Y}\frac{1}{n} \sum_{i=1}^n \| {\mathbf}{\Phi}_c(X_i) \otimes {\mathbf}{\Phi}_c(Y_i) - \Sigma_{XY}\|^2,\end{aligned}$$ where ${\mathbf}{\Phi}_c(X_i)= {\mathbf}{\Phi}(X_i)-\frac{1}{n}\sum_{b=1}^n {\mathbf}{\Phi}(X_b)$. and the kernel covariance operator at point $(X_i, Y_i)$ is then $$\hat{\Sigma}_{YX} (X_i, Y_i)= (k_X(X_i, X_b)- \frac{1}{n}\sum_{b=1}^nk_X(\cdot, X_b))\otimes (k_Y(Y_i, Y_d)- \frac{1}{n}\sum_{d=1}^n k_Y(Y_i, Y_d)).$$ Special case, if Y is equal to X, gives kernel CO.
Robust kernel (cross-) covariance operator
------------------------------------------
It is known that (as in Section \[sec:kernel CCO\]) the kernel ME is the solution to the empirical risk optimization problem, which are the least square type estimators. This type of estimators are sensitive to the presence of [*outliers*]{} in the features, $\Phi(X_i)$. In recent years, the robust kernel ME has been proposed for density estimation [@Kim-12]. Our goal is to extend this notion to kernel CO and kernel CCO. To do these, we estimates kernel CO and kernel CCO based on robust loss functions, [*M-estimator*]{}, and called, [*robust kernel CO*]{} and [*robust kernel CCO*]{}, respectively. Most common example of robust loss functions, $\zeta(t)$ on $t \geq 0$, are Huber’s or Hampel’s loss function. Unlike the quadratic loss function, the derivative of these loss functions is bounded [@Huber-09; @Hampel-86]. The Huber’s function is defined as $$\begin{aligned}
\zeta(t)=
\begin{cases}
t^2/2\qquad \qquad ,0\leq t\leq c
\\
ct-c^2/2\qquad ,c\leq t \nonumber
\end{cases}\end{aligned}$$ and Hampel’s function is defined as $$\begin{aligned}
\zeta(t)=
\begin{cases}
t^2/2\qquad \qquad\qquad\qquad ,0\leq t\le c_1
\\
c_1t-c_1^2/2\qquad\qquad ,c_1\leq t < c_2
\\
c_1(t-c_3)^2/2 (c_2-c_3)+ c_1(c_2+c_3-c_1)/2\qquad ,c_2\leq t < c_3\\
c_1(c_2+c_3-c_1)/2 \qquad\qquad ,c_3\leq t. \nonumber
\end{cases}\end{aligned}$$
The basic assumptions are: (i) $\zeta$ is non-decreasing, $\zeta(0)=0$ and $\zeta(t)/t \to 0$ as $t\to 0$ (ii) $\varphi(t)=\frac{\zeta^\prime(t)}{t}$ exists and is finite, where $\zeta^\prime(t)$ is derivative of $\zeta(t)$. (iii) $\zeta^\prime(t)$ and $\varphi(t)$ are continuous and bounded (iv) $\varphi(t)$ is Lipschitz continuous. Huber’s loss function as well as others hold for all of these assumptions [@Kim-12].
Given weights of robust kernel ME, ${\mathbf}{w}=[ w_1, w_2, \cdots, w_n]^T$, of a set of observations, the points ${\mathbf}{\Phi}_c(X_i):= {\mathbf}{\Phi}(X_i) - \sum_{a=1}^nw_a {\mathbf}{\Phi}(X_a)$ are centered and the centered Gram matrix is $\tilde{K}_{ij}=({\mathbf}{H}{\mathbf}{K}{\mathbf}{H}^T)_{ij}$, where ${\mathbf}{1}_n=[1, 1, \cdots, 1]^T$ and ${\mathbf}{H}={\mathbf}{I}- {\mathbf}{1}_n{\mathbf}{w}^T$.
Eq. (\[EROP2\]) can be written as $$\begin{aligned}
\label{REROP1}
{\operatornamewithlimits{argmin}}_{f\in {\mathcal}{H}_X} \frac{1}{n}\sum_{i=1}^n \zeta(\| {\mathbf}{\Phi}_c(X_i) \otimes {\mathbf}{\Phi}_c(Y_i) - \Sigma_{XY}\|).\end{aligned}$$ As in [@Kim-12], Eq. (\[REROP1\]) does not has a closed form solution, but using the kernel trick the classical re-weighted least squares (IRWLS) can be extended to a RKHS. The solution is then, $$\widehat{\Sigma}_{XY}^{(h)}= \sum_{i=1}^n w_i^{(h-1)}\tilde{k}(X, X_i)\tilde{k}(Y, Y_i),$$ where $w_i^{(h)}=\frac{\varphi(\|{\mathbf}{\Phi}_c(X_i)\otimes {\mathbf}{\Phi}_c(Y_i) - \Sigma_{XY}\|_{{\mathcal}{H}_X\otimes{\mathcal}{H}_Y}}{\sum_{b=1}^n\varphi(\| {\mathbf}{\Phi}_c(X_b)\otimes{\mathbf}{\Phi}_c(Y_b)- \Sigma_{XY}\|_{{\mathcal}{H}_X\otimes{\mathcal}{H}_Y})}\,, \rm{and} \, \varphi(x)=\frac{\zeta^\prime(x)}{x}.$
The algorithms of estimating robust Gram matrix and robust kernel CCO are given in Figure \[Robust.K.M\] and in Figure \[Robust.cross.covariance.M\], respectively.
------------------------------------------------------------------------
Input: $D=\{{\mathbf}{X}_1, {\mathbf}{X}_2 ,\ldots {\mathbf}{X}_n \}$ in ${\mathbb}{R}^m$. The kernel matrix ${\mathbf}{K}$ with kernel $k$ and ${\mathbf}{K}_{{\mathbf}{X}_i}= k(\cdot, {\mathbf}{X}_i)$. Threshold $TH$, (e.g., $10^{-8}$). The objective function of robust mean element is $$M_R=\arg \min_{f\in {\mathcal}{H}} J(f),\qquad \rm {where} \,
J(f)= \frac{1}{n}\sum_{i=1}^n \rho(\|K_{X_i} - f\|_{{\mathcal}{H}} )$$
1. Do the following steps until: $$\frac{|J(M_R^{(h+1)})- J(M_R^{(h)})|}{J(M_R^{(h)})} < TH,$$ where$ M_R^{(h)}=\sum_{i=1}^nw_i{\mathbf}{K}_{{\mathbf}{X}_i}, \,
w_i^{(h)}=\frac{\varphi(\|{\mathbf}{K}_{{\mathbf}{X}_i}- M_R^{(h)}\|_{\mathcal}{H})}{\sum_{i=1}^n\varphi(\|{\mathbf}{K}_{{\mathbf}{X}_i}- M_R^{(h)}\|_{\mathcal}{H})}\,, \rm{and} \, \varphi(x)=\frac{\xi^\prime(x)}{x}$
- Set $h=1$ and $w_i^{(0)}=\frac{1}{n}$.
- Solve $w_i^{(h)}=\frac{ \varphi(\epsilon_i^{[h]})}{\sum_i^n\varphi(\epsilon_i^{[h]})}$ and make a vector ${\mathbf}{w}$ for $i=1, 2, \cdots n$.
- Update the mean element, $M_R^{(h+1)}= [{\mathbf}{w}^{(h)}]^T{\mathbf}{K}$.
- Update error, $\epsilon^{[h+1]} =(\rm{diag}({\mathbf}{K})- 2[{\mathbf}{w}^{(h)}]^T {\mathbf}{K}+ [{\mathbf}{w}^{(h)}]^T {\mathbf}{K}[{\mathbf}{w}^{(h)}]^T{\mathbf}{1}_n )^{\frac{1}{2}} $.
- Update $h$ as $h+1$.
Output: the centered robust kernel matrix, $\tilde{{\mathbf}{K}}_R= {\mathbf}{H}{\mathbf}{K}{\mathbf}{H}^T$ where ${\mathbf}{H}={\mathbf}{I}_n- {\mathbf}{1}_n {\mathbf}{w}^T$
------------------------------------------------------------------------
------------------------------------------------------------------------
Input: $D=\{({\mathbf}{X}_1,{\mathbf}{Y}_1), ({\mathbf}{X}_2, {\mathbf}{Y}_2), \ldots ({\mathbf}{X}_n, {\mathbf}{Y}_n) \}$. The robust centered kernel matrix $\tilde{{\mathbf}{K}}_X$ and $\tilde{{\mathbf}{K}}_Y$ with kernel $k_X$ and $k_Y$, $\tilde{K}_{Xi}$ and, $\tilde{K}_{Yi}$ are the $i$th column of the $\tilde{{\mathbf}{K}}_X$ and $\tilde{{\mathbf}{K}}_Y$, respectively. Also define $\tilde{{\mathbf}{K}}_{{\mathbf}{X}_i}= k_X (\cdot, {\mathbf}{X}_i)$ and $\tilde{{\mathbf}{K}}_{{\mathbf}{Y}_i}= k_y(\cdot, {\mathbf}{Y}_i)$. Threshold $TH$ (e.g., $10^{-8}$). The objective function of robust cross-covariance operator is $$\hat{\Sigma}_R=\arg \min_{{\mathbf}{A}\in {{\mathcal}{H}_X\otimes{\mathcal}{H}_Y}} J({\mathbf}{A}),\qquad \rm {where} \,
J({\mathbf}{A})= \frac{1}{n}\sum_{i=1}^n \rho(\|{\mathbf}{B}_i - {\mathbf}{A}\|_{{\mathcal}{H}_X\otimes{\mathcal}{H}_Y}),$$ $${\mathbf}{B}_i=\tilde{\Phi}({\mathbf}{X}_i)\otimes\tilde{\Phi}({\mathbf}{Y}_i)^T=\tilde{{\mathbf}{K}}_{{\mathbf}{X}_i}\otimes \tilde{{\mathbf}{K}}_{{\mathbf}{Y}_i}$$
1. Do the following steps until: $$\frac{|J\Sigma_R^{(h+1)})- J(\Sigma_R^{(h)})|}{J(\Sigma_R^{(h)})} < TH,$$ where $ \hat{\Sigma}_R^{(h)}=\sum_{i=1}^nw_i^{(h-1)} {\mathbf}{B}_i,\,
w_i^{(h)}=\frac{\varphi(\|{\mathbf}{B}_i- \hat{\Sigma}_R^{(h)}\|_{ {\mathcal}{H}_X\otimes{\mathcal}{H}_Y})}{\sum_{i=1}^n \varphi(\|{\mathbf}{B}_i- \hat{\Sigma}_R^{(h)}\|_{{\mathcal}{H}_X\otimes{\mathcal}{H}_Y})}\, \rm{ and} \, \varphi(x)=\frac{\xi^\prime(x)}{x} $
- Set $h=1$, and $ w_i^{(0)}=\frac{1}{n}$
- Solve $w_i^{(h)}=\frac{ \varphi(\epsilon_i^{[h]})}{\sum_i^n\varphi(\epsilon_i^{[h]})}$ and make a vector ${\mathbf}{w}$ for $i=1, 2, \cdots n$.
- Calculate a $n^2\times 1$ vector, ${\mathbf}{v}^{(h)}= {\mathbf}{B} {\mathbf}{w}^{(h)}$ and make a $n\times n$ matrix ${\mathbf}{V}^{(h)}$, where ${\mathbf}{B}$ is $n^2\times n$ matrix that $i$th column consists of all elements of the $n\times n$ matrix ${\mathbf}{B}_i$.
- Update the robust covariance, $\hat{\Sigma}_R^{(h+1)}= \sum_i^nw_i^{(h)}{\mathbf}{B}_i= {\mathbf}{V}^{(h)}$.
- Update error, $\epsilon^{[h+1]} =(\rm{diag}(\tilde{{\mathbf}{K}}_X\tilde{{\mathbf}{K}}_Y)- 2[{\mathbf}{w}^{(h)}]^T \tilde{{\mathbf}{K}}_X \tilde{{\mathbf}{K}}_X+ [{\mathbf}{w}^{(h)}]^T \tilde{{\mathbf}{K}}_X\tilde{{\mathbf}{K}}_Y[{\mathbf}{w}^{(h)}]^T{\mathbf}{1}_n )^{\frac{1}{2}} $.
- Update $h$ as $h+1$.
Output: the robust cross-covariance operator.
------------------------------------------------------------------------
Influence function of kernel (cross-) covariance operator
=========================================================
To define the notion of robustness in statistics, different approaches have been proposed science 70’s decay for examples, the [*minimax approach*]{} [@Huber-64], the [*sensitivity curve*]{} [@Tukey-77], the [*influence functions*]{} [@Hampel-74; @Hampel-86] and in the finite sample [em breakdown point]{} [@Dono-83]. Due to simplicity, IF is the most useful approach in statistics and in statistical supervised learning [@Christmann-07; @Christmann-04]. In this section, we briefly discuss the notion of IF, IF of kernel ME, IF of kernel CO and kernel CCO. (For details see in Appendix).
Let ($\Omega$, ${\mathcal}{A}$, ${\mathbb}{P}$) be a probability space and $({\mathcal}{X}, {\mathcal}{B})$ a measure space. We want to estimate the parameter $\theta \in \Theta$ of a distribution $F$ in ${\mathcal}{A}$. We assume that exists a functional $R: {\mathcal}{D}(R) \to {\mathbb}{R}$, where $ {\mathcal}{D}(R)$ is the set of all probability distribution in ${\mathcal}{A}$. Let $G$ be some distribution in ${\mathcal}{A}$. If data do not fallow the model $F$ exactly but slightly going toward $G$, the Gâteaux Derivative at $F$ is given by $$\begin{aligned}
\lim_{\epsilon\to 0 }\frac{R[(1-\epsilon)F+\epsilon G] - R(F)}{\epsilon}\end{aligned}$$ Suppose $x\in {\mathcal}{X}$ and $G=\Delta_x$ is the probability measure which gives mass $1$ to $\{x\}$. The [*influence function*]{} (special case of Gâteaux Derivative) of $R$ at $F$ is defined by $$\begin{aligned}
IF(x, R, F)=\lim_{\epsilon\to 0 }\frac{R[(1-\epsilon)F+\epsilon\Delta_x] - R(F)}{\epsilon}\end{aligned}$$ provided that the limit exists. It can be intuitively interpreted as a suitably normalized asymptotic influence of outliers on the value of an estimate or test statistic.
There are three properties of IF: gross error sensitivity, local shift sensitivity and rejection point. They measured the worst effect of gross error, the worst effect of rounding error and rejection point. For a scalar, we just define influence function (IF) at a fixed point. But if the estimate is a function, we are able to express the change of the function value at every points [@Kim-12].
Influence function of kernel mean element and kernel cross-raw moment
---------------------------------------------------------------------
For a scalar we just define IF at a fixed point. But if the estimate is a function, we are able to express the change of the function value at every point.
Let the cross-raw moments $$R(F_{XY}) = {\mathbb}{E}_{XY}[ \langle k_X(\cdot, X), f \rangle_{{\mathcal}{H}_X} \langle k_Y(\cdot,Y), g \rangle_{{\mathcal}{H}_Y}]= {\mathbb}{E}_{XY}[f(X)g(Y)]= \int f(X) g(Y)dF_{XY}.$$
The IF of $R(F_X)$ at $Z^\prime=(X^\prime,Y^\prime) $ for every points $(\cdot)$ is given by
$$IF(\cdot, Z^\prime, R, F_{X Y}) = k_X(\cdot, X^\prime) k_Y(\cdot, Y^\prime) -{\mathbb}{E}_{XY}[ \langle k_X(\cdot, X), f \rangle_{{\mathcal}{H}_X} \langle k_Y(\cdot,Y), g \rangle_{{\mathcal}{H}_Y}],$$ $ \forall k_X(\cdot, X) \in {\mathcal}{H}_X,\, k_Y(\cdot, Y) \in {\mathcal}{H}_Y$, which is estimated with the pairs of data points $ (X_1 Y_1), (X_2,Y_2), \cdots, (X_n,Y_n) \in {\mathcal}{X}\times {\mathcal}{Y}$ at any evaluated point $ (X_i,Y_i) \in {\mathcal}{X}\times {\mathcal}{Y} $ $$\begin{aligned}
k_X(X_i, X^\prime) k(Y_i, Y^\prime)- \frac{1}{n}\sum_{a=1}^n k_X(X_i, X_a)k_Y(Y_i,Y_a) \qquad \forall \, k_X(\cdot, X_a) \in {\mathcal}{H}_{\mathcal}{X},\, k_Y(\cdot, Y_a) \in {\mathcal}{H}_{\mathcal}{Y}. \nonumber\end{aligned}$$
Influence function of complicated statistics
--------------------------------------------
The IF of complicated statistics, which are functions of simple statistics, can be calculated with the chain rule. Say $R(F)=a\{R_1(F), ....., R_s(F)\}$, then $$\begin{aligned}
IF_R(z)=\sum_{i=1}^s \frac{\partial a}{\partial R_i}IF_{R_i }(z). \nonumber
\end{aligned}$$ It can also be used to find the IF for a transformed statistic, given the influence function for the statistic itself.
The IF of kernel CCO, $R(F_{XY})$, with joint distribution, $F_{XY}$, using complicated statistics at $Z^\prime=(X^\prime,Y^\prime)$ is given by $$\begin{aligned}
&\rm{IF}(\cdot, Z^\prime, R, F_{XY}) \nonumber \\
&= \langle k_X(\cdot, X^\prime)-{\mathcal}{M}[F_X], f \rangle_{{\mathcal}{H}_X} \langle k_Y(\cdot,Y^\prime){\mathcal}{M}[F_Y], g \rangle_{{\mathcal}{H}_Y}
\\ &-{\mathbb}{E}_{XY}[ \langle k_X(\cdot, X)-{\mathcal}{M}[F_X], f \rangle_{{\mathcal}{H}_X} \langle k_Y(\cdot,Y)-{\mathcal}{M}[F_Y], g \rangle_{{\mathcal}{H}_Y}],\end{aligned}$$ which is estimated with the data points $(X_1 Y_1), (X_2, Y_2), \cdots, (X_n, Y_n) \in {\mathcal}{X}\times {\mathcal}{Y}$ for every $Z_i = (X_i, Y_i)$ as $$\begin{gathered}
\widehat{\rm{IF}}( Z_i, Z^\prime, R, F_{XY}) \\
= [k_X(X_i, X^\prime)-\frac{1}{n}\sum_{b=1}^n k_X(X_i, X_b)] [k_Y(Y_i, Y^\prime)\\-\frac{1}{n}\sum_{b=1}^n k_Y(Y_i, Y_b)] -
\frac{1}{n}\sum_{d=1}^n[k_X(X_i, X_d)-\frac{1}{n}\sum_{b=1}^n k_X(X_i, X_b)] [k_Y(Y_i, Y_d)-\frac{1}{n}\sum_{b=1}^n k_Y(Y_i, Y_b)]. \nonumber\end{gathered}$$ For the bounded kernels, the above IFs have three properties: gross error sensitivity, local shift sensitivity and rejection point. It is not true for the unbounded kernels, for example, liner and polynomial kernels. We can make similar conclusion for the kernel covariance operator.
Classical and robust kernel canonical correlation analysis
==========================================================
In this Section, we review classical kernel CCA and propose the IF and empirical IF (EIF) of kernel CCA. After that we propose a [*robust kernel CCA*]{} method based on robust kernel CO and robust kernel CCO.
Classical kernel CCA {#sec:CKCCA}
--------------------
[*Classical kernel CCA*]{} has been proposed as a nonlinear extension of linear CCA [@Akaho; @Lai-00]. [@Back-02] has extended the [*classical kernel CCA*]{} with efficient computational algorithm, incomplete Cholesky factorization. Over the last decade, [*classical kernel CCA*]{} has been used for various purposes including preprocessing for classification, contrast function of independent component analysis, test of independence between two sets of variables, which has been applied in many domains such as genomics, computer graphics and computer-aided drug discovery and computational biology [@Alzate2008; @Hardoon2004; @Huang-2009]. Theoretical results on the convergence of kernel CCA have also been obtained [@Fukumizu-SCKCCA; @Hardoon2009].
The aim of [*classical kernel CCA*]{} is to seek the sets of functions in the RKHS for which the correlation (Corr) of random variables is maximized. The simplest case, given two sets of random variables $X$ and $Y$ with two functions in the RKHS, $f_{X}(\cdot)\in {\mathcal}{H}_X$ and $f_{Y}(\cdot)\in {\mathcal}{H}_Y$, the optimization problem of the random variables $f_X(X)$ and $f_Y(Y)$ is $$\begin{aligned}
\label{ckcca1}
\max_{\substack{f_{X}\in {\mathcal}{H}_X,f_{Y}\in {\mathcal}{H}_Y \\ f_{X}\ne 0,\,f_{Y}\ne 0}}{\mathrm}{Corr}(f_X(X),f_Y(Y)).\end{aligned}$$ The optimizing functions $f_{X}(\cdot)$ and $f_{Y}(\cdotp)$ are determined up to scale.
Using a finite sample, we are able to estimate the desired functions. Given an i.i.d sample, $(X_i,Y_i)_{i=1}^n$ from a joint distribution $F_{XY}$, by taking the inner products with elements or “parameters" in the RKHS, we have features $f_X(\cdot)=\langle f_X, \Phi_X(X)\rangle_{{\mathcal}{H}_X}= \sum_{i=1}^na_X^ik_X(\cdot,X_i) $ and $f_Y(\cdot)=\langle f_Y, \phi_Y(Y)\rangle_{{\mathcal}{H}_Y}=\sum_{i=1}^na_Y^ik_Y(\cdot,Y_i)$, where $k_X(\cdot, X)$ and $k_Y(\cdot, Y)$ are the associated kernel functions for ${\mathcal}{H}_X$ and ${\mathcal}{H}_Y$, respectively. The kernel Gram matrices are defined as ${\mathbf}{K}_X:=(k_X(X_i,X_j))_{i,j=1}^n $ and ${\mathbf}{K}_Y:=(k_Y(Y_i,Y_j))_{i,j=1}^n $. We need the centered kernel Gram matrices ${\mathbf}{M}_X={\mathbf}{C}{\mathbf}{K}_X{\mathbf}{C}$ and ${\mathbf}{M}_Y={\mathbf}{C}{\mathbf}{K}_Y{\mathbf}{C}$, where $ {\mathbf}{C} = {\mathbf}{I}_n -\frac{1}{n}{\mathbf}{B}_n$ with ${\mathbf}{B}_n = {\mathbf}{1}_n{\mathbf}{1}^T_n$ and ${\mathbf}{1}_n$ is the vector with $n$ ones. The empirical estimate of Eq. (\[ckcca1\]) is then given by $$\begin{aligned}
\label{ckcca6}
\max_{\substack{f_{X}\in {\mathcal}{H}_X,f_{Y}\in {\mathcal}{H}_Y \\ f_{X}\ne 0,\,f_{Y}\ne 0}}\frac{\widehat{\rm{Cov}}(f_X(X),f_Y(Y))}{[\widehat{\rm{Var}}(f_X(X))+\kappa\|f_X\|_{{\mathcal}{H}_X}]^{1/2}[\widehat{\rm{Var}}(f_Y(Y))+\kappa\|f_Y\|_{{\mathcal}{H}_Y}]^{1/2}} \nonumber \end{aligned}$$ where $$\begin{aligned}
& \widehat{\rm{Cov}}(f_X(X),f_Y(Y))
= \frac{1}{n} {\mathbf}{a}_X^T{\mathbf}{M}_X{\mathbf}{M}_Y {\mathbf}{a}_Y= {\mathbf}{a}_X^T{\mathbf}{M}_X{\mathbf}{W}{\mathbf}{M}_Y {\mathbf}{a}_Y , \\
& \widehat{\rm{Var}}( f_X(X))
=\frac{1}{n} {\mathbf}{a}_X^T{\mathbf}{M}_X^2 {\mathbf}{a}_X= {\mathbf}{a}_X^T{\mathbf}{M}_X {\mathbf}{W} {\mathbf}{M}_X {\mathbf}{a}_X, \, \\ &\widehat{\rm{Var}}( f_Y(Y))=\frac{1}{n} {\mathbf}{a}_Y^T{\mathbf}{M}_Y^2 {\mathbf}{a}_Y= {\mathbf}{a}_Y^T{\mathbf}{M}_Y {\mathbf}{W} {\mathbf}{M}_Y{\mathbf}{a}_Y, \end{aligned}$$ and ${\mathbf}{W}$ is a diagonal matrix with elements $\frac{1}{n}$, and ${\mathbf}{a}_{X}$ and ${\mathbf}{a}_{Y}$ are the directions of $X$ and $Y$, respectively. The regularized coefficient $\kappa > 0$.
### Influence function of classical kernel CCA {#sec:IFKCCA}
By using the IF results of kernel PCA, linear PCA and of linear CCA, we can derive the IF of kernel CCA: kernel CC and kernel CVs,
\[TIFKCCA\] Given two sets of random variables $(X, Y)$ having distribution $F_{XY}$, the influence function of kernel canonical correlation and canonical variate at $Z^\prime = (X^\prime, Y^\prime)$ are given by $$\begin{gathered}
\rm{IF} (Z^\prime, \rho_j^2)= - \rho_j^2 \bar{f}_{jX}^2(X^\prime) + 2 \rho_j \bar{f}_{jX}(X^\prime) \bar{f}_{jY}(Y^\prime) - \rho_j^2 \bar{f}_{jY}^2(Y^\prime), \nonumber\\
\rm{IF} (\cdot, Z^\prime, f_{jX}) = -\rho_j (\bar{f}_{jY}(Y^\prime) - \rho_j \bar{f}_{jX}(X^\prime)){\mathbb}{L} \tilde{k} (\cdot, X^\prime) - (\bar{f}_{jX}(X^\prime) \nonumber\\\qquad \qquad\qquad\qquad\qquad \qquad\qquad\qquad - \rho_j \bar{f}_{jY}(Y^\prime)){\mathbb}{L} \Sigma_{XY}\Sigma^{-1}_{YY} \tilde{k}_Y(\cdot, Y^\prime)+\frac{1}{2}[1- \bar{f}^2_{jX}(X^\prime)]f_{jX}, \nonumber\\
\rm{IF} (\cdot, Z^\prime, f_{jY})
= -\rho_j (\bar{f}_{jX}(X^\prime) - \rho_j \bar{f}_{jY}(Y^\prime)){\mathbb}{L} \tilde{k} (\cdot, Y^\prime)- (\bar{f}_{jY}(Y^\prime)\\ - \rho_j \bar{f}_{jX}(X^\prime)){\mathbb}{L} \Sigma_{YX}\Sigma^{-1}_{XX} \tilde{k}_Y(\cdot, Y^\prime) +\frac{1}{2}[1- \bar{f}^2_{jY}(Y^\prime)]f_{jY}, \nonumber\end{gathered}$$ [where]{} ${\mathbb}{L}= \Sigma_{XX}^{- \frac{1}{2}}(\Sigma_{XX}^{- \frac{1}{2}} \Sigma_{XY} \Sigma_{YY}^{-1} \Sigma_{YX}\Sigma_{XX}^{- \frac{1}{2}}-\rho^2{\mathbf}{I})^{-1}\Sigma_{XX}^{- \frac{1}{2}}$.
To prove Theorem \[TIFKCCA\], we need to find the IF of ${\mathbb}{L}$. All notations and proof are explained in Appendix.
It is known that the inverse of an operator may not exits even exist it may not be continuous operator in general [@Fukumizu-SCKCCA]. While we can derive kernel canonical correlation using correlation operator $ {\mathbf}{V}_{YX}= \Sigma_{YY}^{- \frac{1}{2}}\Sigma_{YX} \Sigma_{XX}^{- \frac{1}{2}}$, even when $\Sigma_{XX}^{- \frac{1}{2}}$ and $\Sigma_{YY}^{- \frac{1}{2}}$ are not proper operators, the IF of covariance operator is true only for the finite dimensional RKHSs. For infinite dimensional RKHSs, we can find IF of $\Sigma_{XX}^{- \frac{1}{2}}$ by introducing a regularization term as follows $$\begin{gathered}
\rm{IF}(\cdot, X^\prime,
(\Sigma_{XX} + \kappa{\mathbf}{I})^{- \frac{1}{2}}) = \frac{1}{2} [(\Sigma_{XX}+\kappa{\mathbf}{I})^{- \frac{1}{2}}- (\Sigma_{XX}+\kappa{\mathbf}{I})^{- \frac{1}{2}}\tilde{k}_X(\cdot, X^\prime)\otimes \tilde{k}_X(\cdot, X^\prime)(\Sigma_{XX} + \kappa{\mathbf}{I})^{- \frac{1}{2}}], \nonumber\end{gathered}$$ where $\kappa > 0$ is a regularization coefficient, which gives empirical estimator.
Let $(X_i, Y_i)_{i=1}^n$ be a sample from the distribution $F_{XY}$. The EIF of kernel CC and kernel CV at $Z^\prime=(X^\prime, Y^\prime)$ for all points $Z_i = (X_i, Y_i)$ are $\rm{EIF} (Z_i, Z^\prime, \rho_j^2) = \widehat{\rm{IF}} (Z^\prime, \hat{\rho}_j^2),
\rm{EIF} (Z_i, Z^\prime, f_{jX}) = \widehat{\rm{IF}} (\cdot, Z^\prime, f_{jX}),
\rm{EIF} (Z_i, Z^\prime, f_{jY}) = \widehat{\rm{EIF}} (\cdot, Z^\prime, \widehat{f}_{jY})$, respectively.
For the bounded kernels the IFs or EIFs, which are stated in Theorem \[TIFKCCA\] and after that, have the three properties: gross error sensitivity, local shift sensitivity and rejection point. But for unbounded kernels, say a linear, polynomial, the IFs are not bounded. In this consequence, the results of classical kernel CCA using the bounded kernels are less sensitive than the results of classical kernel CCA using the unbounded kernels. In practice, classical kernel CCA affected by the contaminated data even using the bounded kernels [@Ashad-10].
Robust kernel CCA {#sec:RKCCA}
-----------------
In this Section, we propose a [*robust kernel CCA*]{} methods based on robust kernel CO and robust kernel CCO. While many robust linear CCA methods have proposed to emphasize on the linear CCA methods that they fit the bulk of the data well and indicate the points deviating from the original pattern for further investment [@Adrover-15; @Ashad-10], there is no general well-founded robust methods of kernel CCA. The classical kernel CCA considers the same weights for each data point, $\frac{1}{n}$, to estimate kernel CO and kernel CCO, which is the solution of an empirical risk optimization problem using the quadratic loss function. It is known that the least square loss function is a no robust loss function. Instead of, we can solve empirical risk optimization problem using the robust least square loss function and the weights are determined based on data via KIRWLS. After getting robust kernel CO and kernel CCO, they are used in classical kernel CCA, which we called a [*robust kernel CCA*]{} method. Figure \[RobustKCCA\] presents detailed algorithm of the proposed methods (except first two steps, all steps are similar as classical kernel CCA). This method is designed for contaminated data as well, and the principles we describe apply also to the kernel methods, which must deal with the issue of kernel CO and kernel CCO.
------------------------------------------------------------------------
Input: $D=\{(X_1,Y_1), (X_2,Y_2), \ldots, (X_n,Y_n) \}$ in ${\mathbb}{R}^{m_1\times m_2}$.
1. Calculate the robust cross-covariance operator, $\hat{\Sigma}_{YX}$ using algorithm in Figure \[Robust.cross.covariance.M\].
2. Calculate the robust covariance operator $\hat{\Sigma}_{XX}$ and $\hat{\Sigma}_{YY}$ using the same weight of cross-covariance operator (for simplicity).
3. Find $ {\mathbb}{B}_{YX} = (\hat{\Sigma}_{YY} + \kappa {\mathbf}{I})^{- \frac{1}{2}} \hat{\Sigma}_{YX} (\hat{\Sigma}_{XX} + \kappa{\mathbf}{I})^{- \frac{1}{2}}$
4. For $\kappa > 0$, we have $\rho^2_j$ the largest eigenvalue of ${\mathbb}{B}_{YX}$ for $j=1, 2,\cdots, n$.
5. The unit eigenfunctions of ${\mathbb}{B}_{YX}$ corresponding to the $j$th eigenvalues are $\hat{\xi}_{jX}\in {\mathcal}{H}_X$ and $\hat{\xi}{j_Y}\in {\mathcal}{H}_Y$
6. The *[j]{}th ($j= 1, 2, \cdots, n$) kernel canonical variates are given by $$\hat{f}_{jX}(X) = \langle \hat{f}_{jX}, \tilde{k}_X(\cdot, X)\rangle \,\rm{and}\,\hat{f}_{jY}(X) = \langle \hat{f}_{jY}, \tilde{k}_Y(\cdot, Y) \rangle$$ where $\hat{f}_{jX} = (\hat{\Sigma}_{XX} + \kappa{\mathbf}{I})^{-\frac{1}{2}}\hat{\xi}_{jX}$ and $f_{jY} = (\hat{\Sigma}_{YY} + \kappa{\mathbf}{I})^{-\frac{1}{2}}\hat{\xi}_{jY}$*
Output: the robust kernel CCA
------------------------------------------------------------------------
Experiments
===========
We generate two types of simulated data, original data and those with $5\%$ of contamination, which are called ideal data (ID) and contaminated data (CD), respectively. We conduct experiments on the synthetic data as well as real data sets. The description of $7$ real data sets are in Sections \[Sec:Visu\] and \[sec:classification\], respectively. The $5$ synthetic data sets are as follows:
**Three circles structural data (TCSD):** Data are generated along three circles of different radii with small noise: $$X_i = r_i \begin{pmatrix} \cos({\mathbf}{Z}_i) \\ \sin({\mathbf}{Z}_i) \end{pmatrix}
+ \epsilon_i, \nonumber$$ where $r_i=1$, $0.5$ and $0.25$, for $i=1,\ldots,n_1$, $i=n_1+1,\ldots,n_2$, and $i=n_2+1,\ldots, n3$, respectively, ${\mathbf}{Z}_{i}\sim U[-\pi, \pi]$ and $\epsilon_i\sim {\mathcal}{N} (0, 0.01\,I_2)$ independently for an ID and ${\mathbf}{Z}_{i}\sim U[-10, 10]$ for the CD.
**Sign function structural data (SFSD):** 1500 data are generated along sine function with small noise: $$X_i = \begin{pmatrix} {\mathbf}{Z}_i \\ 2\sin(2{\mathbf}{Z}_i)\\
\vdots \\10\sin(10{\mathbf}{Z}_i) \end{pmatrix}
+ \epsilon_i, \nonumber$$ where ${\mathbf}{Z}_{i}\sim U[-2\pi,0]$ and $\epsilon_i\sim {\mathcal}{N} (0, 0.01\,I_{10})$ independently for the ID and $\epsilon_i\sim {\mathcal}{N} (0, 10\,I_{10})$ for the CD.
**Multivariate Gaussian structural data (MGSD):** Given multivariate normal data, ${\mathbf}{Z}_i\in{\mathbb}{R}^{12} \sim {\mathbf}{N}({\mathbf}{0},\Sigma)$ ($i= 1, 2, \ldots, n$) where $\Sigma$ is the same as in [@Ashad-15]. We divide ${\mathbf}{Z}_i$ into two sets of variables (${\mathbf}{Z}_{i1}$,${\mathbf}{Z}_{i2}$), and use the first six variables of ${\mathbf}{Z}_i$ as $X$ and perform $\log$ transformation of the absolute value of the remaining variables ($\log_e(|{\mathbf}{Z}_{i2}|))$) as $Y$. For the CD ${\mathbf}{Z}_i\in{\mathbb}{R}^{12} \sim {\mathbf}{N}({\mathbf}{1},\Sigma)$ ($i= 1,2,\ldots, n$).
**Sign and cosine function structural data (SCSD):** We use uniform marginal distribution, and transform the data by two periodic $\sin$ and $\cos$ functions to make two sets $X$ and $Y$, respectively, with additive Gaussian noise: $Z_i\sim U[-\pi,\pi], \,\eta_i\sim N(0,10^{-2}),~\,i=1,2,\ldots, n,
X_{ij}=\sin(j*Z_i)+\eta_i,\, Y_{ij} = \cos(j*Z_i)+\eta_i, j=1,2,\ldots, 100.$ For the CD $\eta_i\sim N(1,10^{-2})$.
**SNP and fMRI structural data (SMSD):** Two data sets of SNP data X with $1000$ SNPs and fMRI data Y with 1000 voxels were simulated. To correlate the SNPs with the voxels, a latent model is used as in [@Parkhomenko-09]). For contamination, we consider the signal level, $0.5$ and noise level, $1$ to $10$ and $20$, respectively.
In our experiments, first we compare classical and robust kernel covariance operators. After that the robust kernel CCA is compared with the classical kernel CCA, hsicCCA and ktaCCA. In all experiments, for the Gaussian kernel we use median of the pairwise distance as a bandwidth and for the Laplacain kernel the bandwidth is equal to 1. The regularized parameters of classical kernel CCA and robust kernel CCA is $\kappa= 10^{-5}$. In robust methods, we consider Hubuer’s loss function with the constant, $c$, equals to the median.
Kernel covariance operator and robust kernel operator covariance
----------------------------------------------------------------
We evaluate the performance of kernel CO and robust kernel CO in two different settings. First, we check the accuracy of both operators by considering the kernel CO with large data (say a population kernel CO). The measure of the kernel CO and robust kernel CO estimators are defined as $$\begin{gathered}
\eta_{kco}=\frac{1}{n^2}\sum_{i=1}^n\sum_{j=1}^nk(X_i,X_j)^2 - 2\frac{1}{Nn}\sum_{i=1}^n\sum_{J=1}^Nk(X_i,X_J)^2+ \frac{1}{N^2}\sum_{I=1}^N\sum_{J=1}^Nk(X_I,X_J)^2,\,\nonumber \rm{and}\end{gathered}$$ $$\begin{gathered}
\eta_{rkco}=\sum_{i=1}^n\sum_{j=1}^n w_iw_jk(X_i,X_j)^2- 2\frac{1}{N}\sum_{i=1}^n w_i\sum_{J=1}^Nk(X_i,X_J)^2 + \frac{1}{N^2}\sum_{I=1}^N\sum_{J=1}^Nk(X_I,X_J)^2, \nonumber \end{gathered}$$ respectively.
In theory, the above two equations become to zero for large population size, $N$, with the sample size, $n\to N$. To do this, we consider the synthetic data, TCSD with $N\in\{1500, 3000, 6000, 9000\}$ and $n\in\{15, 30, 45, 60, 90, 120, 150, 180, 210, 240, 270, 300\}$ ($n = n_1 + n_2 + n_3$). For each $n$ with $5\%$ CD, we consider $100$ samples and the results (mean with standard error) are plotted in Figure \[Fcons\]. Figures show that the both estimators give similar performance in small sample size but for large sample sizes the robust estimator, robust kernel CO shows much better results than kernel CO estimate at all population sizes.
![Accuracy measure of kernel covariance operator (Classical, $\eta_{kco}$) and robust kernel covariance operator (Robust, $\eta_{rkco}$).[]{data-label="Fcons"}](SC.eps){width="\columnwidth" height="12cm"}
Second, we compare kernel CO and robust kernel CO estimators using 5 kernels: linear (Poly-1), polynomial with degree $2$ (Poly-2) and polynomial with degree $3$ (Poly-3), Gaussian and Laplacian on two synthetic data sets: TCSD and SFSD. To measure the performance, we use 4 matrix norms: maximum of absolute column sum (O), Frobenius norm (F), maximum modulus of all the elements (M) and spectral (S) [@Sequeira-11]. We calculate the ratio between ID and CD for the kernel CO and robust kernel CO. For both estimators, we consider the following measure, $$\eta_{co} = | 1- \frac{\| \hat{C_{XX}}^{ID}\|}{\|\hat{C_{XX}}^{CD} \|} |.$$ We repeat the experiment for 100 samples with sample size, $n=1500$. The results (mean $\pm$ standard deviation) of $\eta_{co}$ for kernel CO (Classical) and robust kernel CO (Robust) are tabulated in Table \[NRCRcov\]. From this table, it is clear that the robust estimator performs better than the classical estimator in all cases. Moreover, both estimators using Gaussian and Lapalasian kernels are less sensitive than all polynomial kernels.
------------------------ ----------- --------------------- --------------------- ----------------------- ------------------------
Data
Measure Kerenl Classical Robust Classical Robust
Poly-1 $0.9947\pm 0.0007$ $0.9546\pm 0.0026$ $0.5692\pm 0.1426$ $0.4175\pm 0.1482$
Poly-2 $1.0000\pm 0.0000$ $0.9909\pm 0.0013$ $0.9652\pm 0.0494$ $ 0.6703\pm 0.172$
$\| \hat{C}_{XX} \|_O$ Poly-3 $1.0000\pm 0.0000$ $0.9997 \pm 0.0001$ $0.9971\pm 0.0095$ $ 0.8754 \pm 0.1104$
Gaussian $0.0844\pm 0.0054$ $0.0756\pm 0.0051$ $0.1167\pm 0.0459$ $ 0.0964 0.0424$
Laplacian $0.1133\pm 0.0054$ $0.0980\pm 0.0128$ $0.1332 \pm 0.0830 $ $ 0.1420 \pm 0.0745$
Poly-1 $0.9874\pm 0.0017$ $0.8963\pm 0.0069$ $0.3067\pm 0.1026$ $ 0.1669 \pm 0.0626$
Poly-2 $1.0000\pm 0.0000$ $0.9863\pm 0.0020$ $0.9559\pm 0.0622$ $ 0.5917 \pm 0.1598$
$\| \hat{C}_{XX} \|_F$ Poly-3 $1.0000 \pm 0.0000$ $0.9996\pm 0.0001$ $0.9973\pm 0.0094 $ $ 0.8793 \pm 0.1067$
Gaussian $0.1153\pm 0.0034$ $0.1181\pm 0.0039$ $0.1174\pm 0.0266 $ $ 0.1059 \pm 0.0258$
Laplacian $0.1420\pm 0.0032$ $0.1392\pm 0.0035$ $0.1351\pm 0.0459 $ $ 0.1580 \pm 0.0366$
Poly-1 $0.9993\pm 0.0001$ $0.9940\pm 0.0005$ $0.8074 \pm 0.0838$ $ 0.6944\pm 0.1118$
Poly-2 $1.0000\pm 0.0000$ $0.9996\pm 0.0001$ $0.9921 \pm 0.0122 $ $ 0.9070 \pm 0.0703$
$\| \hat{C}_{XX} \|_M$ Poly-3 $1.0000\pm 0.0000$ $1.0000\pm 0.0000$ $0.9994 \pm 0.0020 $ $ 0.9709 \pm 0.0344$
Gaussian $0.1300\pm 0.0133$ $0.1028\pm 0.0038$ $0.1065 \pm 0.0583 $ $ 0.0735 \pm 0.0370$
Laplacian $0.1877\pm 0.0053$ $0.1474\pm 0.0042$ $0.1065 \pm 0.0583 $ $ 0.0735 \pm 0.0370$
Poly-1 $0.9886\pm 0.0019$ $0.9007\pm 0.0091$ $0.2897\pm 0.1412 $ $ 0.1538\pm 0.0877$
Poly-2 $1.0000\pm 0.0000$ $0.9887\pm 0.0017$ $0.9591 \pm 0.0660 $ $ 0.5927 \pm 0.2002$
$\| \hat{C}_{XX} \|_S$ Poly-3 $1.0000\pm 0.0000$ $0.9997\pm 0.0001$ $0.9975 \pm 0.0090 $ $ 0.8846 \pm 0.1225$
Gaussian $0.1281\pm 0.0051$ $0.1227\pm 0.0048$ $0.1333 \pm 0.0475 $ $ 0.1091 \pm 0.0459$
Laplacian $0.1716\pm 0.0050$ $0.1499\pm 0.0045$ $0.1614 \pm 0.0759 $ $ 0.1696 \pm 0.0613$
------------------------ ----------- --------------------- --------------------- ----------------------- ------------------------
: Mean and standard deviation of the measure, $\eta_{cov}$, of kernel covariance operator (Classical) and robust kernel covariance operator (Robust).[]{data-label="tbl:norm"}
\[NRCRcov\]
Visualizing influential subject using classical kernel CCA and robust kernel CCA {#Sec:Visu}
--------------------------------------------------------------------------------
We evaluate the performance of the propose methods, [*robust kernel CCA*]{}, in three different settings. First, we compare robust kernel CCA with the classical kernel CCA using Gaussian kernel (same bandwidth and regularization). To measure the influence, we calculate the ratio between ID and CD of IF of kernel CC and kernel CV. Based on this ratio, we define two measure on kernel CC and kernel CV $$\begin{aligned}
\eta_{\rho} &=& | 1- \frac{\|EIF(\cdot, \rho^2) ^{ID}\|_F}{\|EIF(\cdot, \rho^2)^{CD}\|_F}| \qquad \rm{and} \\
\eta_{f} &=& | 1- \frac{\|EIF(\cdot, f_X)^{ID}- EIF(\cdot,f_Y)^{ID}\|_F}{\|EIF(\cdot, f_X) ^{CD}-EIF(\cdot, f_Y)^{CD}\|_F}|,\nonumber\end{aligned}$$ respectively. The method, which does not depend on the contaminated data, the above measures, $\eta_{\rho}$ and $\eta_{f}$, should be approximately zero. In other words, the best methods should give small values. To compare, we consider 3 simulated data sets: MGSD, SCSD, SMSD with 3 sample sizes, $n\in \{ 100, 500, 1000\}$. For each sample size, we repeat the experiment for $100$ samples. Table \[tbl:ifnorm\] presents the results (mean $\pm$ standard deviation) of classical kernel CCA and robust kernel CCA. From this table, we observe that robust kernel CCA outperforms than the classical kernel CCA in all cases.
------ --------- ---------------------- --------------------- ----------------------- ---------------------- -- -- -- --
Measure
Data n $Classical$ $Robust$ $Classical$ $Robust$
$100$ $1.9114\pm 3.5945$ $ 1.2445\pm 3.1262$ $ 1.3379\pm 3.5092$ $ 1.3043\pm 2.1842$
MGSD $500$ $ 1.1365\pm 1.9545$ $ 1.0864\pm 1.5963$ $ 0.8631 \pm 1.3324 $ $0.7096\pm 0.7463$
$1000$ $ 1.1695\pm 1.6264$ $ 1.0831\pm 1.8842$ $ 0.6193 \pm 0.7838$ $ 0.5886\pm 0.6212$
$100$ $0.4945\pm 0.5750$ $0.3963\pm 0.4642$ $1.6855\pm 2.1862$ $0.9953\pm 1.3497$
SCSD $500$ $0.2581\pm 0.2101$ $0.2786\pm 0.4315$ $1.3933\pm 1.9546$ $ 1.1606\pm 1.3400$
$1000$ $0.1537\pm 0.1272$ $ 0.1501\pm 0.1252$ $ 1.6822\pm 2.2284$ $1.2715\pm 1.7100$
$100$ $ 0.6455 \pm 0.0532$ $ 0.1485\pm 0.1020$ $ 0.6507 \pm 0.2589 $ $2.6174\pm 3.3295$
SMSD $500$ $0.6449\pm 0.0223$ $0.0551\pm 0.0463$ $ 3.7345 \pm 2.2394$ $ 1.3733 \pm 1.3765$
$1000$ $ 0.6425 \pm 0.0134$ $ 0.0350\pm 0.0312$ $ 7.7497\pm 1.2857$ $ 0.3811 \pm 0.3846$
------ --------- ---------------------- --------------------- ----------------------- ---------------------- -- -- -- --
: Mean and standard deviation of the measures, $\eta_{\rho}$ and $\eta_{f}$ of classical kernel CCA (Classical) and robust kernel CCA (Robust).[]{data-label="tbl:ifnorm"}
Second, we propose a simple graphical display based on EIF of kernel CCA, the index plots (the subject on $x$-axis and the influence, $\eta_{\rho}$, on $y$ axis), to assess the related influence data points in data fusion with respect to EIF based on kernel CCA, $\eta_{\rho}$. To do this, we consider simulated SNP and fMRI data (SMSD) and real SNP and fMRI, Mind Clinical Imaging Consortium (MCIC) Data.
![Influence points of SMDS for ideal data and contaminated data using classical and robust kernel CCA.[]{data-label="SMDSIFOB"}](SMSDIF.eps){width="\columnwidth" height="12cm"}
[**Mind Clinical Imaging Consortium (MCIC) Data:**]{} The Mind Clinical Imaging Consortium (MCIC) has collected two types of data (SNPs and fMRI) from 208 subjects including $92$ schizophrenia patients and $116$ healthy controls. Without missing data the number of subjects is $184$ ($81$ schizophrenia patients and $103$ healthy controls) [@Dongdong-14]. After prepossessing we select $19872$ voxels and $39770$ loci for fMRI data and SNP data, respectively.
The index plots of classical kernel CCA and robust kernel CCA using the SMSD are presented in Figure \[SMDSIFOB\]. The $1$st and $2$nd rows, and columns of this figure are for ID and CD, and classical kernel CCA (Classical KCCA) and robust kernel CCA (Robust KCCA), respectively. These plots show that both methods have almost similar results of the ID. But for CD, it is clear that the classical kernel CCA is affected by the CD in significantly. We can easily identify influence of observation using this visualization. On the other hand, the robust kernel CCA has almost similar results of both data sets, ID and CD.
To detect influential subjects (in schizophrenia patients and healthy controls), we use the EIF of kernel CC of classical and robust kernel CCA methods. For robust kernel CCA, we use robust kernel CC and kernel CVs in Theorem \[TIFKCCA\]. The values of $\eta_{\rho}$ are plotted separately in Figure \[MCICFOB\]. The schizophrenia patients and healthy controls are in $1$st and $2$nd rows, respectively. These plots show that healthy controls have less influence than the schizophrenia patients group. The subject $59$ in Schizophrenia patients has the largest influence over all data and the subject $119$ has the largest influence over healthy controls only. However, both classical and robust kernel CCA have identified similar subject but robust kernel CCA is less sensitive than classical kernel CCA.
![Influence subject of MCIC data set of classical and robust KCCA of the real data.[]{data-label="MCICFOB"}](Rd1.eps){width="\columnwidth" height="12cm"}
Extraction of low-dimensional space for classification {#sec:classification}
------------------------------------------------------
Finally, we use $6$ well-known real datasets for classification from the UCI repository [@UCI]: [*Wine*]{}, [*BUPA liver disorders*]{}, [*Breast tissue*]{}, [*Diabetes*]{}, [*Sona*]{}, and [*Lymphoma*]{} to test the significance of low dimensional canonical features of the input space. We use the features for the classification task. To specify the classes, for an $\ell$-class classification problem, the $\ell$ dimensional binary vectors $(1, 0,\ldots 0), (0, 1,\ldots 0), \ldots (0, 0,\ldots 1)$ are used for $Y$.
Using the low-dimensional canonical features (only $2$, $5$ and $10$) obtained by CCA, we evaluate the classification errors by the kNN classifier ($k = 5$) with $10$-fold cross-validation. In comparison, we use the canonical features given by the classical kernel CCA, robust kernel CCA, hsicCCA and ktaCCA methods. For hsicCCA and ktaCCA methods we use “hsicCCA” R-package. The Table \[tb3:clasification\] presents the results with the number of data and dimensions. By this table, we see that the hsicCCA and ktaCCA methods are not able to extract all of the canonical variates. On top of that, these methods do not work for the high dimensional dataset. In Table \[tb3:clasification\], these situation are noted by $*$ and $**$ respectively. On the other hand, classical kernel CCA and robust kernel CCA can extract all canonical variates as well as for all data sets. In fact, the kernel CCA methods are more faster than the hsicCCA and ktaCCA methods.
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ -- -- -- -- -- -- -- -- -- --
[Dataset]{}&\# [of Data]{}&X&Y&[Methods]{}&$2$&$5$&$10$&$$&$$&$$&Classical KCCA &$ 2.81$&$0.56$&$0.56$Wine&$178$&$13$&$3$&Robust KCCA&$2.81$&$0.56$&$0.56$&$$&$$&$$&hsicCCA&$2.47$&$*$&$*$&$$&$$&$$&ktaCCA&$2.47$&$*$&$*$&$$&$$&$$&Classical kernel CCA &$ 36.79$&$20.76$&$17.92$Breast Tisu&$106$&$10$&$6$&Robust kernel CCA&$38.68$&$19.81$&$18.81$&$$&$$&$$&hsicCCA&$ 24.53$&$20.75$&$*$&$$&$$&$$&ktaCCA&$22.64$&$20.75$&$*$&$$&$$&$$&Classical kernel CCA &$12.41$&$5.52$&$4.14$Diabetes&$145$&$5$&$3$&Robust kernel CCA&$12.41$&$5.52$&$4.14$&$$&$$&$$&hsicCCA&$ 11.03$&$*\tnote{}$&$*$&$$&$$&$$&ktaCCA&$11.03$&$*$&$*$&$$&$$&$$&Classical kernel CCA &$14.90$&$14.90$&$13.94$Sona&$208$&$60$&$2$&Robust kernel CCA&$13.94$&$13.94$&$13.94$&$$&$$&$$&hsicCCA&$**\tnote{}$&$**$&$**$&$$&$$&$$&ktaCCA&$**$&$** $&$**$&$$&$$&$$&Classical kernel CCA &$1.61$&$0.00$&$0.00$Lymphoma&$64$&$4026$&$3$&Robust kernel CCA&$1.61$&$0.00$&$0.00$&$$&$$&$$&hsicCCA&$**$&$**$&$**$&$$&$$&$$&ktaCCA&$**$&$**$&$**$
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ -- -- -- -- -- -- -- -- -- --
: Classification errors by the kNN classifier ($k = 5$) with $10$-fold cross-validation[]{data-label="tb3:clasification"}
\*Functions cannot be evaluated at initial parameters
\*\*Curse of dimensionality
Concluding remark and future research
=====================================
The robust estimator employs robust loss function instead of quadratic loss function to achieve robustness for the contamination of the training sample. The robust estimators are weighted estimators, where smaller weights are given more outlying data points. The weights can be estimated efficiently using a KIRLS approach. In terms of the accuracy and sensitivity, it is clear that the robust estimators (robust kernel CO and robust kernel CCO) perform better than classical estimators (kernel CO and kernel CCO). We propose the IF of kernel CCA: kernel CC and kernel CVs and [*robust kernel CCA*]{} based on robust kernel CO and robust kernel CCO. The proposed IF measures the sensitivity of kernel CCA, which shows that classical kernel CCA is sensitive to contamination. But the proposed, [*robust kernel CCA*]{} is less sensitive to contamination. The visualization methods can identify influential (outlier) data in both synthesized and real imaging genetics analysis data. We also obtain low dimensional subspace for classification by CCA. we evaluate the classification errors by the kNN classifier ($k = 5$) with $10$-fold cross-validation. The proposed [*robust CCA*]{} shows the best performance over hsicCCA and ktaCCA methods.
For the EIF of robust kernel CCA, we use robust kernel CC and kernel CVs in Theorem \[TIFKCCA\]. The theoretical IF of robust kernel CCA is an expected future direction of research. Although our focus was on kernel CCA but we can robustify other kernel methods, which must deal with the issue of kernel CO and kernel CCO. In future work, it would be also interesting to develop robust multiple kernel PCA and robust multiple weighted kernel CCA to apply in genomic analysis.
Appendix
========
We present, derivation of robust centering Gram matrix, robust kernel cross-covariance operator, influence function (IF) of kernel mean elements and kernel cross-covariance operator and proofs which were omitted from the paper.
Derivation of centering Gram matrix using robust kernel mean element
--------------------------------------------------------------------
Given weight of robust kernel mean element ${\mathbf}{w}=[ w_1, w_2, \cdots, w_n]^T$ of a set of observations $X_1, \cdots, X_n$, the points $$\begin{aligned}
{\mathbf}{\Phi}_c(X_i):= {\mathbf}{\Phi}(X_i)- \sum_{b=1}^nw_b {\mathbf}{\Phi}(X_b) \nonumber\end{aligned}$$ are centered. Thus $$\begin{aligned}
\tilde{K}_{ij}&=&\langle {\mathbf}{\Phi}_c(X_i), {\mathbf}{\Phi}_c(X_j)\rangle \nonumber\\
&=& \left\langle {\mathbf}{\Phi}(X_i) - \sum_{b=1}^nw_b {\mathbf}{\Phi}(X_b), {\mathbf}{\Phi}(X_j)- \sum_{d=1}^nw_d {\mathbf}{\Phi}(X_d) \right\rangle \nonumber \\
&=&\langle {\mathbf}{\Phi}(X_i), {\mathbf}{\Phi}(X_j)\rangle - \sum_{b=1}^nw_b \langle {\mathbf}{\Phi}(X_b) , {\mathbf}{\Phi}(X_j)\rangle - \sum_{d=1}^nw_d \langle {\mathbf}{\Phi}(X_i) , {\mathbf}{\Phi}(X_d)\rangle + \sum_{b=1}^n \sum_{d=1}^nw_b w_d \langle {\mathbf}{\Phi}(X_b), {\mathbf}{\Phi}(X_d)\rangle \nonumber\\
&=&{\mathbf}{K}_{ij} - \sum_{b=1}^nw_bK_{bj} - \sum_{d=1} K_{id}w_d+\sum_{b=1}^n\sum_{d=1}^n w_bK_{bd}w_d \nonumber\\
&=&({\mathbf}{K}- {\mathbf}{1}_n {\mathbf}{w}^T {\mathbf}{K}- {\mathbf}{K} {\mathbf}{w}{\mathbf}{1}_n^T+ {\mathbf}{1}_n {\mathbf}{w}^T {\mathbf}{K} {\mathbf}{w}{\mathbf}{1}^T_n)_{ij} \nonumber \\
&=& ( ({\mathbf}{I}- {\mathbf}{1}_n{\mathbf}{w}^T) {\mathbf}{K} ({\mathbf}{I}- {\mathbf}{1}_n {\mathbf}{w}^T)^T)_{ij} \nonumber\\
&=&({\mathbf}{H}{\mathbf}{K}{\mathbf}{H}^T)_{ij},
\label{CM1}\end{aligned}$$ where ${\mathbf}{1}_n=[1, 1, \cdots, 1]^T$ and ${\mathbf}{H}={\mathbf}{I}- {\mathbf}{1}_n{\mathbf}{w}^T$. For a set of test points $ X^t_1, X_2^t, \cdots, X_T^t$, we define two matrices of order $T\times n$ as $K_{ij}^{test}= \langle \Phi(X^t_i), \Phi(X_j) \rangle$ and $\tilde{K}_{ij}^{test}= \langle \Phi(X^t_i)- \sum_{b=1}^nw_b \Phi(X_b), \Phi(X_j)- \sum_{d=1}^nw_d \Phi({\mathbf}{X}_b)\rangle$ As in Eq. (\[CM1\]) the robust centered Gram matrix of test points, $K_{ij}^{test}$, in terms of robust Gram matrix is defined as, $$\begin{aligned}
\label{CM2}
\tilde{K}_{ij}^{test}= K_{ij}^{test}- {\mathbf}{1}_T {\mathbf}{w}^T {\mathbf}{K}- {\mathbf}{K}^{test} {\mathbf}{w}{\mathbf}{1}_n^T+ {\mathbf}{1}_T {\mathbf}{w}^T {\mathbf}{K} {\mathbf}{w}{\mathbf}{1}^T_n \nonumber\end{aligned}$$
Derivation of centering Gram matrix using robust kernel mean element
--------------------------------------------------------------------
Similarly, we can define higher-order moment elements of the feature vector as an empirical risk optimization problem.
\[Kernel *r*th raw moment element\] A mapping ${\mathcal}{M}^{(r)}:= {\mathbb}{E}_X[\Phi(X)\otimes \Phi(X)\otimes \cdots \otimes \Phi(X)] = \otimes^r\Phi(X)={\mathbb}{E}_X[k(X, \cdot)k(X, \cdot)\cdots k(X, \cdot)]= {\mathbb}{E}_X[k^{(r)}(\cdot, X)]$ with ${\mathbb}{E}_X[k^{(r)}(X, X)] < \infty$ is an element of the RKHS, $\otimes^r{\mathcal}{H}_X$. By the reproducing property with $X\in {\mathcal}{X}$ $$\begin{aligned}
\langle {\mathcal}{M}^{(r)}, \otimes^r f \rangle_{\otimes^r{\mathcal}{H}_X} = {\mathbb}{E}_X[\langle {\mathcal}{M}, f \rangle_{{\mathcal}{H}_X} \langle {\mathcal}{M}, f \rangle_{{\mathcal}{H}_X} \cdots \langle {\mathcal}{M}, f \rangle_{{\mathcal}{H}_X}]
={\mathbb}{E}_X[f(X) f(X)\cdots f(X)]
= {\mathbb}{E}_X[f^{(r)}(X)], \end{aligned}$$ for all $f\in {\mathcal}{H}_X$. The mapping $m^{(r)}= \frac{1}{n}\sum_{i=1}^n \otimes^r \Phi(X_i)$ is an empirical *[r]{}th row moment element of the RKHS, $\otimes^r{\mathcal}{H}_X$, $$\langle m^{(r)}, \otimes^r f\rangle_{\otimes^r{\mathcal}{H}_X}= \frac{1}{n}\sum_{i=1}^nf^{(r)}(X_i),$$ where $\otimes^r f= f\otimes f \otimes \cdots \otimes f$ is the tensor product of $r$ functions, $f \in {\mathcal}{H}_X$. The sample *[r]{}th row moment element of the $ {\mathbf}{\Phi}(X_i) $ is a solution of an empirical risk optimization problem $$\begin{aligned}
\label{rmop}
{\operatornamewithlimits{argmin}}_{g\in \otimes^r{\mathcal}{H}_X} \frac{1}{n}\sum_{i=1}^n\| \otimes_{j=1}^r{\mathbf}{\Phi}^{(r)}(X_i) - g\|^2_{\otimes^r{\mathcal}{H}_X}, \end{aligned}$$ at the point $X$, $g(X, X,\cdots X) \in \otimes^r{\mathcal}{H}_X$.**
\[KCME\] A mapping ${\mathcal}{M}_c^{(r)}:= {\mathbb}{E}_X[\tilde{k}(\cdot,X)\tilde{k}(\cdot, X)\cdots \tilde{k}(\cdot, X)]$ with ${\mathbb}{E}_X[\tilde{k}^{(r)}(X, X)] < \infty$ is an element of the RKHS, $\otimes{\mathcal}{H}_X$. By the reproducing property $\forall k(\cdot, X),\, f\in {\mathcal}{H}_X\, \rm{and}\,, X\in {\mathcal}{X}\, \forall f\in {\mathcal}{H}_X$, $$\langle {\mathcal}{M}_c^{(r)} , \otimes^r f_c \rangle_{\otimes{\mathcal}{H}_X}= \langle {\mathcal}{M}_c ,f_c \rangle_{{\mathcal}{H}_X} \langle {\mathcal}{M}, f_c \rangle_{{\mathcal}{H}_X} \cdots \langle {\mathcal}{M}_c[F_X] ,f_c \rangle_{{\mathcal}{H}_X}= {\mathbb}{E}_X[k_c^{(r)}(\cdot, X)]$$ and the empirical *r*th central moment element at every point $X_i$ is defined by $$\langle {\mathcal}{M}_c^{(r)}, \otimes^r f_c\rangle_{\otimes{\mathcal}{H}_X}= \frac{1}{n}\sum_{b=1}^nf_c^{(r)}(X_b)= \frac{1}{n}\sum_{b=1}^nk_c^{(r)}(X_i, X_b).$$
The sample *[r]{}th kernel moment element of the $ {\mathbf}{\Phi}(X_i) $ is a solution of $$\begin{aligned}
\label{cmop}
{\operatornamewithlimits{argmin}}_{\otimes f_c\in \otimes{\mathcal}{H}_X} \frac{1}{n}\sum_{i=1}^n\| \otimes_{j=1}^r{\mathbf}{\Phi}_c^{(r)}(X_i)- \otimes^r f_c\|^2_{\otimes{\mathcal}{H}_X}\nonumber\\
={\operatornamewithlimits{argmin}}_{\otimes g_c\in \otimes{\mathcal}{H}_X} \frac{1}{n}\sum_{i=1}^n\| \otimes_{j=1}^r{\mathbf}{\Phi}_c^{(r)}(X_i)- g\|^2_{\otimes{\mathcal}{H}_X}\end{aligned}$$ where ${\mathbf}{\Phi}_c^{(r)}(X_i)= {\mathbf}{\Phi}^{(r)}(X_i)-\frac{1}{n}\sum_{i=1}^n {\mathbf}{\Phi}^{(r)}(X_i)$, at point $X$ , $g_c(X, X,\cdots X)= f_c(X) f_c(X) \cdots f_c(X)=f_c^{(r)}(X)$ and $\otimes f_c = f_c\otimes f_c \otimes \cdots \otimes f_c$ is the tensor product of $r$ functions $f_c \in {\mathcal}{H}_X$.*
Influence function of mean element and cross-raw moment
-------------------------------------------------------
([**Influence function**]{}). Let ($\Omega$, ${\mathcal}{A}$, ${\mathbb}{P}$) be a probability space and $({\mathcal}{X}, {\mathcal}{B})$ a measure space. We want to estimate the parameter $\theta \in \Theta$ of a distribution $F$ in ${\mathcal}{A}$. We assume that exists a functional $R: {\mathcal}{D}(R) \to {\mathbb}{R}$, where $ {\mathcal}{D}(R)$ is the set of all probability distribution in ${\mathcal}{A}$. Let $G$ be some distribution in ${\mathcal}{A}$. If data do not fallow the model $F$ exactly but slightly going toward $G$, the Gâteaux Derivative at $F$ is given by $$\begin{aligned}
\lim_{\epsilon\to 0 }\frac{R[(1-\epsilon)F+\epsilon G] - R(F)}{\epsilon}\end{aligned}$$ Suppose $x\in {\mathcal}{X}$ and $G=\Delta_x$ is the probability measure which gives mass $1$ to $\{x\}$. The influence function (special case of Gâteaux Derivative) of $R$ at $F$ is defined by
$$\begin{aligned}
IF(x, R, F)=\lim_{\epsilon\to 0 }\frac{R[(1-\epsilon)F+\epsilon\Delta_x] - R(F)}{\epsilon}\end{aligned}$$
provided that the limit exists. It can be intuitively interpreted as a suitably normalized asymptotic influence of outliers on the value of an estimate or test statistic.
The equivalent definition can also be defined using the perturbation theory. Consider the case where $R(\epsilon) = R[(1-\epsilon)F+\epsilon G] - R(F)$ is expanded as a convergent power series of $\epsilon$ as $$\begin{aligned}
R(\epsilon)= R + \epsilon R^{(1)}+ \epsilon^2 R^{(2)}+ O(\epsilon^3) \nonumber \end{aligned}$$ Due to the properties of convergent power series $R(\epsilon)$ is differentiable in a neighborhood of $\epsilon=0$. The IF, $IF(x, R)$ equals to $ R^{(1)}$, the first order term of $\epsilon$. There are three properties of IF: gross error sensitivity, local shift sensitivity and rejection point. They measured the worst effect of gross error, the worst effect of rounding error and rejection point.
For a scalar, we just define influence function (IF) at a fixed point. But if the estimate is a function, we are able to express the change of the function value at every points [@Kim-12].
\[Kernel mean element\] Let $R(F_X)=\langle {\mathcal}{M}, f \rangle_{{\mathcal}{H}_X}= {\mathbb}{E}_X[f(X)]= \int f(X)dF_X= \int k(\cdot, X)dF_X.$ The value of parameter at the contamination model, $W^\epsilon= (1-\epsilon)F_X+\epsilon\Delta_{X^\prime}$ is given by $$\begin{aligned}
R\left[W^\epsilon\right]=R\left[(1-\epsilon)F_X+\epsilon\Delta_{X^\prime}\right]&=& \int f(\tilde{X})d[(1-\epsilon)F_X+\epsilon \Delta_{X^\prime}] \nonumber \\
&=&(1-\epsilon)\int f(\tilde{X})dF_X+\epsilon\int f(\tilde{X})d_{\Delta_{X^\prime}}) \nonumber \\
&=&(1-\epsilon)\int k(\tilde{X},X)dF_X+\epsilon\int k(\tilde{X}, X)d_{\Delta_{X^\prime}} \nonumber \\
&=& (1-\epsilon)\int k(X, \tilde{X})dF_X +\epsilon k(\tilde{X},X^\prime)\nonumber\\
&=& (1-\epsilon) R(F_X) + \epsilon k(\tilde{X},X^\prime) \nonumber\end{aligned}$$ Thus the IF of $R(F_X)$ at point $X^\prime$ for every point $\tilde{X}$ is given by $$\begin{aligned}
IF(X^\prime, \tilde{X}, R, F_X) &=&\lim_{\epsilon\to 0 }\frac{R[W_X^\epsilon] - R(F_X)}{\epsilon} \nonumber\\
&=&\lim_{\epsilon\to 0 }\frac{(1-\epsilon)RF_X+\epsilon k( \tilde{X},X^\prime)-R(F_X)]}{\epsilon} \nonumber\\
&=&\lim_{\epsilon\to 0 }\biggl[k(\tilde{X},X^\prime) - R(F_X) \biggr] \nonumber\\
&=&k( \tilde{X},X^\prime)-R(F_X)\nonumber\\
&=&k(\tilde{X}, X^\prime) - {\mathbb}{E}_X[k\tilde{X}, X)], \qquad \forall k(\cdot, X) \in {\mathcal}{H}_X. \nonumber\end{aligned}$$ Which is estimated with the data points $X_1, X_2, \cdots, X_n \in {\mathcal}{X}$ as $$\begin{aligned}
k(\tilde{X}, X^\prime) - \frac{1}{n}\sum_{i=1}^n k(\tilde{X}, X_i), \qquad \forall k(\cdot, X_i)\in {\mathcal}{H}_X,\,\tilde{X},\, X^\prime \in{\mathcal}{X}. \nonumber\end{aligned}$$
Let $R(F_{XY}) = {\mathbb}{E}_{XY}[ \langle k_X(\cdot, X), f_X \rangle_{{\mathcal}{H}_X} \langle k_Y(\cdot, Y), f_Y \rangle_{{\mathcal}{H}_Y}] = {\mathbb}{E}_{XY}[f_X(X)f_Y(Y)] = \int f_X(X) f_Y(Y)dF_{XY}$. The value of parameter at at $Z^\prime = (X^\prime, Y^\prime)$ the contamination data, for every point $\tilde{Z} = (\tilde{X},\tilde{Y})$ $W_{XY}^\epsilon= (1 - \epsilon)F_{XY} + \epsilon\Delta_{Z^\prime}$ is given by
$$\begin{aligned}
R[W_{XY}^\epsilon] = R[(1-\epsilon)F_{XY} + \epsilon\Delta_{Z^\prime}]&=& \int f_X(\tilde{X}) f_Y(\tilde{Y})d[(1 - \epsilon)F_{XY} + \epsilon \Delta_{Z^\prime}] \nonumber \\
&=&(1 - \epsilon)\int f_X(\tilde{X}) f_Y(\tilde{Y})dF_{X Y} + \epsilon\int f_X(\tilde{X})f_Y(\tilde{Y})d_{\Delta_{Z^\prime}} \nonumber \\
&=& (1-\epsilon)\int f_X(\tilde{X}) f_Y(\tilde{Y})dF_{XY} +\epsilon f_X(X^\prime)f_Y(Y^\prime)\nonumber\\
&=& (1-\epsilon) R(F_{XY}) + \epsilon f_X(X^\prime)f_Y(Y^\prime) \nonumber\end{aligned}$$
Thus the IF of $R(F_{XY})$ is given by $$\begin{aligned}
IF( \cdot, Z^\prime, R, F_{XY}) &=&\lim_{\epsilon\to 0 }\frac{R[W_{XY}^\epsilon] - R(F_{XY})}{\epsilon} \nonumber\\
&=& \lim_{\epsilon\to 0 }\frac{(1-\epsilon)RF_{XY}+\epsilon f_X(X^\prime)f_Y(Y^\prime) - R(F_{XY})}{\epsilon} \nonumber\\
&=& \lim_{\epsilon\to 0 }\biggl[f_X(X^\prime)f_Y(Y^\prime) - R(F_{XY}) \biggr] \nonumber\\
&=& f_X(X^\prime)f_Y(Y^\prime) - R(F_{XY})\nonumber\\
&=& k_X(\tilde{X}, X^\prime) k_Y(\tilde{Y}, Y^\prime) -{\mathbb}{E}_{XY}[\langle k_X(\tilde{X}, X), f_X \rangle_{{\mathcal}{H}_X} \langle k_Y(\tilde{Y},Y), f_Y \rangle_{{\mathcal}{H}_Y}]. \nonumber\end{aligned}$$ Which is estimated as $$\begin{aligned}
k_X(X_i, X^\prime) k_Y(X_i, Y^\prime) - \frac{1}{n}\sum_{b=1}^n k_X(X_i, X_b)k_Y(Y_i, Y_b)\nonumber\end{aligned}$$
An cross-covariance operator of $(X, Y)$, $\Sigma_{YX}: {\mathcal}{H}_X \to {\mathcal}{H}_Y$ is defined as $$\begin{aligned}
R(F_{YX}) &=& \langle f_Y, \Sigma_{YX}f_X \rangle_{{\mathcal}{H}_Y} ={\mathbb}{E}_{XY}[ \langle k_X(\cdot, X)-{\mathcal}{M}[F_X], f_X \rangle_{{\mathcal}{H}_X} \langle k_Y(\cdot,Y)-{\mathcal}{M}[F_Y], f_Y \rangle_{{\mathcal}{H}_Y}] \nonumber \\
&=& {\mathbb}{E}_{XY}[(f_X(X)- {\mathbb}{E}_{X}[f(X)]) (f_Y(Y)- {\mathbb}{E}_{Y}[f_Y(Y)])] \nonumber \\
&=&{\mathbb}{E}_{XY}[f_X(X) g_Y(Y)] - {\mathbb}{E}_{X}[f_X(X)] {\mathbb}{E}_Y[g_Y(Y)] \nonumber\end{aligned}$$ for $f_X\in {\mathcal}{H}_X$ and $f_Y\in {\mathcal}{H}_Y$. The IF of $R(F_{XY})$ at $Z^{\prime}= (X^\prime, Y^\prime)$ using the rule of IF of complicated statistics is given by $$\begin{aligned}
\rm{IF}(\cdot, Z^\prime, R, F_{XY}) &=& f_X(X^\prime)f_Y(Y^\prime)- {\mathbb}{E}_{XY}[f_X(X)f_Y(Y)]- {\mathbb}{E}_{Y}[ f_Y(Y)][f(X) - {\mathbb}{E}_{X}[ f_X(X)]]\nonumber \\
&-& {\mathbb}{E}_{X}[ f_X(X)][ f_Y(X )- {\mathbb}{E}_{X}[ f_Y(Y)] \nonumber \\
&=& [ f_Y(X)- {\mathbb}{E}_{X}[ f_X(X)]][ f_Y(Y)- {\mathbb}{E}_{Y}[ f_Y(Y)]]- R(F_{XY}) \nonumber \\
&=& \langle k_X(\cdot, X)-{\mathcal}{M}[F_X], f_X \rangle_{{\mathcal}{H}_X} \langle k_Y(\cdot,Y){\mathcal}{M}[F_Y], f_Y \rangle_{{\mathcal}{H}_Y} \nonumber \\
&-&{\mathbb}{E}_{XY}[ \langle k_X(\cdot, X)-{\mathcal}{M}[F_X], f_X \rangle_{{\mathcal}{H}_X} \langle k_Y(\cdot,Y)-{\mathcal}{M}[F_Y], f_Y \rangle_{{\mathcal}{H}_Y}] \nonumber
\end{aligned}$$
Which is estimated with the data points $(X_1, Y_1), (X_2, Y_2), \cdots, (X_n, Y_n) \in {\mathcal}{X}\times {\mathcal}{Y}$ as $$\begin{aligned}
\left[k_X(X_i, X^\prime) - \frac{1}{n}\sum_{b=1}^n k_X(X_i, X_b)\right] \left[k_Y(Y_i, Y^\prime)-\frac{1}{n}\sum_{b=1}^n k_Y(Y_i, Y_b)\right]- \nonumber\\
\frac{1}{n}\sum_{j=1}^n\left[k_X(X_i, X_j) - \frac{1}{n}\sum_{b=1}^n k_X(X_i, X_b)] [k_Y(Y_i, Y_j) - \frac{1}{n}\sum_{d=1}^n k_Y(Y_i, Y_d)\right] \nonumber\end{aligned}$$
Robust kernel cross-covariance operato
--------------------------------------
\[lemma1\] Under the assumptions (i) and (ii) the Gâteaux differential of the objective function $J$ at $g_1\in \otimes^r {\mathcal}{H}$ and incremental $g_2 \in \otimes^r {\mathcal}{H}$ is $$\delta J(g_1, g_2)=-\langle G(g_1), g_2\rangle_{\otimes^r {\mathcal}{H}},$$ where $G: \otimes^r {\mathcal}{H} \to \otimes^r {\mathcal}{H}$ is defined as $$G(g_1)=\frac{1}{n}\sum_{i=1}^n\phi (\|\otimes^r{\mathbf}{\Phi}_c(X_i) - g_1\|_{\otimes^r {\mathcal}{H}})\cdot (\otimes^r{\mathbf}{\Phi}_c(X_i) - g_1).$$ The necessary condition for $g_1=\hat{f} ^{(r)}$, the kernel central moment element is $G(g_1)={\mathbf}{0}$
Under the same assumption of Lemma \[lemma1\], $r$th robust kernel central moment element (robust kernel CME) at $X$ is given as $$\hat{f}_c^{(r)} (X, X, \cdots, X) ^{(h)} = \sum_{i=1}^n w_i^{(h-1)}\tilde{k}(X, X_i) \tilde{k}(X, X_i)\cdots \tilde{k}(X, X_i)$$ where $w_i^{(h)}=\frac{\varphi(\| \otimes_{b=1}^r{\mathbf}{\Phi}_c^{(b)}(X_i)- g_c\|_{\otimes{\mathcal}{H}_X})}{\sum_{a=1}^n\varphi(\| \otimes_{b=1}^r{\mathbf}{\Phi}_c^{(b)}(X_i)- g_c\|)_{\otimes{\mathcal}{H}_X}}\,, \rm{and} \, \varphi(x)=\frac{\zeta^\prime(x)}{x}.$ Putting the different value of $r$, we get the different robust kernel moment estimates.
Under the same assumption of Lemma \[lemma1\], kernel CO at $(X,X)$ and kernel CCO at $(X, Y)$ are estimated as $$\widehat{\Sigma}_{X X}^{(h)}= \sum_{i=1}^n w_i^{(h-1)}\tilde{k}(X, X_i)\tilde{k}(X, X_i),\qquad
\widehat{\Sigma}_{X Y}^{(h)}= \sum_{b=1}^n w_b^{(h-1)}\tilde{k}(X, X_b) \tilde{k}(Y, Y_i),$$ respectively and $w_b^{(h)}$ is the same as in Lemma Eq. (\[lemma1\]) with $r=2$.
### Proof of Theorem $4.1$: Influence function of kernel CCA
As in [@Fukumizu-SCKCCA], using the cross-covariance operator of (X,Y), $\Sigma_{XY}: {\mathcal}{H}_Y\to {\mathcal}{H}_X$ we can reformulate the optimization problem of classical kernel canonical correlation (classical kernel CCA) as follows: $$\begin{aligned}
\label{ckcca2}
\sup_{\substack{f_{X}\in {\mathcal}{H}_X, f_{Y}\in {\mathcal}{H}_Y \\ f_{X}\ne 0,\,f_{Y}\ne 0}}\langle f_X,\Sigma_{XY}f_Y\rangle_{{\mathcal}{H}_X}\qquad
\text{subject to}\qquad
\begin{cases}
\langle f_X, \Sigma_{XX}f_X\rangle_{{\mathcal}{H}_X}=1,
\\
\langle f_Y, \Sigma_{YY}f_Y\rangle_{{\mathcal}{H}_Y}=1.
\end{cases}\end{aligned}$$ Using generalized eigenvalue problem, we can derive the solution of Eq. (\[ckcca2\]) as with liner CCA [@Anderson-03]. $$\begin{aligned}
\label{ckcca3}
\begin{cases}
\Sigma_{XY}f_X - \rho\Sigma_{YY} f_Y = 0,
\\
\Sigma_{XY}f_Y - \rho\Sigma_{XX}f_X = 0.
\end{cases} \nonumber\end{aligned}$$ After some simple calculation, we can reset the solution as a single matrix equation for $f_X$ or $f_Y$. $$\begin{aligned}
\label{ckcca4}
\begin{cases}
(\Sigma_{XY} \Sigma_{YY}^{-1}\Sigma_{XY} - \rho^2 \Sigma_{XX})f_X = 0,
\\
(\Sigma_{XY} \Sigma_{XX}^{-1}\Sigma_{XY}-\rho^2 \Sigma_{YY})f_Y = 0.
\end{cases}\end{aligned}$$ The generalized eigenvalue problem in Eq. (\[ckcca4\]) (for simplicity we use first equation only) can be formulated as a simple eigenvalue problem using [*j*]{}th eigenfunction. $$\begin{aligned}
\label{ckcca5}
(\Sigma_{XX}^{- \frac{1}{2}} \Sigma_{XY} \Sigma_{YY}^{-1}\Sigma_{YX}\Sigma_{XX}^{- \frac{1}{2}} - \rho_j^2I) \Sigma_{XX}^{\frac{1}{2}}f_{jX}&=&0 \nonumber \\
\Rightarrow (\Sigma_{XX}^{-\frac{1}{2}} \Sigma_{XY} \Sigma_{YY}^{-1}\Sigma_{YX}\Sigma_{XX}^{- \frac{1}{2}} - \rho_j^2I) f_{jX}& = &0\end{aligned}$$
To use the results of IF of liner principle components analysis [@Tanaka-88], IF of liner canonical correlation analysis [@Romanazii-92] and IF of kernel principle component analysis [@Huang-KPCA] for the finite dimension and for the infinite dimension, respectively, we convert generalized eigenvalue problem of kernel canonical correlation analysis into a simple eigenvalue problem. Thus, we need to find, the IF of $ \Sigma_{XX}^{- \frac{1}{2}} \Sigma_{XY} \Sigma_{YY}^{-1}\Sigma_{YX}\Sigma_{XX}^{- \frac{1}{2}}$ and henceforth IF of $\Sigma_{YY}^{-1}, \Sigma_{XX}^{\frac{1}{2}}$ and $\Sigma_{XY}$. Let $\Sigma_{XY}$ be the covariance of the random vectors $k_X(\cdot, X)$ and $k_Y(\cdot, Y)$ on RKHS i.e., kernel covariance operator, $\Sigma_{YX}: {\mathcal}{H}_X\to{\mathcal}{H}_Y$, for all $f_X\in {\mathcal}{H}$ and $f_Y\in {\mathcal}{H}_Y$ we have $$\begin{aligned}
&{\mathbb}{E}_{XY}[ \langle f_X, k_X(\cdot, X)-{\mathcal}{M}_X\rangle_{{\mathcal}{H}_X} \langle k_Y(\cdot, Y)-{\mathcal}{M}_Y, f_Y\rangle_{{\mathcal}{H}_Y}]\nonumber\\
&= {\mathbb}{E}_{XY}[ \langle f_Y, ((k_X(\cdot, X)-{\mathcal}{M}_X)\otimes(k_Y(\cdot, Y)-{\mathcal}{M}_Y)) f_X\rangle_{{\mathcal}{H}_Y}]\nonumber\\
&= \langle f_Y, {\mathbb}{E}_{XY}(k_X(\cdot, X) - {\mathcal}{M}_X)\otimes(k_Y(\cdot, Y) - {\mathcal}{M}_Y)) f_X\rangle_{{\mathcal}{H}_Y}\nonumber\\
&=\langle f_Y, \Sigma_{YX}f_X\rangle_{{\mathcal}{H}_Y} \end{aligned}$$ where ${\mathcal}{M}_X$ is kernel mean elements in ${\mathcal}{H}_X$ and $\Sigma_{YX}= (k_X(\cdot, X) - {\mathcal}{M}_X)\otimes(k_Y(\cdot, Y) - {\mathcal}{M}_Y)$, since $((T_1\otimes T_2)(x) = \langle x,T_2\rangle T_1$). Using simple algebra we have at $Z^\prime= (X^\prime, Y^\prime)$
$$\begin{aligned}
&\rm{IF}(\cdot, X^\prime, \Sigma_{XX})= (k_X(\cdot, X^\prime)-{\mathcal}{M}_X) \otimes(k_X(\cdot, X^\prime)-{\mathcal}{M}_X)- \Sigma_{XX}, \nonumber\\
&\rm{IF}(\cdot, Y^\prime, \Sigma_{YY})= (k_Y(\cdot, Y^\prime)-{\mathcal}{M}_Y) \otimes (k_Y(\cdot, Y^\prime)-{\mathcal}{M}_Y)- \Sigma_{YY},\nonumber\\
&\rm{IF}(\cdot, Z^\prime, \Sigma_{XY})= (k_X(\cdot, X^\prime) -{\mathcal}{M}_X)\otimes (k_Y(\cdot, Y^\prime)-{\mathcal}{M}_Y)- \Sigma_{XY} \, \rm{and} \nonumber\\
&\rm{IF}(Z^\prime, X^\prime,
\Sigma_{XX}^{-\frac{1}{2}} )= \frac{1}{2} [\Sigma_{XX}^{-\frac{1}{2}}- \Sigma_{XX}^{-\frac{1}{2}}(k_X(\cdot, X^\prime)-{\mathcal}{M}_X)\otimes (k_X(\cdot, X^\prime)-{\mathcal}{M}_X) \Sigma_{XX}^{-\frac{1}{2}}]. \nonumber
\label{IFIVCOV}\end{aligned}$$
For simplicity, let us define $\tilde{k}_X (\cdot, X^\prime):= k_X(\cdot, X^\prime)-{\mathcal}{M}_X, \, \tilde{k}_Y (\cdot, {\mathbf}{y}^\prime):= k_X(\cdot, Y^\prime) - {\mathcal}{M}_Y\, \rm{and}\, {\mathbb}{A}:= \Sigma_{XY} \Sigma_{YY}^{-1}\Sigma_{YX}, \, {\mathbb}{B}:= \Sigma_{XX}^{- \frac{1}{2}}{\mathbb}{A}\Sigma_{XX}^{- \frac{1}{2}}$, and ${\mathbb}{L}= \Sigma_{XX}^{- \frac{1}{2}}(\Sigma_{XX}^{- \frac{1}{2}} \Sigma_{XY} \Sigma_{YY}^{-1} \Sigma_{YX}\Sigma_{XX}^{- \frac{1}{2}}-\rho^2{\mathbf}{I})^{-1}\Sigma_{XX}^{- \frac{1}{2}}$ Now, $$\begin{aligned}
&\rm{IF} (\cdot, Z^\prime, {\mathbb}{A})= \rm{IF}({\mathbf}{x}^\prime, Y^\prime, \Sigma_{XY}) \Sigma^{-1}_{YY}\Sigma_{YX} + \Sigma_{XY} \rm{IF}(X^\prime, Y^\prime, \Sigma_{YY}^{-1})\Sigma_{YX}+ \Sigma_{XY} \Sigma_{YY}^{-1}\rm{IF}(X^\prime, Y^\prime, \Sigma_{XY}) \nonumber\\
&= \left[\tilde{k}_X (\cdot, X^\prime)\otimes \tilde{k}_Y (\cdot, Y^\prime)- \Sigma_{XY}\right]\Sigma_{YY}^{-1}\Sigma_{YX}
+ \Sigma_{XY}\left[\Sigma_{YY}^{-1}-\Sigma_{YY}^{-1}\tilde{k}_Y (\cdot, Y^\prime)\otimes \tilde{k}_Y (\cdot, Y^\prime) \Sigma_{YY}^{-1}\right] \Sigma_{YX}\nonumber\\
&\qquad \qquad\qquad \qquad\qquad \qquad\qquad \qquad+ \Sigma_{XY}\Sigma_{YY}^{-1}\left[\tilde{k}_X (\cdot, X^\prime)\otimes \tilde{k}_Y (\cdot, Y^\prime)- \Sigma_{YX}\right]\nonumber\\
&=2 \Sigma_{XY}\Sigma^{-1}_{YY}\left[\tilde{k}_X(\cdot, X^\prime)\otimes \bar{k}_Y(\cdot, Y^\prime)- \Sigma_{XY}\right]
+ \Sigma_{XY}\left[\Sigma_{YY}^{-1}-\Sigma_{YY}^{-1}\tilde{k}_Y (\cdot, Y^\prime)\otimes \tilde{k}_Y (\cdot, Y^\prime) \Sigma_{YY}^{-1}\right] \Sigma_{YX} \nonumber\end{aligned}$$ Then, $$\begin{gathered}
\Sigma_{XX}^{-\frac{1}{2}} \rm{IF}(Z^\prime, {\mathbb}{A}) \Sigma_{XX}^{-\frac{1}{2}}= 2 \Sigma_{XX}^{-\frac{1}{2}}\Sigma_{XY}\Sigma^{-1}_{YY}[ \tilde{k}_X (\cdot, X^\prime)\otimes \tilde{k}_Y (\cdot, Y^\prime)- \Sigma_{XY}] \Sigma_{XX}^{-\frac{1}{2}}\nonumber\\
+\Sigma_{XX}^{-\frac{1}{2}}\Sigma_{XY}[\Sigma_{YY}^{-1}-\Sigma_{YY}^{-1}[\tilde{k}_Y (\cdot,{\mathbf}{y}^\prime)\otimes \tilde{k}_Y (\cdot,{\mathbf}{y}^\prime)] \Sigma_{YY}^{-1}] \Sigma_{YX}\Sigma_{XX}^{-\frac{1}{2}} \nonumber\end{gathered}$$ and $$\begin{gathered}
\rm{IF}(X^\prime, \Sigma_{XX}^{-\frac{1}{2}}){\mathbb}{A} \Sigma_{XX}^{-\frac{1}{2}}+ \Sigma_{XX}^{-\frac{1}{2}}{\mathbb}{A} \rm{IF}(X^\prime, \Sigma_{XX}^{-\frac{1}{2}})
= 2 \rm{IF}(X^\prime, \Sigma_{XX}^{-\frac{1}{2}}){\mathbb}{A} \Sigma_{XX}^{-\frac{1}{2}}
= [\Sigma_{XX}^{-\frac{1}{2}}- \Sigma_{XX}^{-\frac{1}{2}}\tilde{k}_X (\cdot, X^\prime)\otimes \tilde{k}_X (\cdot, X^\prime) \Sigma_{XX}^{-\frac{1}{2}}] {\mathbb}{A} \Sigma_{XX}^{-\frac{1}{2}}\nonumber\end{gathered}$$ The influence of ${\mathbb}{B}$ is given by $$\begin{aligned}
&\rm{IF} (X^\prime, Y^\prime,{\mathbb}{B})= 2\rm{IF} (X^\prime, Y^\prime, \Sigma_{XX}^{-\frac{1}{2}}) \Sigma_{XY} \Sigma_{YY}^{-1}\Sigma_{XY} \Sigma_{XX}^{-\frac{1}{2}}+ \Sigma_{XX}^{-\frac{1}{2}} \rm{IF}(X^\prime, Y^\prime, {\mathbb}{A}) \Sigma_{XX}^{-\frac{1}{2}} \nonumber\\
&= [\Sigma_{XX}^{-\frac{1}{2}}- \Sigma_{XX}^{-\frac{1}{2}}\tilde{k}_X (\cdot, X^\prime)\otimes \tilde{k}_X (\cdot, X^\prime)\Sigma_{XX}^{-\frac{1}{2}}] \Sigma_{XY} \Sigma_{YY}^{-1}\Sigma_{XY}\Sigma_{XX}^{-\frac{1}{2}} \nonumber\\
& \qquad \qquad \qquad\qquad + 2 \Sigma_{XX}^{-\frac{1}{2}}\Sigma_{XY}\Sigma^{-1}_{YY}[ \tilde{k}_X (\cdot, X^\prime)\otimes \tilde{k}_Y (\cdot, Y^\prime) - \Sigma_{XY}] \Sigma_{XX}^{-\frac{1}{2}}\nonumber\\
&\qquad\qquad\qquad\qquad+ \Sigma_{XX}^{-\frac{1}{2}}\Sigma_{XY}[\Sigma_{YY}^{-1} - \Sigma_{YY}^{-1}\tilde{k}_Y (\cdot, Y^\prime)\otimes \tilde{k}_Y (\cdot, Y^\prime) \Sigma_{YY}^{-1}] \Sigma_{YX}\Sigma_{XX}^{-\frac{1}{2}}\nonumber \\
&= - \Sigma_{XX}^{-\frac{1}{2}}\tilde{k}_X (\cdot, X^\prime)\otimes \tilde{k}_X (\cdot, X^\prime) \Sigma_{XX}^{-\frac{1}{2}} \Sigma_{XY} \Sigma_{YY}^{-1}\Sigma_{XY}\Sigma_{XX}^{-\frac{1}{2}} \nonumber \\
& \qquad \qquad \qquad\qquad+ 2 \Sigma_{XX}^{-\frac{1}{2}}\Sigma_{XY}\Sigma^{-1}_{YY} \tilde{k}_X (\cdot, X^\prime)\otimes \tilde{k}_Y (\cdot, X^\prime) \Sigma_{XX}^{-\frac{1}{2}}\nonumber \\
&\qquad\qquad \qquad\qquad- \Sigma_{XX}^{-\frac{1}{2}}\Sigma_{XY}\Sigma_{YY}^{-1}\tilde{k}_Y (\cdot, Y^\prime)\otimes \tilde{k}_Y (\cdot, Y^\prime) \Sigma_{YY}^{-1} \Sigma_{YX}\Sigma_{XX}^{-\frac{1}{2}}\end{aligned}$$ We convert generalized eigenvalue problem as a eigenvalue problem and use the Lemma 1 of [@Huang-KPCA] to define the IF of kernel CC, $\rho_j^2$ and kernel CVs, $f_X(X)$ and, $f_Y(Y)$. Then the IF of kernel $\rho_j^2$ is defined as $$\begin{aligned}
&\rm{IF} (Z^\prime,\rho_j^2)= \langle\tilde{f}_{jX}, \rm{IF} (Z^\prime, {\mathbb}{B}) \tilde{f}_{jX}\rangle_{{\mathcal}{H}_X\otimes {\mathcal}{H}_Y}\nonumber\\
&=-\langle \tilde{f}_{jX}, \Sigma_{XX}^{-1}\tilde{k}_X (\cdot, X^\prime)\otimes \tilde{k}_X (\cdot, X^\prime) \Sigma_{XX}^{-\frac{1}{2}} \Sigma_{XY} \Sigma_{YY}^{-1}\Sigma_{XY}\Sigma_{XX}^{-\frac{1}{2}} \tilde{f}_{jX}\rangle_{{\mathcal}{H}_X\otimes {\mathcal}{H}_X}\nonumber\\
& \qquad \qquad \qquad\qquad+\ 2\langle \tilde{f}_{jX}, \Sigma_{XX}^{-\frac{1}{2}}\Sigma_{XY}\Sigma^{-1}_{YY}\tilde{k}_X (\cdot, X^\prime)\otimes \tilde{k}_Y (\cdot, Y^\prime) \Sigma_{XX}^{-\frac{1}{2}} \tilde{f}_{jX}\rangle_{{\mathcal}{H}_X\otimes {\mathcal}{H}_Y}\nonumber\\
&\qquad \qquad \qquad\qquad-\langle \tilde{f}_{jX}^T, \Sigma_{XX}^{-\frac{1}{2}}\Sigma_{XY}\Sigma_{YY}\tilde{k}_Y (\cdot, Y^\prime)\otimes \tilde{k}_Y (\cdot, Y^\prime)\Sigma_{YX}\Sigma_{XX}^{-\frac{1}{2}} \tilde{f}_{jX}\rangle_{{\mathcal}{H}_Y\otimes {\mathcal}{H}_Y}\nonumber\\
&=-\langle\tilde{f}_{jX}, \Sigma_{XX}^{-1}\tilde{k}_X (\cdot, X^\prime)\otimes \tilde{k}_X (\cdot, X^\prime) \Sigma_{XX}^{-\frac{1}{2}} \Sigma_{XY} \Sigma_{YY}^{-1}\Sigma_{XY}\Sigma_{XX}^{-\frac{1}{2}} \tilde{f}_{jX}\rangle_{{\mathcal}{H}_X\otimes {\mathcal}{H}_X}\nonumber\\
& \qquad \qquad \qquad\qquad +2 \langle \tilde{f}_{jX}^T, \Sigma_{XX}^{-\frac{1}{2}}\Sigma_{XY}\Sigma^{-1}_{YY}\tilde{k}_X (\cdot, X^\prime)\otimes \tilde{k}_Y (\cdot, Y^\prime) \Sigma_{XX}^{-\frac{1}{2}} \tilde{f}_{jX}\rangle_{{\mathcal}{H}_X\otimes {\mathcal}{H}_Y}\nonumber\\
& \qquad \qquad \qquad\qquad -\langle\tilde{f}_{jX}, \Sigma_{XX}^{-\frac{1}{2}}\Sigma_{XY}[\Sigma_{YY}\tilde{k}_Y (\cdot, Y^\prime)\otimes \tilde{k}_Y (\cdot, Y^\prime) \Sigma_{YY}^{-1}] \Sigma_{YX}\Sigma_{XX}^{-\frac{1}{2}} \tilde{f}_{jX}\rangle_{{\mathcal}{H}_Y\otimes {\mathcal}{H}_Y}
\label{IFKCCA1}\end{aligned}$$ For simplicity we calculate in parts of Eq. (\[IFKCCA1\]). The first part derive as $$\begin{aligned}
&\langle\tilde{f}_{jX}, \Sigma_{XX}^{-1}\tilde{k}_X (\cdot, X^\prime)\otimes \tilde{k}_X (\cdot, X^\prime) \Sigma_{XX}^{-\frac{1}{2}} \Sigma_{XY} \Sigma_{YY}^{-1}\Sigma_{XY}\Sigma_{XX}^{-\frac{1}{2}} \tilde{f}_{jX}\rangle_{{\mathcal}{H}_X\otimes {\mathcal}{H}_X} \nonumber\\
&= \langle \Sigma_{XX}^{-\frac{1}{2}}\tilde{f}_{jX},\tilde{k}_X (\cdot, X^\prime)\otimes \tilde{k}_X (\cdot, X^\prime) \Sigma_{XX}^{-\frac{1}{2}} \Sigma_{XY} \Sigma_{YY}^{-1}\Sigma_{XY}\Sigma_{XX}^{-\frac{1}{2}}\Sigma_{XX}^{-\frac{1}{2}} \tilde{f}_{jX}\rangle_{{\mathcal}{H}_X\otimes {\mathcal}{H}_X} \nonumber\\
&= \langle f_{jX}, \tilde{k}_X(\cdot, X^\prime)\rangle_{{\mathcal}{H}_X} \langle \tilde{k}_X(\cdot, X^\prime), \Sigma_{XX}^{-\frac{1}{2}} \Sigma_{XY} \Sigma_{YY}^{-1}\Sigma_{XY}\Sigma_{XX}^{-\frac{1}{2}} f_{jX}\rangle_{{\mathcal}{H}_X}\nonumber\\
&= \rho_j^2 \bar{f}_{jX}^2(X^\prime),
\label{IFKCCA2}\end{aligned}$$ in the last equality, we use Eq. (\[ckcca5\]). The 2nd part of the Eq. (\[IFKCCA1\]) derive as $$\begin{aligned}
&\langle f_{jX}, \Sigma_{XX}^{-\frac{1}{2}}\Sigma_{XY}\Sigma^{-1}_{YY}[ \tilde{k}_X (\cdot, X^\prime)\otimes \tilde{k}_Y (\cdot, Y^\prime)] \Sigma_{XX}^{-\frac{1}{2}} f_{jX}\rangle_{{\mathcal}{H}_X\otimes {\mathcal}{H}_Y} \nonumber\\
&= \langle \Sigma_{XX}^{-\frac{1}{2}} \tilde{f}_{jX}, \tilde{k}_X (\cdot, X^\prime)\otimes \tilde{k}_Y (\cdot, Y^\prime)\Sigma_{XY}\Sigma^{-1}_{YY} \Sigma_{XX}^{-\frac{1}{2}} f_{jX}\rangle_{{\mathcal}{H}_X\otimes {\mathcal}{H}_Y} \nonumber\\
&=\langle f_{jX},\tilde{k}_X (\cdot, X^\prime)\otimes \tilde{k}_Y (\cdot, Y^\prime) \Sigma_{XY}\Sigma^{-1}_{YY} f_{jX}\rangle_{{\mathcal}{H}_X\otimes {\mathcal}{H}_Y} \nonumber\\
&= \rho_j \langle f_{jX}, \tilde{k}_X (\cdot, X^\prime)\rangle_{{\mathcal}{H}_X} \langle \tilde{k}_Y (\cdot, Y^\prime), f_{jY}\rangle_{{\mathcal}{H}_Y}
\nonumber\\
&= \rho_j \bar{f}_{jX}( X^\prime) \bar{f}_{jY}(Y^\prime),
\label{IFKCCA3}\end{aligned}$$ in the last second equality, we use Eq.(\[ckcca3\]). Similarly, we can write the 3rd term as $$\begin{aligned}
\label{IFKCCA4}
\langle \tilde{f}_{jX}, \Sigma_{XX}^{-\frac{1}{2}}\Sigma_{XY}[\Sigma_{YY}\tilde{k}_Y (\cdot, Y^\prime)\otimes \tilde{k}_Y (\cdot, Y^\prime) \Sigma_{YY}^{-1}] \Sigma_{YX}\Sigma_{XX}^{-\frac{1}{2}} \tilde{f}_{jX}\rangle_{{\mathcal}{H}_Y\otimes {\mathcal}{H}_Y} = \rho_j^2 \bar{f}_{jY}^2(Y^\prime)\end{aligned}$$ where $\bar{f}_{jX}= f_{jX}(X^\prime)= \langle f_{jX}, \tilde{k}_X (\cdot, X^\prime)$ and similar for $\bar{f}_{jY}$. Therefore, substituting Eq. (\[IFKCCA2\]), (\[IFKCCA3\]) and (\[IFKCCA4\]) into Eq. (\[IFKCCA1\]) the IF of kernel CC is given by $$\begin{aligned}
\rm{IF} (X^\prime, Y^\prime,\rho_j)= - \rho_j^2 \bar{f}_{jX}^2(Y^\prime)+2 \rho_j\bar{f}_{jX}(X^\prime) \bar{f}_{jY}(Y^\prime) -\rho_j^2 \bar{f}_{jY}^2(Y^\prime)\end{aligned}$$ Now we derive the IF of kernel Cvs. To this end first we need to derive $$\begin{aligned}
\rm{IF} (X^\prime, f_{jx})= \rm{IF} (X^\prime, \Sigma_{XX}^{-\frac{1}{2}}f_{jX})
= \Sigma_{XX}^{-\frac{1}{2}} \rm{IF} (X^\prime, f_{jX})+\rm{IF} (X^\prime, \Sigma_{XX}^{-\frac{1}{2}})f_{jX}
\label{CV1}\end{aligned}$$ By the first term of Eq. (\[CV1\]) we have $$\begin{aligned}
&\Sigma_{XX}^{-\frac{1}{2}} \rm{IF} (X^\prime, Y^\prime,f_{jX})= \Sigma_{XX}^{-\frac{1}{2}} ( {\mathbb}{B}-\rho^2{\mathbf}{I})^{-1}\rm{IF} (X^\prime, Y^\prime,{\mathbb}{B})f_{jX} \nonumber\\
&= - \Sigma_{XX}^{-\frac{1}{2}} ( {\mathbb}{B}-\rho^2{\mathbf}{I})^{-1}\big[ \Sigma_{XX}^{-\frac{1}{2}}\tilde{k}_X(\cdot, X^\prime)\otimes \tilde{k}_X(\cdot, X^\prime)\Sigma_{XX}^{-\frac{1}{2}} \Sigma_{XY} \Sigma_{YY}^{-1}\Sigma_{XY}\Sigma_{XX}^{-\frac{1}{2}}\nonumber\\
&+ 2 \Sigma_{XX}^{-\frac{1}{2}}\Sigma_{XY}\Sigma^{-1}_{YY}\tilde{k}_X(\cdot, X^\prime)\otimes \tilde{k}_Y(\cdot, Y^\prime) \Sigma_{XX}^{-\frac{1}{2}}- \Sigma_{XX}^{-\frac{1}{2}}\Sigma_{XY}\Sigma_{YY}\tilde{k}_Y(\cdot, Y^\prime) \Sigma^{-1}_{YY}\Sigma_{YX}\Sigma_{XX}^{-\frac{1}{2}}\big]\bar{f}_{jX}
\label{CV2}\end{aligned}$$
We derive each terms of Eq. (\[CV2\]), respectively. The first term of Eq. (\[CV2\]) is given by $$\begin{aligned}
&\Sigma_{XX}^{-\frac{1}{2}} ({\mathbb}{B}-\rho^2{\mathbf}{I})^{-1}[ \Sigma_{XX}^{-\frac{1}{2}} \langle \tilde{k}_X(\cdot, X^\prime)\otimes \tilde{ k}_X(\cdot, X^\prime), \Sigma_{XX}^{-\frac{1}{2}} \Sigma_{XY} \Sigma_{YY}^{-1}\Sigma_{XY}\Sigma_{XX}^{-\frac{1}{2}} f_{jX}\rangle \nonumber\\
&={\mathbb}{L}\langle\tilde{ k}_X(\cdot, X^\prime)\otimes \tilde{ k}_X(\cdot, X^\prime), \Sigma_{XX}^{-\frac{1}{2}} \Sigma_{XY} \Sigma_{YY}^{-1}\Sigma_{YX} f_{jX}\rangle \nonumber\\
&= {\mathbb}{L} \rho_j^2\langle\tilde{ k}_X(\cdot, X^\prime)\otimes \tilde{ k}_X(\cdot, {\mathbf}{x}^\prime), f_{jX}\rangle \\ \nonumber
&= {\mathbb}{L} \rho_j^2\langle\tilde{ k}_X(\cdot, X^\prime)\otimes \tilde{ k}_X(\cdot, X^\prime), f_{jX}\rangle \\ \nonumber
&= {\mathbb}{L} \rho_j^2\bar{f}(X^\prime) \tilde{k}(\cdot, X^\prime) \nonumber\end{aligned}$$ 2nd term of Eq. (\[CV2\]) is $$\begin{aligned}
&2{\mathbb}{L} \Sigma_{XX}^{-\frac{1}{2}}\Sigma_{XY}\Sigma^{-1}_{YY}\tilde{k}_X(\cdot, X^\prime) \tilde{k}_Y(\cdot, Y^\prime) \Sigma_{XX}^{-\frac{1}{2}} f_{jX} \nonumber\\
&= {\mathbb}{L} \langle \Sigma_{XX}^{-\frac{1}{2}}\Sigma_{XY}\Sigma^{-1}_{YY} f_{jX}, \tilde{k}_Y(\cdot, Y^\prime)\rangle \tilde{k}_X(\cdot, X^\prime) + {\mathbb}{L} \langle \Sigma_{XX}^{-\frac{1}{2}}\Sigma_{XY}\Sigma^{-1}_{YY}\tilde{k}_X(\cdot, X^\prime), f_{jX}\rangle \tilde{k}_Y(\cdot, Y^\prime)\nonumber\\
&= {\mathbb}{L} \rho_j \bar{f}_{jY}(Y^\prime)(\tilde{k}_X(\cdot, X^\prime)+ {\mathbb}{L} \Sigma_{XX}^{-\frac{1}{2}}\Sigma_{XY}\Sigma^{-1}_{YY} \bar{f}_{jX} (X^\prime) \tilde{k}_Y(\cdot, Y^\prime) \nonumber\end{aligned}$$ and the 3rd term of Eq. (\[CV2\]) is $$\begin{aligned}
{\mathbb}{L} \Sigma_{XY}\Sigma^{-1}_{YY}\tilde{k}_Y(\cdot, Y^\prime) \otimes \tilde{k}_Y(\cdot, Y^\prime) \Sigma_{YY}^{-1} \Sigma_{YX}\Sigma_{XX}^{-\frac{1}{2}}]f_{jX}
&=&{\mathbb}{L} \langle \Sigma_{XY}\Sigma^{-1}_{YY}\tilde{k}_Y(\cdot, Y^\prime) , \Sigma_{YY}^{-1}\Sigma_{YX}f_{jX}\rangle \tilde{k}_Y(\cdot, Y^\prime)\rangle \nonumber\\
&=&{\mathbb}{L} \langle \Sigma_{XY}\Sigma^{-1}_{YY}\tilde{k}_Y(\cdot, Y^\prime) , \rho _jf_{jX}\rangle \tilde{k}_Y(\cdot, Y^\prime) \rangle\nonumber\\
&=&{\mathbb}{L} \rho _j \Sigma_{XY}\Sigma^{-1}_{YY} \bar{f}_{jY} (Y^\prime) \tilde{k}_Y(\cdot, Y^\prime) \nonumber \end{aligned}$$
By substituting the above three equations into Eq. (\[CV2\]) we have $$\begin{aligned}
\Sigma_{XX}^{-\frac{1}{2}} \rm{IF} (\cdot, Z^\prime, f_{jX})&=& \Sigma_{XX}^{-\frac{1}{2}} ( {\mathbb}{B} - \rho^2{\mathbf}{I})^{-1}\rm{IF} (\cdot, Z^\prime,{\mathbb}{B})f_{jX} \nonumber\\
&=&-\rho_j (\bar{f}_{jY}(Y^\prime)-\rho_j \bar{f}_{jX}(X^\prime)){\mathbb}{L} \tilde{k} (\cdot, X^\prime) - (\bar{f}_{jX}(X^\prime) - \rho_j \bar{f}_{jY}(Y^\prime)) \Sigma_{XY}\Sigma^{-1}_{YY} \tilde{k}_Y(\cdot, Y^\prime)\nonumber\\
\label{CV6}\end{aligned}$$ The 2nd term of the Eq. (\[CV1\]) is give by $$\begin{aligned}
&\rm{IF} (X^\prime, \Sigma_{XX}^{-\frac{1}{2}}) f_{jX} \nonumber \\
&= - \langle f_{jX}, \Sigma_{XX}^{-1} f_{jX}\rangle \Sigma_{XX}^{-\frac{1}{2}} \rm{IF} (X^\prime, \Sigma_{XX}^{\frac{1}{2}}) \Sigma_{XX}^{-\frac{1}{2}} \tilde{f}_{jX} \nonumber \\
&= \langle f_{jX}, \Sigma_{XX}^{-\frac{1}{2}} \rm{IF} (X^\prime, \Sigma_{XX}^{\frac{1}{2}}) f_{jX}\rangle f_{jX}\nonumber \\
&= -\frac{1}{2}[\langle f_{jX}, \Sigma_{XX}^{-\frac{1}{2}} \rm{IF} (X, \Sigma_{XX}^{\frac{1}{2}}) f_{jX}\rangle + \langle f_{jX}, \rm{IF} (X^\prime, \Sigma_{XX}^{\frac{1}{2}}) \Sigma_{XX}^{-\frac{1}{2}} f_{jX}\rangle]f_{jX} \nonumber \\
&= - \frac{1}{2}[\langle f_{jX}, \Sigma_{XX}^{-\frac{1}{2}} \rm{IF} (X^\prime, \Sigma_{XX}) f_{jX}\rangle]f_{jX} \nonumber \\
&= -\frac{1}{2}[ \langle f_{jX}, (\tilde{k}_X(\cdot, X^\prime)-\Sigma_{XX} f_{jX})]f_{jX} \nonumber \\
&= -\frac{1}{2}[\bar{f}_{jX}(X^\prime)- \langle f_{jX},\Sigma f_{jX}\rangle]f_{jX}\nonumber\\
&= \frac{1}{2}[1- \bar{f}_{jX}(X^\prime)] f_{jX}
\label{CV7}\end{aligned}$$ Therefore, substituting Eq. (\[CV6\]) and Eq. (\[CV7\]) into Eq. (\[CV1\]) we get the IF of kernel canonical variate (CV) of $$\begin{gathered}
\rm{IF} (\cdots, X^\prime, Y^\prime,f_{jX})=-\rho_j (\bar{f}_{jY}(Y^\prime)-\rho_j \bar{f}_{jX}(X^\prime)){\mathbb}{L} \tilde{k} (\cdot, X^\prime)- (\bar{f}_{jX}(X^\prime)-\rho_j \bar{f}_{jY}(Y^\prime)){\mathbb}{L} \Sigma_{XY}\Sigma^{-1}_{YY} \tilde{k}_Y(\cdot, Y^\prime)\\ +\frac{1}{2}[1- \bar{f}^2(Y^\prime)]f_{jX} \nonumber\end{gathered}$$ Similarly, we can derive $\rm{IF} (\cdot, X^\prime, Y^\prime,f_{jY})$.
Let $({\mathbf}{X}_i, {\mathbf}{Y}_i)_{i=1}^n$ be a sample from the distribution $F_{XY}$. The empirical estimator of Eq. (\[ckcca2\]) and Eq. (\[ckcca4\]) are $$\begin{aligned}
\sup_{\substack{f_{X}\in {\mathcal}{H}_X,f_{Y}\in {\mathcal}{H}_X \\ f_{X}\ne 0,\,f_{Y}\ne 0}}\langle f_Y,\hat{\Sigma}_{YX}f_X\rangle_{{\mathcal}{H}_Y}\qquad
\text{subject to}\qquad
\begin{cases}
\langle f_X, (\hat{\Sigma}_{XX} + \kappa{\mathbf}{I}) f_X\rangle_{{\mathcal}{H}_X} = 1,
\\
\langle f_Y, (\hat{\Sigma}_{YY}+\kappa{\mathbf}{I})f_Y\rangle_{{\mathcal}{H}_Y} = 1,
\end{cases}
\label{ckcca7}\end{aligned}$$
$$\begin{aligned}
\label{ckcca8}
\begin{cases}
(\hat{\Sigma}_{XY} (\hat{\Sigma}_{YY} + \kappa{\mathbf}{I})^{-1}\hat{\Sigma}_{XY} - \rho^2 (\hat{\Sigma}_{XX}+\kappa{\mathbf}{I}))f_X = 0,
\\
(\hat{\Sigma}_{YX} (\hat{\Sigma}_{XX} + \kappa{\mathbf}{I})^{-1}\hat{\Sigma}_{YX} - \rho^2 (\hat{\Sigma}_{YY}+\kappa{\mathbf}{I}))f_Y = 0,
\end{cases}\end{aligned}$$
respectively.
Using the above equations, the empirical IF (EIF) of kernel CC and kernel CVs at $Z^\prime = (X^\prime, Y^\prime)$ for all points $Z_i=(X_i, Y_i)$ are $\rm{EIF} (Z^\prime, \rho_j^2)= \hat{\rm{IF}} (Z^\prime, \hat{\rho}_j^2),
\rm{EIF} (\cdot, Z^\prime, f_{jX}) = \hat{\rm{IF}} ( Z_i, Z^\prime, f_{jX}),
\rm{EIF} (\cdot, {\mathbf}{Z}^\prime, f_{jY})= \hat{\rm{EIF}} (Z_i, Z^\prime, \hat{f}_{jY})$, respectively.
| 2024-05-31T01:26:57.367551 | https://example.com/article/4688 |
Tag: employment
Indemnification provisions are present in many types of contracts- service agreements, leases, insurance contracts, and employment agreements – to name just a few. To “indemnify” someone is to promise to pay the cost of future damage, loss, or injury and therefore to protect them from future monetary damage. Indemnity provisions…
This month marks the 25th anniversary of the Americans with Disabilities Act (the “ADA”). This was the first comprehensive law to protect the civil rights of Americans with Disabilities. We are now used to seeing wheelchair ramps, handicapped parking spaces and telephone or ATM options for people with disabilities.…
If you are a stay home parent a divorce can be especially frightening. A person who is going through a divorce will experience major changes in his/ her personal life but work life may be consistent. However, a stay at home parent who needs to enter or reenter the… | 2024-03-07T01:26:57.367551 | https://example.com/article/6096 |
This invention relates to a control handle for controlling the speed and direction of travel of pallet trucks and other related vehicles.
Specifically, this invention relates to a control handle that may be used by an operator both while riding on a pallet truck and while walking along the side the truck. Such control handles have typically included a pair of twist grips which are rotated to control the speed and direction (forward and reverse) of travel of the truck. The twist grips are intended to be rotated in the same direction as the drive wheel of the truck; thus, rotating the twist grips forward will cause the vehicle to move in the forward direction, and rotating the twist grips down or in the reverse direction will cause the truck to move in the reverse direction. Operating the twist grips in the direction opposite to the direction of rotation of the drive wheel can also accomplish a slowing or braking function, although a separate brake is available to the operator. Moving the control handle from side to side controls steering of the truck.
It is sometimes desirable to operate a truck while walking along side of it, as shown in U.S. Pat. No. 5,245,144. In this mode of operation, one of the operator's hands grips the control handle to provide positive steering control of the truck, and some mechanism is provided to control speed and direction of travel. In the U.S. Pat. No. 5,245,144, a steering control unit includes a pair of switches on either side of the control handle which are actuated to move the truck at a predetermined slow speed; vertical movement of the steering control arm controls braking of the truck.
In other trucks, the steering control arm is oriented toward the operator's platform. In this type of vehicle, the control handle will be located over the truck's power unit, not ahead of the truck and waist high. In such vehicles, control of the truck has been by use of existing twist grips, butterfly buttons on the steering arm, separate push/pull buttons and the like.
It would be desirable to provide an integrated control handle for a pallet truck which permits walk-along control while making use of the existing twist grips and without the need for separate, complex control switches or circuits. | 2024-04-07T01:26:57.367551 | https://example.com/article/2503 |
Comments about Kay Ryan
Nothing Ventured
Nothing exists as a blockand cannot be parceled up.So if nothing's venturedit's not just talk;it's the big wager.Don't you wonderhow people thinkthe banks of space and time don't matter?How they'll drainthe big tanks down to slime and salamandersand want thanks? | 2024-06-18T01:26:57.367551 | https://example.com/article/1870 |
Large carnivores recorded via trapcameras in the Silesian Beskids Mountains by scientists from Association for Nature "Wolf" / Duże drapieżniki zarejestrowane za pomocą fotopułapek w Beskidzie Ślskim przez Stowarzyszenie dla Natury "Wilk".
click to wach video | 2024-06-15T01:26:57.367551 | https://example.com/article/4697 |
Whenever I travel abroad, I like to arrive with a few phrases in the local tongue. Fluency is another matter. The son of a language teacher, I learned French as a boy. I picked up Spanish during a ramble through Mexico in my 20s. Since then, I’ve been busy. Despite a lifelong yearning for Italian, I reached middle age without getting much beyond per favore and grazie. Recently, an invitation to a conference near Milan made me eager to venture further. But my days were crammed with work deadlines and family obligations; there was no room for an evening course or a regimen of home-instruction through an online app. Maybe, I conjectured, I could master la bella lingua by listening to recordings in my sleep.
Almost a century ago, a fad for sleep-learning swept the industrialised world, ending only after neuroscientists determined it was physiologically impossible. Yet today, a growing body of research suggests they were wrong. Sleep-learning appears to be heading for a revival, on a far more solid scientific basis than its earlier incarnation. By subjecting sleep to a few engineering fixes, we could minimise the time our brains are offline each night, gaining precious hours for absorbing information. Over many nights, we could vastly expand our stock of knowledge and skills, or even treat stubborn addictions and psychological traumas. All of which raises an unsettling question: should the prospect be welcomed or dreaded? If we harness sleep for self-improvement, will we lose something essential about ourselves?
The idea that humans can learn during slumber dates back at least to biblical times, when God gave Jacob a glimpse of his destiny in a dream of angels climbing a ladder to heaven. But the first person to make money from the concept was Alois Benjamin Saliger, a Czech-born New York-based businessman and inventor – ‘tall, spare, thin-lipped’, according to a contemporary account, ‘with piercing eyes and a wide forehead’ – who in 1932 patented the Psycho-Phone. A phonograph fitted with a repeating mechanism and a tiny acoustic horn, the device was meant to sit by a sleeper’s bed and replay spoken-word recordings at the volume of a whisper. It was marketed with disks whose titles included Prosperity, Inspiration, Normal Weight, and Mating. ‘I desire an ideal mate,’ Saliger intoned on the latter record. ‘I radiate love. I have a fascinating and attractive personality. I have a strong sex appeal.’ If the machine functioned as advertised, the user would awaken filled with irresistible confidence, ready to stride off and conquer his chosen territory.
The Psycho-Phone operated on the premise that people are as suggestible while asleep as they are under hypnosis, an unproven theory picked up by Aldous Huxley in his dystopian novel, Brave New World (1932). There, recorded messages are used to train sleeping children in the values of a soulless future society. A proud official in the book calls the new method, dubbed ‘hypnopaedia’ by Huxley, the ‘greatest moralising and socialising force of all time’.
Although hypnopaedia was never employed for mass indoctrination in the real world, it came to be widely used as a tool for teaching new skills or changing unwanted behaviour. Scientific studies seemed to show it worked. In one study, a group of sleeping men heard a recorded list of Chinese words and their English translations; the next day, they scored significantly better on a comprehension test than a control group. In another study, 20 boys with a nail-biting habit were played the phrase ‘my fingernails taste terribly bitter’ 300 times a night for 54 nights; by the end of the trial, 40 per cent had reportedly overcome their vice. The technique grew particularly popular in the Soviet Union, where whole villages were said to learn foreign languages while dozing.
The backlash came in the 1950s, when scientists began using electroencephalography (EEG), in which electrodes are used to measure the brain’s electrical activity. With this technique, they could finally determine whether subjects were, in fact, asleep rather than drifting near sleep or just resting. When the Rand Corporation researchers William Emmons and Charles Simon repeatedly played a list of 10 words to men whose EEGs showed an absence of alpha waves (a reliable gauge of sleep), their performance on a memory test upon awakening was no better than chance. Other EEG-monitored trials drew similar results. Scientific consensus soon concluded that the sleeping brain was incapable of absorbing outside information, and hypnopaedia was consigned to the realm of quackery.
Now the pendulum is swinging again. Although no practical method yet exists (despite the claims of online hucksters), recent studies suggest that hypnopaedia might be possible in principle. And, if certain technical challenges can be overcome, it could truly usher in a brave new world.
The new interest in sleep-learning comes from a deeper understanding of what our brains do as we lie inert and drooling. Experiments have long shown that people who get a good night’s sleep are better able to recall what they learned the day before than those who don’t. But why?
One theory held that the brain rehearsed the day’s new information during dreams, characterized by rapid eye movement (REM). Lab research eventually dispensed with that idea, however. By the mid-1990s, the data was pointing another way: memories are rehearsed mainly during the deep, dreamless stage known as slow-wave sleep (SWS), so-called because the brain cycles leisurely between peaks and valleys of electrical activity. Researchers found that during SWS, specialised ‘place cells’ in a rat’s hippocampus replay the route of a maze that the animal learned to run during the day. Later studies indicated that humans rehearse new memories in much the same way. The hippocampus functions as a temporary storehouse for memories until slower-growing but more permanent connections for those same memories form in the cerebral cortex, where language, sensory perception and thought reside.
As it became clear how much activity the sleeping brain devoted to digesting the day’s learning, the notion of learning while asleep came to seem less far-fetched.
In 1996, Japanese researchers put a rudimentary form of sleep-learning to the test by eliciting what psychologists call a conditioned response – linking two stimuli so that repeating the second one triggers a response ordinarily associated with the first. They delivered mild electric shocks to the legs of five sleepers while playing a tone; after awakening, the subjects experienced increased heart rates when they heard the tone alone. The study proved that the sleeping subjects recalled the tone, at least subconsciously.
A much more ambitious conditioning trial came in 2007, when a team, led by the medical psychologist Jan Born at the University of Lübeck, Germany, used olfactory cues to trigger fully conscious, or declarative, memories. They began by piping the odour of roses into subjects’ nostrils while they memorised the position of objects on a computer screen. Then, some of the subjects were exposed to the fragrance again during SWS. When they awoke, they recalled the locations of the objects 15 per cent more accurately than a control group not exposed to the smell during sleep.
In subsequent experiments, the neuroscientist Ken Paller at Northwestern University in Chicago taught subjects to play a short melody on a keyboard; later, as the subjects napped, the tune was played repeatedly to half of them. After waking up, that group played it with fewer mistakes than the cohort who had slept in silence. Another study was directly relevant to my linguistic quest: at the Swiss National Science Foundation, subjects who were taught Dutch words during the day proved better able to recall and translate them after those words were replayed during SWS at night. These music and language trials show that auditory cues trigger memory rehearsal for certain tasks directly – no need for conditioned association at all.
Anyone seeking to master a musical instrument or explore the intricacies of particle physics could do so with almost magical ease
Together, these studies offered something close to a proof of concept. But one element was missing: all the experiments involved skills or information that had first been learned during waking. To demonstrate hypnopaedia in action, scientists would have to teach a sleeping subject something new.
In March 2015, researchers reported in the journal Nature Neuroscience that they had done just that. The team, led by the neurobiologist Karim Benchenane of the National Centre for Scientific Research in Paris, began by implanting electrodes in the brains of mice, and recording the firing of place cells as the animals navigated a metre-wide circular platform. The researchers chose one cell for each mouse, then waited until it reactivated during sleep. At that moment, they zapped the brain’s reward centre with a low-voltage electric current, creating a sensation of intense pleasure. When the mice awoke, they rushed to the location associated with the place cell, lingering there in apparent expectation of another reward. The scientists had created a false memory, changing the animals’ behaviour in ways that did not result from prior waking experience.
Benchenane’s study used positive reinforcement to train its subjects. Meanwhile, neuroscientists at the Weizmann Institute of Science in Israel took the opposite approach with human nicotine addicts, exposing them during SWS to the odours of cigarettes and rotten eggs or fish. Thanks to that aversive conditioning, participants smoked 30 per cent less the following week.
And at Northwestern University, the neurology post-doc Katherina Hauner devised a way to erase negative associations during SWS. First, she showed volunteers images of faces while delivering an unpleasant electrical shock; at the same time, she exposed them to odours of mint, lemon or pine. The subjects soon learned to associate the faces with pain. Then, while they slept, Hauner exposed them to the odours alone. Initially, they responded with fear (as measured by tiny amounts of sweat on the skin), but the reaction diminished with repetition. When the subjects awakened, their anxiety at seeing the faces was reduced as well.
It’s easy to see how transformative such techniques could be. ‘We’re at the cutting edge of trying to figure out what the sleeping brain is capable of doing,’ Paller said when I called him in Chicago. In the near term, ‘we might improve learning by taking advantage of the processing that already happens during sleep’. An updated Psycho-Phone, designed to deliver cues during SWS, might help a student master a subject faster, though daytime study would still be necessary. The device might be paired with a headband delivering transcranial electric stimulation, which has been shown to deepen SWS at certain frequencies. A deluxe package could include an aroma-generating machine to boost the overall effect.
Sleep-learning could be used in therapy to replace traumatic memories with less charged recollections of the same events. That might require drugs to pry open the neural gates, allowing a Super-Psycho-Phone to funnel content directly into the hippocampus. Scientists at the University of California, Berkeley recently devised a method for recording a subject’s neural patterns as she watches a video, translating them into computer code, and generating a fuzzy approximation of the original imagery. Some day, perhaps, the process will be refined and reverse-engineered, producing neuro-downloadable multimedia courses in Italian for beginners.
The potential is obvious, and not only for over-scheduled language geeks such as me. Students in every field might achieve proficiency twice as fast as they do now, and go on to learn twice as much. Anyone seeking to acquire new job skills, master a musical instrument or explore the intricacies of particle physics could do so with almost magical ease.
But the risks seem palpable, too. Hypnopaedia could undermine the restorative functions that normally occur during sleep – for example, the pruning of excess neural connections to make room for memories to come. After a night of unconscious exam-cramming, it might be harder to learn anything new the next day. Learning while sleeping could also make less brain energy available for consolidating long-term memories, perhaps leading to the erasure of last year’s trip to Istanbul. It might disrupt the nocturnal cleansing of the brain’s metabolic wastes by glial cells, increasing the learner’s vulnerability to neurodegenerative ailments such as Alzheimer’s disease.
During sleep, the brain rebalances both the immune and endocrine systems; that’s why sleep disorders are associated with ills including depression, obesity, diabetes, cardiovascular disease and cancer. Tinkering with the controls might have nasty consequences.
Less tangible costs must also be factored in. I am old enough to remember life before smartphones, when being disconnected at certain times of day, or even for days at a time, was perfectly acceptable. That state of tranquil unreachability is a lost Eden. Now, if my editor emails an urgent query, I have no excuse not to respond during waking hours. Sleeping hours, at least, remain off-limits. But in the hypnopaedic future, that boundary, too, could fall away.
Imagine your boss asking you to download data for tomorrow morning’s presentation into your brain after lights-out. Imagine your coworkers discussing the TV shows they binge-watched in the arms of Morpheus. Picture your Facebook friends posting updates on their nightly absorption of Mandarin. What will we sacrifice by succumbing to the demands of 24-hour accessibility, availability, productivity?
In his book Healing Night (2006), the sleep psychologist Rubin Naiman tells of a game he played with his mother as a child. She would ask: ‘What is the best thing in the world?’ Little Rubin would shout out guesses (Toys! Cartoons! Ice‑cream!) until she revealed the correct answer: ‘Night.’ Naiman’s mother had spent four years in a Nazi concentration camp; during that hellish sojourn, she had learned to cherish the hours of darkness as a promised land. ‘Night brought sleep,’ he writes, ‘a vital daily measure of peace. Sleep, in turn, served as a natural bridge to dreams. And dreaming opened a mysterious portal to a more malleable and compassionate reality.’
‘We talk about falling in love and falling asleep. Both require a kind of surrender’
Such reverence is common in traditional societies, Naiman observes, but has been largely discarded in the over-illuminated West. Most of us see night as an inconvenience and sleep as nothing more than a means of recharging for the following day. We get by on as little sleep as possible, and take pills to knock ourselves out when it doesn’t arrive on schedule. Then we fuel ourselves with stimulants to compensate for the deficit in genuine rest. Naiman argues that our ‘wake-centric’ worldview is undermining our spiritual as well as our physical and psychological health. We could benefit, he suggests, by ‘restoring a sense of sacredness to our nights and night consciousness’.
When I asked Naiman his opinion of hypnopaedia, he likened it to using lovemaking to burn calories, or eating while using the toilet. Sleep is a time for digesting data, not ingesting it, he told me. It’s also a time for abandoning our diurnal strivings, for wandering among the inner mysteries. ‘We talk about falling in love and falling asleep,’ Naiman said. ‘Both require a kind of surrender.’ One of the chief pleasures of sleep – as of love – is that it pulls us out of time, into a realm where every moment is its own reward.
Hypnopaedia, by contrast, is about refusing to fall. It represents wake-centrism’s ultimate aim: to conquer night entirely, reducing it to a fully exploitable colony. In such a territory, going to bed without your headphones might become as unthinkable as leaving the house without your phone. Somewhere, surely, a latter-day Saliger is plotting to profit from such a scenario.
I hereby post my ‘Do not disturb’ sign. Please, Alois: just let me sleep. | 2024-07-11T01:26:57.367551 | https://example.com/article/2995 |
---
bibliography:
- 'DynamicAnalysis.bib'
title: 'Ball-Larus Path Profiling Across Multiple Loop iterations'
---
\[programmer workbench\] \[diagnostics, tracing\]
Algorithms, Measurement, Performance. Profiling, dynamic program analysis, instrumentation.
| 2024-05-29T01:26:57.367551 | https://example.com/article/3779 |
Q:
What is the difference between Atom and Neatbeans?
I've been working with netbeans for 6 months, but it doesn't seems to be very popular for web development (for example PHP), still I find that is a very good and complete tool. Someone told me to try Atom (which is very popular), because of esthetic and efficiency in therms of using what you actually need.
Do you think that Neatbeans is outdated compare to Atom?
Is there any important reason to change from an IDE to an Editor for web development on php?
A:
NetBeans is a full-featured IDE where you can e.g debug, Atom is - as you already wrote - a rich text editor, which is extendable with plugins (maybe also to debug PHP, but I really don’t know), I’m also not aware if Atom supports code completion, but in general I would say use what you like, I’m a Java developer and I do not want to miss NetBeans a single day, I also used it for PHP development and it’s also not bad, but if you fine working with Atom, than go with Atom, otherwise stay with NetBeans or discover other tools like IntelliJ Webstorm or Brackets. I think noone can give you a concrete answer to your question, sorry.
| 2023-08-01T01:26:57.367551 | https://example.com/article/4218 |
Pi(e) Day Party
Tricia
9 years ago
Advertisements
Pie Makers - Nov 2007
After I was reminded of its existence by a friend who sent along this post on “Nerdigras“, the family decided that on 3/14 at 1:59, we would host a Pi(e) Day Party. This appealed greatly to my math geek 11yo, as well as his party- and sweets-loving brothers. A party all about eating dessert? What’s not to love!
On very short notice, we invited 7 neighbor families and 9 school friend families. More than half came (about 30 people), most bearing pie (including pizza pie and a Serbian meat pie) – but I totally forgot to take pictures!
Great fun was had by all, so we might be making this an annual event! And even though we went to great effort to clean up and clear off the living room coffee table, and put the leaf in the dining room table, all the adults mostly hung out in the kitchen. Good thing we also cleared off flat surfaces in the kitchen! (We’re a family of pilers, so flat surfaces quickly become covered with stuff.) The dining room table did get used by the kids, and the coffee table was used today by a boy consortium playing some paper D&D style game, so it was not all for naught.
I made chocolate chip pie (favorite of T-boy and his mama) using this recipe (and a combo of Trader Joe’s and Spartan chocolate chips – sorry Nestle!), as well as pecan pie (favorite of Jonski Papa, C-boy, and Z-boy) using Grandma P’s recipe and Grandpa P’s pecans. As usual, I used the flaky buttery pastry recipe from The Heritage of Southern Cooking for the crusts.
Had I been truly inspired, I would have made another gooseberry pie using the berries in my freezer, but I wasn’t. Jonski Papa made 2 pizza pies (a cinnamon sugar one and a pepperoni) using his sourdough crusts, and a loaf of chocolate sourdough (not a pie at all, but it was round). As I said, I don’t have pictures of this weekend’s events, but I did take photos when T-boy and Jonski Papa made pecan pie for Thanksgiving 2007, so they will have to do. | 2024-05-27T01:26:57.367551 | https://example.com/article/4807 |
Bi-Folding Doors
Bi-Folding doors can create a whole new feel for your home, linking rooms together and seamlessly joining inside living with outside space. Versatile and aesthetically pleasing, bi-fold doors are the ideal replacement for existing patio or French doors to flood any room with natural light.
Both stylish and versatile, bi-folding doors offer a revolution in the creation of open space – making stylish 21st century living a reality in virtually any home. But, bi-folding doors need to do more than look good, they need to be secure too.
Unfortunately, the design of the locking hardware used on other bi-folds is actually far less secure than would be used on standard residential doorsand French doors. Yet, home owners are unwittingly expected to pay thousands of pounds to have them installed, leaving their homes vulnerable to intruders.
No Compromise – Superior Strength!
Doorsare at the heart of your home’s security, and we design our products to ensure the visitors to your home are only the welcome ones.
All of our bi-folds are designed and constructed to work as a complete system. Most other products differ as they are created by fixing together individual residential door profiles, which can produce inherent design flaws and areas of weakness that are detrimental to security performance.
Another added benefit to our bi-folding doors are the fully retractable locks – making our doors extremely reliable and minimising costly call backs for adjustments to the doors to ensure smooth operation is achieved. | 2023-09-05T01:26:57.367551 | https://example.com/article/6574 |
It said members have reported getting calls from someone claiming to be a PCSO asking for a financial contribution towards setting up a community scheme to help stop youths vandalising farms.
In return, the caller said the farm’s details would be added to a publication.
A similar scam has also been reported by CLA members in Nottinghamshire.
Nicola Currie, CLA eastern regional director, said: “The police do not cold call in order to request money for advertisements. If you receive a suspicious call and are unable to authenticate it, you are advised not give out any personal details, bank details or funds.
“Any representative from Suffolk’s force should always state their full name, job role and give their contact details, and give the person they are calling the option to call them back through the main switchboard to authenticate the call.”
Suffolk Police’s Rural Crime Officer Mark Bryant said: “It’s a scam, 100%, and farmers should just put the phone down. If you receive one of these calls you can alert the police by calling 101, as well as reporting it to BT as a nuisance call.”
Calls can also be reported to Action Fraud, a one-stop reporting centre handling all reports of fraud in the UK. It can be contacted online at www.actionfraud.police.uk or by calling 0300 123 2040. | 2023-08-07T01:26:57.367551 | https://example.com/article/2286 |
User Reviews
Kingdom of Dumpling
Posted by Anonymous on 04/07/2006
I like it!
I really enjoyed eating at David's Kitchen. The food was great, the price was right, and the service was friendly. What I *loved* about this very tiny restaraunt was its personal touch. The chef actually came and asked how our food was! My only problem with it was the atmosphere. This place has real potential: It is small, intimate, and has good windows. But the owner needs to work on the lighting and lack of decorations on the walls. It could feel very cozy, with only a small amount of work. | 2023-11-30T01:26:57.367551 | https://example.com/article/9442 |
Q:
Need help to change vertical spacing and margin in my messy document
Over time I have added a number of commands at the beginning of my document and I am getting lost and I now fail to simply modify vertical spacing and margin.
Current Document
Here are my packages and things
% ---------------------------
DOCUMENTCLASS
% ---------------------------
\documentclass[letterpaper,12pt]{scrartcl}
% --------------------------
BIBLIOGRAPHY
% ---------------------------
\usepackage[round,mcite]{natbib}
\bibliographystyle{plainnat}
% --------------------------
INPUTSTRUCTURE
% ---------------------------
\input{structure.tex}
% --------------------------
OTHER
% ---------------------------
\usepackage{pifont}% http://ctan.org/pkg/pifont
\usepackage{hhline}
\usepackage{multirow}
\newcommand{\xmark}{\ding{55}}
\hypersetup{citecolor=brown}
\hyphenation{Fortran hy-phen-ation}
and here is my structure.tex
\usepackage[
nochapters, % Turn off chapters since this is an article
beramono, % Use the Bera Mono font for monospaced text (\texttt)
eulermath,% Use the Euler font for mathematics
pdfspacing, % Makes use of pdftex’ letter spacing capabilities via the microtype package
dottedtoc % Dotted lines leading to the page numbers in the table of contents
]{classicthesis} % The layout is based on the Classic Thesis style
\usepackage{arsclassica} % Modifies the Classic Thesis package
\usepackage[T1]{fontenc} % Use 8-bit encoding that has 256 glyphs
%\usepackage[utf8]{inputenc} % Required for including letters with accents
\usepackage{graphicx} % Required for including images
\graphicspath{{Figures/}} % Set the default folder for images
\usepackage{enumitem} % Required for manipulating the whitespace between and within lists
\usepackage{lipsum} % Used for inserting dummy 'Lorem ipsum' text into the template
\usepackage{subfig} % Required for creating figures with multiple parts (subfigures)
\usepackage{amsmath,amssymb,amsthm} % For including math equations, theorems, symbols, etc
\usepackage{varioref} % More descriptive referencing
%----------------------------------------------------------------------------------------
% THEOREM STYLES
%---------------------------------------------------------------------------------------
\theoremstyle{definition} % Define theorem styles here based on the definition style (used for definitions and examples)
\newtheorem{definition}{Definition}
\theoremstyle{plain} % Define theorem styles here based on the plain style (used for theorems, lemmas, propositions)
\newtheorem{theorem}{Theorem}
\theoremstyle{remark} % Define theorem styles here based on the remark style (used for remarks and notes)
%----------------------------------------------------------------------------------------
% HYPERLINKS
%---------------------------------------------------------------------------------------
\hypersetup{
%draft, % Uncomment to remove all links (useful for printing in black and white)
colorlinks=true, breaklinks=true, bookmarks=true,bookmarksnumbered,
urlcolor=webbrown, linkcolor=RoyalBlue, citecolor=webgreen, % Link colors
pdftitle={}, % PDF title
pdfauthor={\textcopyright}, % PDF Author
pdfsubject={}, % PDF Subject
pdfkeywords={}, % PDF Keywords
pdfcreator={pdfLaTeX}, % PDF Creator
pdfproducer={LaTeX with hyperref and ClassicThesis} % PDF producer
}
Question
How can I modify vertical spacing (in between lines spacing) and margins?
What I tried
The following does not seem to have any impact on the document
\usepackage[body={4.8in,7.5in},
top=1.2in, left=1.8in]{geometry}
Same for
\usepackage[a4paper, total={6in, 8in}]{geometry}
I tried to switch from scrartcl to article but I did not achieve much.
I also tried things like
\usepackage[utf8]{inputenc}
\usepackage[english]{babel}
\setlength{\parindent}{4em}
\setlength{\parskip}{1em}
\renewcommand{\baselinestretch}{1.5}
but I often ended up receiving an error message although I failed to understand why and when.
A:
Firstly, you have:
% ---------------------------
DOCUMENTCLASS
% ---------------------------
This is missing a comment character before DOCUMENTCLASS which will result in an error (though I'm hoping that this is just an error in copy/pasting).
Now regarding modifying the margins, as you figured out, it's easiest to use the geometry package. Note that if another package is changing the page layout, you may want to use the \geometry command just before \begin{document}:
\geometry{vmargin=2cm,hmargin=3cm}
\begin{document}
Regarding the spacing between lines, Chris already provided an answer in his comment using \linespread{1.1} for example . You can also have a look at the setspace package which provides \singlespacing, \onehalfspacing and doublespacing (though that's typically used for creating proof-reading copies).
Finally, you should have a look through the log file, because you are getting a lot of warnings. For example, the KOMA classes (such as scrartcl) provide a lot of features which overlap with other packages (cfr also mentioned that in his comment just then). The overlap of these packages can result in features being partially overriden and then neither packages working properly.
Finally, you should typically prefer the subcaption package over the subfig package which you are using.
| 2024-03-04T01:26:57.367551 | https://example.com/article/2361 |
import { debugger } from "foo";
| 2024-07-24T01:26:57.367551 | https://example.com/article/3945 |
import pytest
import flask_praetorian
from flask import Flask, jsonify
from flask_sqlalchemy import SQLAlchemy
from flask_praetorian import Praetorian
from flask_mail import Mail
_db = SQLAlchemy()
_guard = Praetorian()
_mail = Mail()
class User(_db.Model):
"""
Provides a basic user model for use in the tests
"""
id = _db.Column(_db.Integer, primary_key=True)
username = _db.Column(_db.Text, unique=True)
password = _db.Column(_db.Text)
email = _db.Column(_db.Text)
roles = _db.Column(_db.Text)
@property
def rolenames(self):
try:
return self.roles.split(',')
except Exception:
return []
@classmethod
def lookup(cls, username):
return cls.query.filter_by(username=username).one_or_none()
@classmethod
def identify(cls, id):
return cls.query.get(id)
@property
def identity(self):
return self.id
class ValidatingUser(User):
__tablename__ = 'validating_users'
id = _db.Column(_db.Integer, primary_key=True)
username = _db.Column(_db.Text, unique=True)
password = _db.Column(_db.Text)
roles = _db.Column(_db.Text)
is_active = _db.Column(_db.Boolean, default=True, server_default='true')
def is_valid(self):
return self.is_active
__mapper_args__ = {'concrete': True}
@pytest.fixture(scope='session')
def app(tmpdir_factory):
"""
Initializes the flask app for the test suite. Also prepares a set of routes
to use in testing with varying levels of protections
"""
app = Flask(__name__)
# In order to process more requests after initializing the app,
# we have to set degug to false so that it will not check to see if there
# has already been a request before a setup function
app.debug = False
app.config['TESTING'] = True
app.config['SECRET_KEY'] = 'top secret'
_guard.init_app(app, User)
_mail.init_app(app)
app.mail = _mail
db_path = tmpdir_factory.mktemp(
'flask-praetorian-test',
numbered=True
).join('sqlite.db')
app.config['SQLALCHEMY_DATABASE_URI'] = 'sqlite:///' + str(db_path)
"""
Set to False as we don't use it, and to suppress warnings,
as documented by SQLALCHEMY.
"""
app.config['SQLALCHEMY_TRACK_MODIFICATIONS'] = False
_db.init_app(app)
with app.app_context():
_db.create_all()
@app.route('/unprotected')
def unprotected():
return jsonify(message='success')
@app.route('/protected')
@flask_praetorian.auth_required
def protected():
return jsonify(message='success')
@app.route('/protected_admin_required')
@flask_praetorian.auth_required
@flask_praetorian.roles_required('admin')
def protected_admin_required():
return jsonify(message='success')
@app.route('/protected_admin_and_operator_required')
@flask_praetorian.auth_required
@flask_praetorian.roles_required('admin', 'operator')
def protected_admin_and_operator_required():
return jsonify(message='success')
@app.route('/protected_admin_and_operator_accepted')
@flask_praetorian.auth_required
@flask_praetorian.roles_accepted('admin', 'operator')
def protected_admin_and_operator_accepted():
return jsonify(message='success')
@app.route('/undecorated_admin_required')
@flask_praetorian.roles_required('admin')
def undecorated_admin_required():
return jsonify(message='success')
@app.route('/undecorated_admin_accepted')
@flask_praetorian.roles_accepted('admin')
def undecorated_admin_accepted():
return jsonify(message='success')
@app.route('/reversed_decorators')
@flask_praetorian.roles_required('admin', 'operator')
@flask_praetorian.auth_required
def reversed_decorators():
return jsonify(message='success')
@app.route('/registration_confirmation')
def reg_confirm():
return jsonify(message='fuck')
return app
@pytest.fixture(scope='session')
def user_class():
"""
This fixture simply fetches the user_class to be used in testing
"""
return User
@pytest.fixture(scope='session')
def validating_user_class():
"""
This fixture simply fetches the validating user_class to be used in testing
"""
return ValidatingUser
@pytest.fixture(scope='session')
def db():
"""
This fixture simply fetches the db instance to be used in testing
"""
return _db
@pytest.fixture(autouse=True)
def db_setup(app, db):
"""
Prepares the testing database to hold testing data by creating the schema
anew for each test function and then dropping all data after the test runs
"""
with app.app_context():
db.create_all()
db.session.commit()
yield
db.drop_all()
db.session.commit()
@pytest.fixture(scope='session')
def default_guard():
"""
This fixtures fetches the flask-praetorian instance to be used in testing
"""
return _guard
@pytest.fixture(scope='session')
def mail():
"""
This fixture simply fetches the db instance to be used in testing
"""
return _mail
@pytest.fixture(autouse=True)
def clean_flask_app_config(app):
"""
This fixture ensures a clean `app.config` is available for each round
of testing.
"""
with app.app_context():
stock_config = app.config.copy()
yield
app.config = stock_config.copy()
| 2024-06-09T01:26:57.367551 | https://example.com/article/6629 |
Criterion A of the AMPD in HiTOP.
The categorical model of personality disorder classification in the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders (5th ed. [DSM-5]; American Psychiatric Association, 2013 ) is highly and fundamentally problematic. Proposed for DSM-5 and provided within Section III (for Emerging Measures and Models) was the Alternative Model of Personality Disorder (AMPD) classification, consisting of Criterion A (self-interpersonal deficits) and Criterion B (maladaptive personality traits). A proposed alternative to the DSM-5 more generally is an empirically based dimensional organization of psychopathology identified as the Hierarchical Taxonomy of Psychopathology (HiTOP; Kotov et al., 2017 ). HiTOP currently includes, at the highest level, a general factor of psychopathology. Further down are the five domains of detachment, antagonistic externalizing, disinhibited externalizing, thought disorder, and internalizing (along with a provisional sixth somatoform dimension) that align with Criterion B. The purpose of this article is to discuss the potential inclusion and placement of the self-interpersonal deficits of the DSM-5 Section III Criterion A within HiTOP. | 2023-09-06T01:26:57.367551 | https://example.com/article/7382 |
London-based BABB App Ltd. plans to soft launch its blockchain-based banking platform in Q2’17. The company is currently applying for a full banking license and is seeking to enter the UK’s Financial Conduct Authority (FCA) sandbox.
Built on a distributed ledger using Ethereum smart contracts, the BABB banking platform will operate a “decentralized information model” to minimize costs while maximizing data fidelity and security, the company said.
A multi-sided platform, BABB is designed to provide banking and other financial services to individuals and small businesses. Customer data will be encrypted and stored security on the blockchain where the permissions for who can access or use that will be entirely controlled directly by the customer.
It aims to provide “universal banking services for all” by using biometrics to facilitate client on-boarding and transaction authentication, Big Data analytics to optimize the quality of match-making and control risk, as well as AI-based customer self-service.
“I believe BABB’s plan to bring banking services to everyone, through the blockchain and the latest mobile technology, will both provide new features not seen in the high street banks and include many who have been excluded from the traditional financial system,” said Guido Branca, CEO of BABB App Ltd.
The platform will allow users to “leverage their social connectivity and their money in new ways” with more control and greater transparency. It will connect users to each other, allowing peer-to-peer transactions over the network. Beneficiaries and senders will be connected directly, without resorting to unnecessary middlemen or middle-layers of technologies.
BABB will be accessible through a mobile app. Users who download the app will need to provide a selfie and a voice prints to get a UK-based account upon activation. The company doesn’t consider itself as a challenger bank but rather as an enabler helping existing banks and financial institutions to innovate.
BABB App Ltd. is the latest company seeking to apply blockchain technology into banking. Another UK-based startup that intends to transform the industry is Secco Aura, a location-based, financial social networking app that allows users to exchange digital tokens that replace money in real-world transactions. It hopes to reinvent commerce and create a social network based on digital peer-to-peer currency.
Secco Aura is currently in the process of raising Series A funding with a target of March this year for a public beta launch. Pilots have been agreed for offices, campuses and shopping centers in the coming months, Chris Gledhill, CEO of Secco Aura, told FSTech earlier this month.
BABB App Ltd. is currently regulated as an Approved Payment Institution by the FCA. In June 2016, it raised seed funding and was accepted into the Level39 Accelerator Space in Canary Wharf. | 2024-07-19T01:26:57.367551 | https://example.com/article/7366 |
Divalent cations modulate human colon cancer cell adhesion.
Iatrogenic tumor implantation within surgical sites can compromise curative cancer surgery. Cancer cell adhesion to extracellular matrix proteins is mediated by diverse matrix receptors, most notably integrins. Divalent cations may modulate integrin-ligand interactions in some cells. We studied adhesion of SW620 and Caco-2 human colon cancer cells to collagen I, the dominant collagen of the interstitial matrix, and confirmed our results in primary human colon cancer cells from surgical specimens. Single cell suspensions in either HEPES/NaCl buffer or media supplemented with 0-1 mM Mn2+ or Mg2+, and 0-10 mM Zn2+ or Ca2+ were plated onto collagen-I-precoated dishes for 30 min. Supplementation of the HEPES/NaCl/BSA buffer with 1 mM Mn2+, Mg2+, Zn2+, or Ca2+ affected adhesion differently. Mn2+ (1 mM) markedly promoted SW620 adhesion vs control (21.17 +/- 0.08-fold). Mg2+ (1 mM) had a similar but lesser effect (14.71 +/- 0.02-fold). However, 1-10 mM Ca2+ inhibited basal cell adhesion by 22.0 +/- 3.1 to 88.0 +/- 7.3 % inhibition. Ca2+ (2.5-10 mM) also inhibited Mn2+-induced adhesion. Zn2+ stimulated basal adhesion slightly at lower concentrations but inhibited Mn2+-stimulated adhesion similarly to Ca2+ at higher concentrations. Results were duplicated in conventional serum containing culture medium supplemented with these cations. Caco-2 cells and primary cancer cells yielded similar results. All results are significant to P < 0.01. Integrin-mediated colon cancer cell adhesion is affected by extracellular divalent cation concentrations. Washing the surgical site with dilute calcium or zinc solutions might diminish perioperative tumor implantation. | 2024-05-21T01:26:57.367551 | https://example.com/article/8751 |
I am trying my hardest to get back over here and post more often. I can not seem to keep my head above the water with all the projects I have going!!!!
I have a request from a friend. Her 1st grandson was just born. Her son has been a member of the fire dept for years... she got to playing on etsy and saw some adorable firefighter dress up/prop outfits for pictures and asked me if I could make them. I have never done hats or clothes for an infant so I really would like to have a pattern that would give me an approx gauge, etc.
Does anyone know where I can find the pattern for a firefighter outfit... hat, diaper cover with suspender straps, and boots. A free pattern would be great but I know that is not likely! I have no problem buying the pattern as long as it is available PDF since I need to get this made up ASAP!!!!
Thank you sooooooo much to anyone that can help me with finding this pattern. I did a quick search online but it just continues to send me to pictures of the outfits on etsy. I know I can most likely just wing it and make it by guessing at the pattern but I really, really would love to a successful pattern written out in front of me to follow!!!!
You could use a basic baby cardigan pattern and just make it black or tan with bright yellow stripes. Same with booties, just make the cuff a bit longer and add a stripe to the toe and the top. Google 'fireman photo' to get an idea, the clothing is pretty simple just stripey.
Edit, I red too fast and missed the diaper covers with suspenders. Hmm, maybe have grandma measure for you and just add crocheted strips crossed in the back to a plain diaper cover pattern? | 2024-01-17T01:26:57.367551 | https://example.com/article/8307 |
Ashley Fosberg. Executive Director of the RAVE Foundation came to the Alliance Council meeting and recapped RAVE’s programs. She told the Council that RAVE received $12,000 in December as they were glassybaby’s charity of the month. Relatively soon, The Sounders will even have their own glassybaby candle that they can sell in the Pro Shop.
IV. Concessions
Ashli Brown, Michael Johnson, and Dawn Wheeler from concessions came to the council to discuss the changes coming to concessions. The Alliance Council was able to try some of the food and met the new Executive Chef, who comes over to CenturyLink from Safeco Field. The Seahawks have decided that they don’t want a concessionaire and, as such, they will do it themselves. They have created First and Goal Hospitality.
→ Executive Chef Michael Johnson
The key to his move from Safeco Field was the promise that the concessions would have a focus on sustainable and local. In the last 8 months, they have tripled the amount of local businesses with whom they do business.
– Sustainability: Compost from the stadium will go to Cedar Grove Compost and then will be used on a farm in Redmond. They will grow all the vegetables for the stadium. They will also have bees and will produce honey. If they have too much food in this process, the plan is to donate the excess to local food banks. The crops are being planted in the next 10-15 days.
→ 2016 Concession Specific Fan Survey
– Fans are looking for more variety.
– Fans are looking for speed of service.
– Fans are looking for better quality. Not as concerned with pricing.
Because like the team is dedicated to tickets and food for a family being available for under $100, there will be more family packs and combos to help achieve this goal.
They will take a cross-section of of Seattle and bring it in to CenturyLink. They will have a range of unique items that are just at CenturyLink Field and replica items that are available. They have been out to Eastern Washington to meet with many ranchers and growers. CenturyLink is already ranked #1 in the MLS in terms of food.
– Vegetarian & Vegan Options: Cameron Collins spoke on vegans options and the ways that the extenuating circumstances that made the vegan food fail at Safeco Field. He also asked that they make the ingredients of the food available at the various concessions and grab and go stands. They said that ingredients will be available, will have food maps, and will do a better job of educating employees on the various food options.
Martin Buckley spoke on the lack of quality in the vegetarian options and expressed hope that they would be improved moving forward. He also spoke on the hope that the options will be better advertised to fans. Taylor Graham said that the Sounders are dedicated to communicating those options.
– Beverage Selection: Stephanie Steiner made a request for better non-alcoholic beverage options and cited a the growing concern about the coming diabetes epidemic. Coke, as the beverage partner, could hopefully help move this forward.
– Local Alcohol Options: Alex Eagleton spoke on the hope that the Sounders could incorporate local wines in to the fold.
→ Other plans moving forward
– The chefs want to be visible during the matches. This can help them better connect with the consumers and let them receive comments and criticism directly.
– There will be some more concessions that are sold in the stands directly to the fans. For example, Papa Johns. The Sounders want to be sure this does not become so prevalent hat it gets in the way of the game, though.
– As they moved over from the previous concessionaire to creating their own, one of the keys was keeping a lot of the staff on board. They hope to continue to grow as an organization this way.
– For suites, they are streamlining the menu and making it better.
– They are investing in the back of the house and hope to continue to grow the equipment in order to increase quality.
– The food trucks will continue to be in the North Plaza during the matches
V. Constitution Articles
A) Article 6: Honorary Chairman (First Reading) – Stephanie
→ adds process, removal, and duties to 6.1 of the Honorary Chair. Makes it clear that if we don’t do this, it does not technically preclude us from doing anything listed here, but will make it easier for future councils. By adding (s) to the title, it means we can name multiple people, like the entire team. This puts it out there so future councils will know even when all of us are gone.
→ Why the executive committee should have the ability to remove the distinction, beyond the power of the council as a whole, is that it lets the EC act immediately, instead of waiting until the next meeting, when it could be necessary to act. Like they are convicted of drunk driving, commit some other egregious act.
→ “The executive shall make best efforts to survey the council for input prior to voting outside of a regular Business Meeting. ” added to 6.1.2 in case the executive committee needs to act (see above)
→ Other small tidying up done to ensure language follows rest of constitution
B) Article 8: Supporter Groups (Second Reading) – Stephanie
→ There was discussion on how the language in the article as previously amended lacks definition to make clear who an “affected” supporter group would be. In the end, however, because there is a group coming together to come to a solution for Bylaw 7, most feel it is likely irrelevant.
Vote to make amendments to Article 8 passes unanimously 11-0.
→ Language added to the Article to make clear the Front Office determines Supporter Group sections, not the Alliance Council.
Motion to amend Article 8 is seconded.
Amendments to Article 8 passes 11-0.
Vote to pass Article 8, as amended, passes 11-0.
VI. Alliance Council Initiatives for 2017
The Alliance has collected these initiatives from various sources, including email, social media, in person, and any other way fans choose to communicate with members of the Alliance Council. Undoubtedly, more initiatives will come in to existence throughout the year.
→ Memorializing “The Save” (Nick willing to take lead of Save Project, Cameron willing to hop on as backup.)
→ Hosting MLS Cup & What will happen to Season Ticket Holders (Stephanie is lead) | 2024-07-09T01:26:57.367551 | https://example.com/article/4401 |
C’est le jackpot pour les actionnaires des entreprises françaises. Au deuxième trimestre de cette année, ils ont franchi un nouveau seuil : 40,6 milliards de dollars de dividendes encaissés, soit une progression de 6,1 % par rapport à la même période un an plus tôt (hors variation de changes). Des chiffres extraits du «Global Dividend Index», réalisé par la firme d’investissement britannique Janus Henderson. De tous les Etats de l’Union européenne, l’Hexagone est celui où l’on distribue, en volume, le plus gros montant de dividendes. En Grande-Bretagne, les détenteurs d’actions n’ont eu à se mettre sous la dent «que» 32,5 milliards de dollars sur la même période et les Allemands 34,1 milliards. La cuvée française aurait pu être encore meilleure si Total, gros contributeur en la matière, n’avait décalé de quelques semaines le versement de son dividende. Qu’à cela ne tienne, le laboratoire pharmaceutique Sanofi et BNP Paribas figurent dans le classement mondial des six plus gros contributeurs aux gains de leurs actionnaires au cours du dernier trimestre.
«Bonne santé»
Dans le microcosme financier, ces nouveaux sommets ne surprennent guère. L’économiste en chef de la banque d’investissement américaine JP Morgan, Vincent Juvyns, y voit une sorte de retour à la norme : «La crise financière a eu lieu il y a tout juste dix ans, le secteur s’est réformé depuis et les entreprises ont renoué avec la rentabilité. C’est un signe de leur bonne santé.» Il est vrai que les dividendes versés en 2017 aux actionnaires sont calculés sur la base des bénéfices encaissés en 2016. Or ils ont été plantureux pour les grosses pointures de l’économie française regroupées dans le CAC 40. Leurs profits cumulés atteignent 75 milliards d’euros, soit 32 % de mieux qu’en 2015. Le trio de tête est composé de BNP Paribas (7,7 milliards), suivi d’Axa (5,8 milliards) et de Total (5,6 milliards). Dans l’industrie automobile, Renault a dégagé un bénéfice record de 3,5 milliards et celui de PSA a bondi de 79 %.
Avec de tels résultats, les actionnaires réclament leur part du gâteau et l’obtiennent ne serait-ce que pour éviter qu’ils ne soient volages. «L’épargnant européen n’est pas très friand des marchés financiers, il préfère le Livret A ou les placements en obligations», analyse Vincent Juvyns, de JP Morgan. Or des actionnaires qui vendent un titre entraînent le plus souvent sa chute sur les marchés financiers. Le meilleur moyen de le cajoler passe donc par les dividendes.
Toute la question est de fixer sa hauteur ? En 2009, un président de la République pas vraiment réputé pour être un fan de l’économie administrée proposait une règle, néanmoins non écrite : «Cela fait bien longtemps que je pense que la règle des trois tiers est une bonne règle. Sur 100 de bénéfices, il devrait y en avoir 33 % qui reviennent aux salariés, 33 % qui vont directement dans la poche de l’actionnaire et 33 % qui servent à être réinvestis dans l’entreprise», indiquait Nicolas Sarkoy au début de son mandat à une question du Monde.
Qu’en est-il véritablement ? D’après les études menées par l’économiste Eric Galiègue, sur 291 entreprises européennes cotées en Bourse (1), «la part de leurs bénéfices distribués aux actionnaires est comprise entre 42 % et 44 %». La règle des trois tiers n’est donc plus vraiment dans l’air du temps. En dix ans, ce ratio a progressé de 20 %.
Une telle inflation ne s’explique pas uniquement par la volonté de fidéliser les actionnaires. Le dividende attaché à chaque titre est certes ratifié par l’assemblée générale des actionnaires mais en réalité décidé dans un cénacle plus réduit : celui des conseils d’administration. Et ce sont les cadres du plus haut rang qui y siègent. «C’est ce que l’on appelle le capitalisme managérial, les dirigeants salariés ont un intérêt tout particulier à mettre en œuvre des politiques court-termistes de rémunération des actionnaires car leur intéressement est lié au cours de Bourse, poursuit Eric Galiègue. C’est un cas typique de conflit d’intérêts entre managers et actionnaires. Carlos Ghosn [PDG de Renault-Nissan, ndlr] s’est bien plus enrichi grâce aux actions qu’il a reçues qu’avec son salaire.» En clair, présidents et directeurs généraux, dont les bonus sont calculés en partie suivant l’évolution du cours de Bourse, ont tout intérêt à le faire grimper. Et pour y parvenir, rien de mieux qu’un confortable bénéfice par action décidé en conseil d’administration.
Libération a tenté d’interroger un certain nombre d’administrateurs indépendants sur le montant des dividendes proposés dans les conseils où ils siégeaient. Aucun d’entre eux n’a souhaité s’exprimer, pas plus que la représentante de l’Institut français des administrateurs. Un organisme dont l’une des missions est justement de travailler sur ces questions. Le président de l’Association française des grandes entreprises (Afep), Laurent Burelle, n’a pas non plus souhaité répondre à nos questions. Sous couvert de l’anonymat, un ex-PDG du CAC 40 indique qu’il avait fixé un plafond à son conseil d’administration. «Pas plus de 20 % des bénéfices», avait-il demandé. Manque de chance, la société qu’il dirigeait a vu ses résultats plonger après 2011.
Pierre Gattaz à Jouy-en-Josas le 29 août. Photo Albert Facelly pour Libération.
«les salariés ne prennent pas le même risque»
Du côté des entreprises de l’économie numérique, on assume des dividendes conséquents, mais à condition de les partager. Jean-Baptiste Danet, le président de Croissance plus, qui compte parmi ses membres des jeunes pousses des technologies appelées à grossir très vite, a fixé des règles : «Pour adhérer chez nous, il faut que l’entreprise ait une politique d’intéressement des salariés, ainsi qu’un système qui leur permette de devenir actionnaires. Au-delà de 30 % des bénéfices distribués, il faut veiller à ce qu’une part cohérente revienne aux salariés.» Fondateur et dirigeant de Voyageurs du monde, un tour-opérateur haut de gamme coté en Bourse, Jean-François Rial a «attendu dix ans avant de se verser des dividendes». Aujourd’hui, il distribue un tiers de ses bénéfices à ses équipes mais n’hésite pas à rappeler que «les salariés ne prennent pas le risque de l’actionnaire». Pour autant, ces politiques d’intéressement et de distribution d’actions à l’ensemble de l’entreprise demeurent à la discrétion des dirigeants. L’ensemble des salariés du secteur privé ont vu leur rémunération augmenter d’environ 2 % au cours des douze derniers mois, soit… trois fois moins que les dividendes.
(1) Base de données financières Factset. | 2023-10-31T01:26:57.367551 | https://example.com/article/2860 |
As America and Europe were in the clutches of eleventh-hour Obamania, a tragedy was unfolding in the heart of Africa: Up to 250,000 Congolese had been driven from their homes in the latest fighting that still threatens to engulf more African nations, driven into fetid refugee camps that supposedly offer U.N. peacekeeper protections from the carnage. On Halloween, U.N. Secretary-General Ban Ki-moon asked the U.N. Security Council president to consider sending peacekeeping reinforcements to the Democratic Republic of the Congo. The Security Council didn’t take it up until this week, approving a French resolution to send 2,785 more troops and 300 more police to the beef up the sorely understaffed force of about 17,000 already there.
Oh, and that valiant Ban begging for more troops? The United Nations Organization Mission in the Democratic Republic of the Congo, known by its French acronym MONUC, actually asked the U.N. for more troops — properly equipped and trained to react rapidly to the deteriorating situation there, at that — back on Oct. 3. And those freshly approved troops? No volunteers as of this writing. And pencil in another two months or so for their arrival, not likely by MONUC’s Dec. 31 mandate expiration. No rush or anything, U.N.
But France’s ambassador to the U.N., Jean-Maurice Ripert, got it right on the nose: While quantity is lacking, so is quality. “The rules of engagement, if they are strong enough, they are not being used strongly enough,” he said. No one knows better than the beleaguered Congolese.
As Americans went to the polls on Nov. 4, civilians in the town of Kiwanja, which had an ostensibly protective U.N. base nearby, answered knocks on their doors only to be met by gunfire. The slayings continued into the next day, as Americans and Europeans were snapping up their Obama-headlined newspapers. Human Rights Watch estimated at least 50 killings in Kiwanja; the Congolese Red Cross said it could be as high as 200. Reuters reporter Emmanuel Braun wrote, in part:
In Kiwanja, one distraught woman, crying hysterically, asked journalists to “come and see the five dead bodies in my house.” One was that of her husband. Two more bodies lay outside.
…Journalists asked the U.N. peacekeepers, who have a base nearby, why they had not intervened. They did not reply.
The U.N. later said that they only had 120 peacekeepers at the base, and were too “pinned down by fighting,” according to the Guardian, to save lives.
Bush has been vilified as an idiot and a war monger for years, and everybody knows that the world hates the USA for its imperialism. Why would Bush confirm the terrible opinion of him and his country by waging war in the noble Third World? I’m sure when the One is coronated on 20 January, the evil murderers in equatorial Africa will lay down their weapons and weep in shame at their behavior – problem solved. Meanwhile, “human shields”, deprived of the opportunity to protect a genocidal dictator in Iraq, could deploy to the Congo to protect the unfortunate UN “peacekeepers” from their rock-throwing attackers. All problems dissolve in the clear white light of hope and change emanating from the One!
It’s a shame the U.S. won’t withdraw from the U.N. and instead, use the money saved to hire groups such as Blackwater to protect people groups in danger of attack and death. That would be change you can believe in.
Would that the MSM (or even Fox) compare the sexual predacity of these “peacekeepers” with the minor abuses of Abu Ghraib! They are a mob disguised as a militia, sent to do the job of a gendarmarie and dignified by the name of an army. Anyone who doubts the honor of the American military tradition should be forced to confront this uncivilized, undisciplined, abominable and depraved abuse of the trust and power given for the defense of civilization but used for rape and pillage. Never, even at their lowest, have American (or Confederate) soldiers come near this grade of depravity.
The UN “peacekeeping” forces have been ineffectual in the DRC and elsewhere for two simple reasons as is shown by their operational history. Reason one: Most of the contingents are tactically inept, poorly trained, poorly equipped and badly officered. Reason two: The rules of engagement are overly restrictive and do not allow for any taking of the offensive initiative even by those few contingents both capable and willing to do so. The ROE problem has two parts, the UN rules and those imposed by the political leadership of the national governments.
There are other difficulties as well. Usually the political goal of the deployment is not well defined. There is no definition of success. Well meaning hand-wringers of the NGO community are allowed too much leeway to criticise the actions of the deployed forces. The old stand-by charges of “neo-colonialism” and “racism” are hurled quickly by the local government (if any) and the local wannabe governmental contenders. The ever-so-sensitive types at the UN listen too quickly and too well to complaints of this nature.
Then there are the minor difficulties such as no unity of command. No common doctrine. No interoperability of equipment. Poor communications. Language barriers. And, in the estimate of the snuffies, the guys who do the killing and dying, damn little reason to put one’s butt in the line of fire. Orders from above are not a good reason for getting killed particularly when one’s own country has no dog in the fight.
The above is just for openers, the observations of a military historian and one time grunt who actually spent some time up close and personal with UN peacekeeping ops.
You say that UN peacekeeping has long been a sick joke. I’ll go you one better. The UN has long been a sick joke ( since about 1945 ). We need to get out. Financing this kleptomaniac dictators’ club is a crime against humanity. The US needs to stop doing it.
I suggest that the neo New Left drop their linen, quit their grinnen, and pick up an AK47 on the the way to Darfur to stop the genocide now: American Express, never leave home with out it. Besides an icon of the left, the AK47 is plentiful in Africa, and it can be picked for a good price, unlike San Francisco where only Sean Penn and ilk can acquire a concealed weapons permit. Less talk and more walk from the likes of Michael Moore and Jane Fonda, saviors of the common folk, successful capitalists passing for communists, and perhaps the African genocides would end at the point of a gun. If only the rank and file Lear jet liberals of the Daily Kos would show up for the Darfur peace keeping rather than London Concerts against global warming and the genocide would stop. With that victory under their belts, then they could move on to the rumble in the jungle in the Congo. You go girls! After all, you can’t adopt all those babies and the Netroots have never stopped a civil war in progress.
Any question what’ll happen once the UN, the American Left, the EU, and the Arab League convince the Israeli government to capitulate (one can argue that acceptance of a Two State Solution was the first capitulation)? We’ll have the good people in UN Peace-Keeper blue protesting in good bureaucratic fashion that they are there to keep the peace and not stop a genocide.
History Geek – There is no definition of success. Well meaning hand-wringers of the NGO community are allowed too much leeway to criticise the actions of the deployed forces.
It goes well past criticism to anti-Western human rights lawyers and “progressive” Leftists like Kenneth Roth openly gunning for trials for “evil” Western soldiers committing “acts of violence” against “possibly innocent people” living with or alongside terrorists and insurgents.
Peacekeepers from Canada and The Netherlands have had full trials for chickenshit like beating Somali thieves raining their camps and stealing supplies..and it brought down the Dutch government.
Marines from Haditha fell victim to US media, the insurgents counter-propaganda campaign and were brought back to the USA in shackles to stand trial as Leftist Jews from Roth to George Soros to Barbara Boxer to the NYTimes presumed their guilt and demanded their blood, along with Nancy Pelosi’s pet, John Murtha.
And of course it isn’t just Leftist Jews gunning for Marines, you have the European side of the “human rights lawyers” community gunning for them as well as European peacekeepers and their true speciality – denouncing and suing Israel for “atrocities” while turning a blind eye to the “persecuted Muslims engaged in resistance activities.”
This has made many governments hyper-cautious or outright unwilling to send their forces into the face of lawsuits and media frenzy that could occur if they so much as poke their heads out of “safe compounds” in Afghanistan, Congo, Iraq and say “boo!” to the bad guys.
Meanwhile, the US has been beaten up enough by the self-righteous Left domestically and internationally, seen our troops demoralized by overly-restrictive ROE….that it may be in our national interest to let a big messy bloodbath just happen with the US not lifting a finger to stop it until the Left stops seeing human suffering as something that “traps” their opponents into involvement and which they then can denounce and exploit for political and legal actions.
A good place to start is the Congo. Let ‘em die….Then get all the Samantha Powers, Soros, Roth, lawyer types and the Euroweenies together and say “The Congo was a lesson.” Do not say “Something Must Be Done!! if your real objective is West and soldier-bashing.”
“If you want intervention, you must agree to support a “Good Samaritan Law” for our country’s armed forces” Because just as countries were compelled to do on learning motorists and medical people, after wide knowledge the slightest error in rescue could lead to a pack of lawyers suing their asses – found it safer to walk or drive by injured people and let them die, you don’t send soldiers into harms way and attempt to sue or criminalize them for any error.
History Geek recognizes the present system is unworkable. Be a Marine, risk your ass for vicious 3rd worlders you must rescue when it it is not in your Country’s vital interest, see yourself or other Marines clamped in irons for a mistake or non-mistake that NGOs and spokesmen for the enemy call “murder, atrocity!!”. Want to sign up for that?
****************The U.S. should withdraw from the U.N. and start building a league of democracies willing to defend it or assist it around the world.
We ran into problems with that with the “Coalition of the Willing”. No one really wants to be in a US-led “League of Lackeys” doing the declining Great Power’s bidding. We ran into other democracies having quite different ideas of what is in their best interest in Iraq, Afghanistan, and in past Wars like Vietnam.
And democracies are not agreed on what their “coalition of democratic Freedom-Lovers!!!” would do. Be a pact where we defend certain democracies against non-democracies, but not others? A pact where “democracries” with no vital interest at stake, defend unquestioningly “other democracies”?
500 hundred mercenaries could put down the insurrection and killing in a couple of days but its not politically correct. Highly trained, well armed merc’s would have no problem with this but it can’t happen. White troops and black murderers, no way; the people have to die. Same thing happened in Rwanda, 800,000 people died when the white merc were taken out by the UN. The poor blacks of Africa were better off under colonialism, they had laws, they had safety, they had jobs and dignity, now they have death by machete.
The UN is by far the largest and most damaging actor on the continent of africa. They give money to petty dictators who in turn use this money to raise armies and attack their neighbors or their own ppl… which in turn the UN uses as an excuse to exist… its nothing more then a giant scam thats being played on the world most of all the US tax payer. Sending money to africa or even aid of any form simply makes problems worse there. The only way to fix things is to start on one corner of the continent and take over one country at a time setting up a government that can run. The only short term fix for many of these minor wars and genocide are merc units dropped in with orders to kill everything that is acting stupid period.
UN peacekeepers have never in the history of africa stopped a war or genocide… however they have a long history of helping both start and/or standing by and doing nothing. If UN peacekeepers could fix africa’s problems they would have 20+ years ago. Africa is still a hellhole and UN peacekeepers are still there… funny how that is…
That’s globalism for you. Looks as though USA is destined to become a vassal of the UN.
“…[T]o please foreigners …, he will back measures that limit U.S. sovereignty. …
In the November 17 issue of the National Review (not available online to my knowledge), John Fonte of the Hudson Institute identifies four “transnational power grabs” that Obama is likely to push for They are: the Law of the Sea Treaty, the Rights of the Child Treaty, the Convention for the Elimination of All Forms of Discrimination against Women (CEDAW), and the International Criminal Court. Agreement by the U.S. to these arrangements would make us more popular with foreigners, but only at a cost to our national security, our right of self-governance, and our rights under the Constitution.”
(Power Line Blog: John Hinderaker, Scott Johnson, Paul Mirengoffhttp://www.powerlineblog.com, November 8, 2008 Posted by Paul at 10:16 PM)
Seems like we voted for it.
OK every one, I understand the sarcasm of referring to Obama as the “One” or as the “Messiah”. But these are serious issues, and for my part I find little humor in it all.
Reading the article, there is one obvious answer to the whole issue of the ineffectiveness of the UN forces. The UN needs more power to accomplish this task and that power will only come by superseding national sovereignty, which includes that of the US as well. Our original constitution was “the Articles of Confederation and Perpetual Union”, which failed because it failed to create a power that could rule over the states. In short some one needs to be in control. Now the question is should the UN be given such power to rule over all the nations of the World. No, No, & No! Be careful because this is where it is leading to.
Any effectual response to crises of those happening in the Congo will be led by a sovereign state and a leader that is more concerned about the lives of people than what “the world” thinks. Bush has been such a leader and it is a shame that he has been demonized. With hindsight you may be able to point out some mistakes, but it is easier to coach from the bleachers and comment after the game than to actually call the shots. Whether America or any other country should enter into such a conflict is for every Sovereign state to decide. And often the issues are much more complicated than they first appear. | 2023-11-28T01:26:57.367551 | https://example.com/article/5635 |
Q:
Saving data from a JSON response
I currently have a JSON response, fairly simple one. But I couldn't find a good guide or kicking off point for getting the JSON response and saving it within a model that I have e.g posts.
"Grab a JSON feed containing posts and save each one within the posts
table in rails"
Is there a simple way to do this with rails?
Any help is greatly appreciated.
A:
Not a lot to work with...but let's assume the json string is represented by the variable json_str.
parsed = JSON.parse(json_str)
The parsed string should now essentially just be key value pairs like any other hash. To get the value, simply use the key.
parsed["some_key"]
Will return the value. To make your post from this, you can take the values you need and pass them in one by one, like so:
Post.create(some_value: parsed["some_key"], # etc)
Or, if all of your keys happen to share names with your attributes, you can pass the params all at once by saying:
post = Post.new(parsed)
and then calling:
post.save
Let me know if you have trouble.
| 2024-01-12T01:26:57.367551 | https://example.com/article/8956 |
Above: Class of 2020 participates in an 8th grade Youth Frontiers Courage Retreat. The retreat is designed to build positive school communities and help young people to recognize the importance of respecting themselves and others.
The Clayton School Counseling program strives to meet the comprehensive needs of ALL students through individual, small group, and classroom counseling. Our focus is in three key areas: academic success, post-secondary planning, and socio-emotional heath. | 2024-06-04T01:26:57.367551 | https://example.com/article/6055 |
---
abstract: 'Expansion for the free energy functionals of the Potts spin glass models with 3, 4 and 5 states up to the fourth order in $\delta q_{\alpha \beta }$ around the replica symmetric solution (RS) is investigated using a special quadrupole-like representation. The temperature dependence of the 1RSB order parameters is obtained in the vicinity of the point $T=T_c$ where the RS solution becomes unstable. The crossover from continuous to jumpwise behavior with increasing of number of states is derived analytically. The comparison is made of the free energy expansion for the Potts spin glass with that for other models.'
address:
- '$^1$ Institute for High Pressure Physics, Russian Academy of Sciences, Troitsk 142190, Moscow Region, Russia'
- '$^2$ Institute for Computational Physics, University of Stuttgart, Pfaffenwaldring 27, D-70569 Stuttgart, Germany'
author:
- 'N.V. Gribova$^{1,2}$, T.I. Schelkacheva$^1$, E.E. Tareyeva$^1$'
title: |
Potts spin glasses with 3, 4 and 5 states near $T=T_c$:\
expanding around the replica symmetric solution
---
The spin glass corresponding to the random variant of the Potts model [@Wu] occured to be one of the first models where the scenario of the transition to the nonergodic state was discovered [@ElSh83; @GoEl85] to be quite different from the standard one, that is of the Sherrington–Kirkpatrick (SK) model [@sk]. Nevertheless, it remains to be in the focus of investigations till now and a number of questions still are not answered (see, e.g., the recent papers Ref.[@Janis; @Al] that are also dealing with the mean field Potts glasses with few number of states).
It is well known that in the case of SK model the so callled full replica symmetry breaking (FRSB) [@Par] takes place at the instability point $T_c$ of RS solution. All other models with two-particle interaction and with reflection symmetry behave in the same way [@Full]. The Potts glass models (with $p>2$) are usually considered to be classical examples of models without reflection symmetry: FRSB does not work at $T_c$ [@ElSh83; @GoEl85; @Gross; @NoSh93; @Ritort; @Gribova] and there is a region where the 1RSB solution is stable (see, however, Ref.[@Janis]). It is known (see, e.g., Ref.[@Ritort] and references therein) from the numerical solution, that if the number of states $p\leq4$, then the 1RSB solution appears continuously at $T_c$ – the bifurcation point of the RS solution. For $p>4$ the 1RSB solution appears jumpwise. The Potts model with $p=4$ presents a boundary between two types of phase transitions. In this case the order parameter $m$ defining the number of groups in 1RSB construction achieves its maximum value $m=1$ just at the point $T_c=T_{1RSB}$. It is worthwhile to follow in an analytic way, using the bifurcation theory, how the crossover from continuous to jumpwise behavior of the order parameter happens. This paper is devoted to this problem. To some extent we proceed in a same line as the authors of the paper Ref.[@A.Crisantione] where the SK model was considered.
It is worth to notice that the Potts symmetry leads to the cancellation of some terms in the free energy. This by-turn results in a trivial solution for the order parameters in the RS approximation. This is why the Potts models differ from other models without reflection symmetry, in particular, from the models similar to SK model, that is with the Hamiltonian $$H=-\frac{1}{2}\sum_{i\neq j}J_{ij} U_i U_j, \label{SK}$$ with quenched interactions $J_{ij}$ distributed with Gaussian probability, but without reflection symmetry: $$Tr\left[{U}^{(2k+1)}\right]\neq 0 \label{odd}$$ for some integer $k$. In this case the disorder smears out the phase transition: there is no trivial RS solution at high temperature (see, e.g. [@Luchinskaya; @Hydrogen]).
At the same time the Potts symmetry leads to the fact that the order parameter is one and the same for all of $p$ states, that makes it related to two–particle generalized SK–type models and allows to show that the dependence of the free energy on the order parameters in the vicinity of the bifurcation point $T_c$ is universal in both cases.
We use a special representation of Potts models which originates from the analogy between three–state Potts spin glass and isotropic quadrupole glass (see below). In our representation the number of integrations is less than in usual approach and so it is more convenient for actual calculations. Moreover, this representation demonstrates explicitly the appearence of cubic terms in the free energy expansion. These terms play a crucial role for the order-parameter behavior.
The Potts glass model is defined by the Hamiltonian $$\hat H=-\frac{p}{2}\sum_{i\neq j}J_{ij} {\delta_{\sigma_{i},\sigma_{j}}}.
\label{one1}$$ where $p$ is the number of states and variables $\sigma_i$, $\sigma_j$ can take the values $0,1,...,p-1$. Here $J_{ij}$ are random interactions distributed with Gaussian probability $$P(J_{ij})=\frac{\sqrt{N}}{\sqrt{2\pi}
\tilde{J}}\exp\left[-\frac{(J_{ij})^{2}N}{ 2\tilde{J}^{2}}\right],
\label{two1}$$ where the factor $N$ insures a sensible thermodynamic limit.
The 3-state Potts spin glass was described earlier in our papers (see, e.g., Ref.[@L; @Lu; @Gribova]) using the representation in terms of the operators of quadrupole momenta: $Q=\left(3{J_{z}}^{2}-2\right)$, $V=\sqrt{3}\left({J_{x}}^{2}-{J_{y}}^{2}\right)$, $\textbf{J}=1$, $J_{z}=1,0,-1$. The operators $Q$ and $V$ commute and in the representation where both of them are diagonal have the following form $$Q =
\left(
\begin{array}{ccc}
-2& 0& 0\\
\phantom{x}0& 1& 0\\
\phantom{x}0& 0& 1\\
\end{array}
\right)
%$$ V = (
[cccc]{} 0& 0& 0\
0& & 0\
0& 0& -
) $$ Throughout this paper we shall use the short-cut notations $Q_{k} =(-2,1,1)$ and $V_{k }=(0,\sqrt{3},-\sqrt{3})$. The equation $$\left[Q_{k
}Q_{l}+V_{k }V_{l }\right]+2=6\delta_{k ,l }.
\label{two13}$$ provides the equivalence of the Potts model with $p=3$ to the model with the Hamiltonian $$H=-\frac{1}{2}\sum_{i\neq j}\left[J_{ij}Q_{i}Q_{j}+G_{ij}{V}_{i}{V}_{j}\right] ,
\label{two834}$$ (with $J_{ij}\equiv G_{ij}$). The quenched interactions are distributed with Gaussian probability.
Let us note that this representation is very useful for actual calculations and allowed us to obtain, in particular, the number of metastable states at zero temperature [@Lu] and to determine the low–temperature boundary of the stability of the 1RSB solution [@Gribova].
Using the replica method we obtain the averaged free energy of the system in the form
$$\begin{aligned}
\label{two934}
F/NkT=\lim_{n \rightarrow
0}\frac{1}{n}\left[-2t^{2}n+\frac{t^{2}}{2}\sum_{\alpha
\neq\beta}\left(q^{\alpha \beta }\right)^{2} \right.\nonumber
\\
\left.-\ln\Tr\exp\left[\frac{t^{2}}{2}\sum_{\alpha \neq\beta}q^{\alpha
\beta }
\left({Q}^{\alpha}{Q}^{\beta}+{V}^{\alpha}{V}^{\beta}\right)\right]
\right],\end{aligned}$$
where $q^{\alpha \beta }$ is the glass order parameter, $t={\tilde{J}}/kT$. Here indices $\alpha$ and $\beta$ label replicas. The Potts symmetry permits to describe the glass state with one glass order parameter for both kinds of operators.
Using the standard procedure we perform the first stage of the replica symmetry breaking (1RSB) according to Parisi ($n$ replicas are divided into $n/m$ groups with $m$ replicas in each group) and obtain the free energy in the form (with $q^{\alpha \beta }= r_1$ if $\alpha $ and $\beta $ are from different groups and $q^{\alpha \beta }= r_1+v$ if $\alpha $ and $\beta $ belong to the same group).
$$\begin{aligned}
\label{two137}
F_{\mathrm{1RSB}}=-NkT\times \nonumber
\\
\biggl\{t^2\left(2+(1-m)\frac{v^{2}}{2}-2v^{2}\right)+ \nonumber
\\
\frac{1}{m}\ln\int dz^G\int ds^G \left[\Psi\right]^{m}\biggr\}.\end{aligned}$$
Here $$\begin{aligned}
\Psi=\exp{\left(-2{\theta}_{1}\right)}+\exp{\left({\theta}_{1}\right)}\left[\exp{\left({\theta}_{2}\right)}
+\exp{\left(-{\theta}_{2}\right)}\right],\\
{\theta}_{1}=zt\sqrt{v},{\theta}_{2}=st\sqrt{3v},
\\
\int\frac{dz}{\sqrt{2\pi}}(\ldots)\exp\left(-\frac{z^2}{2}\right)\equiv \int dz^G(\ldots).\end{aligned}$$
The Potts symmetry provides also the trivial solution $q_{RS}=0$ and $r_{1}=0$ (see e.g. Ref [@Ritort]). Let us note that in general case of the Hamiltonian (\[two834\]) with $J_{ij}\neq G_{ij}$ one of the glass order parameter can not be zero. This fact can be easily seen by analyzing the high–temperature expansions for the RS equations for the order parameters [@LuTa87]. Let us note that the Eq.(\[two137\]) can be obtained from the appropriate equation of Ref. [@Ritort] by the integration over suitable linear combinations of variables.
Expanding the free energy (\[two137\]) near $T_c$ up to $v^4$ and solving the equations that define the extremum conditions for the free energy one can obtain [@L]: $$v = 8\tau -84\tau ^2,\hspace{3mm} m = \frac{1}{2} - \frac{9}{2} \tau,
\hspace{3mm}\Delta F= 16\tau ^3,
\label{mvf3}$$ where $\Delta F=F_{1RSB}-F_{RS}$ and $ \tau = t-t_c$.
Let us consider now the case $p=4$. In this case one can use the following operators $Q'\sim 3{J_{z}}^{2}-J(J+1)$, $V'\sim J_{z}$ for $\textbf{J}=3/2$ and the third operator $P'$ defined as to be orthogonal to $Q'$ and $V'$. Namely, $2Q'_k= (1,-1,-1,1)$, $2\sqrt{5}V'_k=
(-3,1,-1,3)$, $2\sqrt{5}P'_k= (1,3,-3,-1).$
It is easy to see that the following equations hold: $$Q'^2=\frac{1}{4};~~
V'^2=\frac{1}{4}+\frac{2}{5}Q';~~P'^2=\frac{1}{4}-\frac{2}{5}Q',$$ $$V'P'=-\frac{3}{10}Q';~~Q'V'= -\frac{3}{10}P'+\frac{2}{5} Q';$$ $$Q'P'=\frac{2}{5}P'-\frac{3}{10}Q',$$ so that $$Q'_kQ'_l+V'_kV'_l+P'_kP'_l +\frac{1}{4} = \delta _{k,l}.
\label{4444prs}$$ There is no reflection symmetry and we have $$\Tr P'^2Q'=-\frac{2}{5};\hspace{2mm}\Tr V'^2Q'=\frac{2}{5};\hspace{2mm}\Tr P'V'Q'=-\frac{3}{10},
\label{kuby}$$ that leads to the appearance of cubic terms in the free energy expansion[^1].
Now, using the replica method one can write the averaged free energy for the Hamiltonian $$H=-\frac{1}{2}\sum_{i\neq
j}J_{ij}[Q'_{i}Q'_{j}+V'_{i}V'_{j}+P'_iP'_j] ,
\label{ham4}$$ in the form analogous to (\[two934\]) and perform the 1RSB. The extremum conditions for this 1RSB free energy have the form of three equations that look different only at the first glance. In fact they are the equations from Ref.[@Ritort], integrated over different variables and so have a common solution $v\equiv \langle \langle
Q'_\alpha Q'_\beta \rangle \rangle = \langle \langle V'_\alpha
V'_\beta \rangle \rangle = \langle \langle P'_\alpha P'_\beta
\rangle \rangle$. Expanding the 1RSB free energy in the neighborhood of $t_c$ up to $v^4$ we obtain (in terms of standard variables (as, e.g., in [@Ritort])): $$\begin{aligned}
\label{df4}
\Delta F/Nkt= -\frac{1}{8}(-1+m)t^2 v^2 \times \nonumber
\\
\times [-6+6t^2+4(m-1)t^4v+(3m^2-3m-7)t^6v^2].\end{aligned}$$ Here $t=t_c+\tau $, $m=m_0+\delta m$ and $\tau $, $v$ and $\delta m$ are small. The extremum conditions relative to $v$ and $m$ give the system of equations: $$\begin{aligned}
\label{min4}
-\frac{1}{2}(-1+m)t^2v\times \nonumber
\\
\times[-3 +3t^2+3(m-1)t^4v+(3m^2+3m-7)t^6v^2]=0;\nonumber
\\
\frac{1}{8}t^2v^2[6-6t^2-8(m-1)t^4v+(4+12m-9m^2)t^6v^2]=0.\end{aligned}$$ To determine the relative orders of the small values $\tau $, $v$ and $\delta m$, one has to use the so-called bifurcation equation (see, e.g. Ref.[@tren]) for the system (\[min4\]). One can also check directly that the equations (\[min4\]) become incompatible if $v$ and $\delta m$ are supposed to be series in integer powers of $\tau $. If $\delta
m=m_1 \sqrt{\tau }$ and $v=v_1\sqrt{\tau }$ (as it follows from the bifurcation equation) then we obtain $t_c=1$, $m_0=1$, and for $m_1$ and $v_1$ the following system of equations can be obtained from (\[min4\]): $$6+3m_1v_1-7v_1^2=0;$$ $$12+8m_1v_1-7v_1^2=0.$$ The system has two solutions that differ only by the sign. Taking into account that $$\Delta F/NkT=\frac{1}{2}m_1^2v_1^3\tau ^{5/2},$$ and that on the physical branch $m<1$, we choose $$v=\sqrt{12\tau /35}; ~~~m=1-\sqrt{21\tau /5}.$$ So, in the case of Potts spin glass with 4 states $m=1$ at the point of bifurcation $t_c$, that is $t_{1RSB}=t_c$.
Let us consider now the Potts glass with $p=5$ in an analogous way. We shall use the following diagonal operators: $P_k^{(1)}=(-4,1,1,1,1)$, $P_k^{(2)}=(0,-3,1,-1,3)$, $P_k^{(3)}=\sqrt{5}(0,1,-1,-1,1)$, $P_k^{(4)}=(0,1,3,-3,-1)$. The equivalence to the Potts model follows from the relation: $$P_k^{(1)}P_l^{(1)}+P_k^{(2)}P_l^{(2)}+P_k^{(3)}P_l^{(3)}+P_k^{(4)}P_l^{(4)}+4 =
20\delta _{k,l}.$$ The nonzero cubic terms are $${P^{(1)}}^3=-12;~~{P^{(2)}}^2P^{(1)}=4;~~{P^{(3)}}^2P^{(1)}=4;$$ $${P^{(4)}}^2P^{(1)}=4;~~
{P^{(2)}}^2P^{(3)}=16/\sqrt{5};$$ $${P^{(4)}}^2P^{(3)}=-16/\sqrt{5};~~ P^{(1)}P^{(3)}P^{(4)}=-12/\sqrt{5}$$
Now the 1RSB free energy expansion near $t_c$ has the following form (in standard variables): $$\begin{aligned}
\label{df5}
\Delta F/NkT= -\frac{1}{12}(-1+m)t^2 v^2 \times \nonumber
\\
\times [-12+12t^2+4(2m-1)t^4v+(6m^2+6m-35)t^6v^2].\end{aligned}$$ The extremum conditions are $$\begin{aligned}
\label{min5}
-\frac{1}{3}(-1+m)t^2v\times \nonumber
\\
\times[-6 +6t^2+3(2m-1)t^4v+(6m^2+6m-35)t^6v^2]=0, \nonumber
\\
\frac{1}{12}t^2v^2[12-12t^2-4(4m-3)t^4v+(41-18m^2)t^6v^2]=0.\end{aligned}$$
It is easy to obtain that now $v$ and $\delta m$ can be presented as series in integer degrees of $\tau $: $v=v_1\tau +v_2 \tau ^2$, $m=m_0+m_1\tau $. Here $v_1$, $v_2$, $m_0$ and $m_1$ satisfy the following equations: $$6-3v_1+4m_0v_1=0,$$ $$4-v_1+2m_0v_1=0$$ and $$6v_2-12m_1-92=0,$$ $$6v_2-16m_1-43=0,$$ so that finally we obtain $$t_c=1;\hspace{2mm}m=\frac{3}{2}+\frac{49}{4}\tau ;\hspace{2mm} v=-2\tau
+\frac{239}{6}\tau ^2$$ and the transition from RS to 1RSB can not take place at the bifurcation point.
Here we would like to make a remark about the 5-state model without any pretend to be rigorous. Basing on the fact that the numerically obtained [@Ritort] value of $t_{1RSB}$ for $m=1$ is very close to our $t_c$, one can hope to obtain the jump to the 1RSB solution already from the equation (\[df5\]) and to estimate qualitatively the characteristics in question. In fact, if we expand (\[df5\]) in the vicinity of $m=1$ and let its first and second derivatives relative to $v$ be equal to zero, we obtain $t_{1RSB}^2= 184/187$ and the jump $\Delta v = 561/8464$ (compare with the table in [@Ritort]).
Qualitative behavior of the solutions for $v$ near $t_c$ for $p=3,4,5$ is presented on the figure \[fig:scheme\].
![(Colour online) Schematic representation of the order parameter $v$ in the vicinity of the bifurcation point $t_c$ for $p=3,4,5$. Solid lines for $m\leq 1$, dotted lines for $m>1$. []{data-label="fig:scheme"}](potts1.eps){width="8"}
Now let us consider the 1RSB free energy expansion near the bifurcation point $t_c$ from a more general point of view. It is interesting that the form of the series for $\Delta F$ in small deviations $\delta q^{\alpha \beta}$ from $q_{RS}$ up to the third order is one and the same for different models and coincides with the one that can be written [@Hydrogen] for the random generalized Hamiltonian (\[SK\]) where $U$ is arbitrary diagonal operator. The fact is that in the most general case (including (\[SK\]) and (\[one1\])) $\Delta F$ depends on the following (the only nonzero) sums: $$\begin{aligned}
&\lim_{n \rightarrow 0}\frac{1}{n}
{\sum_{\alpha,\beta}}^{'}(\delta
q^{\alpha\beta})^{2} =-[r-(m-1)v]^{2}-m(1-m)v^{2},\label{11eq:JP}\\
&\lim_{n \rightarrow 0}\frac{1}{n}
{\sum_{\alpha,\beta,\delta}}^{'}\delta
q^{\alpha\beta}\delta
q^{\alpha\delta} =[r-(m-1)v]^{2},
\label{51eq:JP}\\
&\lim_{n \rightarrow 0}\frac{1}{n}
{\sum_{\alpha,\beta}}^{'}(\delta
q^{\alpha\beta})^{3}=-[r-(m-1)v]^{3}+\nonumber\\
&3m(m-1)[r-(m-1)v]v^{2}+m(m-1)(2m-1)v^{3}, \label{61eq:JP}\\
&\lim_{n \rightarrow
0}\frac{1}{n}{\sum_{\alpha,\beta,\gamma}}^{'}\delta
q^{\alpha\beta}\delta q^{\beta\gamma}\delta
q^{\gamma\alpha}=2[r-(m-1)v]^{3} - \nonumber\\
&3m(m-1)[r-(m-1)v]v^{2}-m^{2}(m-1)v^{3}, \label{711eq:JP}\\
&\lim_{n \rightarrow
0}\frac{1}{n}{\sum_{\alpha,\beta,\gamma}}^{'}(\delta
q^{\alpha\beta})^{2}\delta
q^{\alpha\gamma}=[r-(m-1)v]^{3}- \nonumber
\\
&[r-(m-1)v]v^{2}, \label{71eq:JP}\\
&\lim_{n \rightarrow
0}\frac{1}{n}{\sum_{\alpha,\beta,\gamma,\delta}}^{'}\delta
q^{\alpha\beta}\delta q^{\alpha\gamma}\delta q^{\alpha\delta}=\nonumber\\
&\lim_{n \rightarrow
0}\frac{1}{n}{\sum_{\alpha,\beta,\gamma,\delta}}^{'}\delta
q^{\alpha\beta}\delta q^{\alpha\gamma}\delta q^{\beta\delta}=
-[r-(m-1)v]^{3},\label{81eq:JP}\end{aligned}$$ with $r=r_{1}-q_{RS}$. The prime on the sum means that only the superscripts belonging to the same $\delta q$ are necessarily different in $\sum'$ .
So, the deviation $\Delta F$ of the free energy $F_{1RSB}$ from its replica symmetric part in the most general case is
$$\begin{aligned}
\label{1011frs}
\frac{\Delta F}{NkT}=\frac{t^2}{4}\left[1-t^{2}W\right]\left\{-\left[r-(m-1)v\right]^{2}-\right.
\nonumber\\
\left.v^{2}m(1-m)\right\}-\frac{t^{4}}{2}L\left[r-(m-1)v\right]^{2}- \nonumber\\
t^{6}\left\{C\left[r-(m-1)v\right]^{3}+\right. \nonumber
\\
D\left[r-(m-1)v\right]v^{2}m(m-1)-B_{3}v^{3}m^{2}(m-1)+\nonumber
\\
\left. B_{4}v^{3} m(m-1)(2m-1)\right\}+...\end{aligned}$$
where $t=t_c+\tau $, and parameters $W, L, C, D, B_3,
B_4$ are some combinations of operators averaged over the RS solution. For example, the coefficient $L$ enters $\Delta F$ with the sum (\[51eq:JP\]): $\frac{1}{n}
{\sum_{\alpha,\beta,\delta}}^{'}\delta q^{\alpha\beta}\delta
q^{\alpha\delta}$
Let us note that in the case of zero RS solutions for the order parameters the expansion does not contain the terms where some indices occur only once. In the case of reflection symmetry there is no terms where some indices occur odd number of times. If there is no reflection symmetry, using the extremum conditions for the free energy for Hamiltonian (\[one1\]) and taking into account the fact that $L|_{t=t_c}\neq {0}$, the bifurcation condition gives[@Hydrogen] $r-(m-1)v=0+o(\Delta t)^{2}$, i.e. the condition that there is no linear term for the glass order parameters. In fact, there is no other linear term because $$\left[1-t^{2}W\right]|_{t=t_c}=\lambda_{\rm RS repl}|_{t=t_c}=0
\label{40510prs}$$ at the bifurcation point.
In the case of the Potts spin glass model the situation is quite different, although the reflection symmetry is absent. Now $L=0$ because it is zero RS solution that bifurcates so that single indices mean that $L$ is composed of zero values of averaged $\Tr Q$,$\Tr V$.... Analogously the coefficients in front of (\[71eq:JP\]) and (\[81eq:JP\]) are also zero. So, we have $C=2B_{3}-B_{4}$ and $D=3(-B_{3}+B_{4})$. Here $B_{3}$ is the coefficient in front of (\[711eq:JP\]), and $B_{4}$ that in front of (\[61eq:JP\]). Since $L|_{t=t_{0}}= {0}$, then the condition $r-(m-1)v=0$ is not fulfilled. (If $p>2$ then $r_{1}=r=0$.[@Ritort])
In general case we obtain from the extremum conditions for the free energy (\[1011frs\]) the following equations for the order parameters: $$\begin{aligned}
& 2m(m-1)vZ\Delta t =
t_c^{6}m(m-1)v\left\{3\left[-B_{4}+\right.\right.\nonumber\\
&\left.\left. m(-B_{3} +2B_{4})\right]v+2D\left[r-(m-1)v\right]\right\} \label{30prs},\\
\nonumber \\
&(2m-1)v^{2}Z\Delta t =
t_c^{6}v^{2}\left\{(2m-1)\left[-B_{4}+\right.\right. \nonumber \\
&\left.m(-B_{3}+2B_{4})\right]v+m(m-1)(-B_{3}+2B_{4})v+\nonumber\\
&\left.D(2m-1)\left[r-(m-1)v\right]\right\}
\label{40prs},\end{aligned}$$ where $$\label{B123eq:JP}
Z = \frac{d}{dt}{\left[\frac{t^2}{4}\left[1-t^{2}W\right]
\right]}_{\mid_{t_c}}.$$
Hence we obtain from (\[30prs\])-(\[40prs\]): $$m(m-1)\left[mB_{3}-B_{4}\right]=0.
\label{401prs}$$
In fact, in the case of the Potts spin glass with $p=3$ for the Hamiltonian (\[two834\]) at $t=t_{c}$ we have: $$\begin{aligned}
W &= \frac{1}{2}\left[\left(TrQ^2/3\right)^{2}+
\left(TrV^2/3\right)^{2}\right]=4,\nonumber\\
B_{3}&=\frac{1}{6}\left[\left(TrQ^2/3\right)^{3}+
\left(TrV^2/3\right)^{3}\right]=\frac{8}{3},\nonumber\\
B_{4}&=\frac{1}{12}\left[\left(TrQ^3/3\right)^{2}+
3\left(TrQV^2/3\right)^{2}\right]=\frac{4}{3},\nonumber\\
D&=-4.
\label{0298prs}\end{aligned}$$
In this case in the neighborhood of the bifurcation point one can get from the equations (\[30prs\])-(\[40prs\]) that $v\sim \tau $, and from the equation (\[401prs\]) $m=\frac{1}{2}$.
Analogously, in the case of 4 states (for the Hamiltonian (\[ham4\]) we obtain: $D=0$, $B_{3} = B_{4} = \frac{1}{128}$ so that $m=1$. The equation (\[30prs\]) is the identity now. The r.h.s. of (\[40prs\]) becomes zero and one has to take into account higher order terms in the expansion of $F_{1RSB}$. This leads to $v\sim\sqrt{\tau }$. For $p=5$ we have $D=64, B_{3}=\frac{128}{3}, B_{4}=64$, so that $m=\frac{3}{2}$ and $v\sim -\tau $.
It is worth noticing that from the free–energy expansion (\[1011frs\]) it follows directly that the stability of the RS state is determined by the sign of $\lambda_{\rm RS repl}$, just as in the case of SK model [@Almeida]. To prove this fact let us consider small deviations from the RS solution: $q^{\alpha \beta}=q_{RS}+\delta q^{\alpha \beta}$ at arbitrary temperature and let us perform 1RSB. Using the notations introduced before, (that is $\delta q^{\alpha \beta }= r$ if $\alpha $ and $\beta $ are from different groups and $\delta q^{\alpha \beta }= r+v$ if $\alpha $ and $\beta $ belong to the same group) we can write $F_{1RSB}-F_{RS}$ up to the second order in $\delta q$ in the form (\[1011frs\]) with arbitrary $t$. Since $\left[1-t^{2}W\right]=\lambda_{\rm RS repl}$, it is easy to see that for $m<1$ and $L \geq {0}$ the RS state is stable ($\Delta F<0$) for the values of temperature such that $\lambda_{\rm RS repl}>0$. The similar result is valid for all subsequent stages of RSB because of the Parisi rule $$\lim_{n \rightarrow 0}\frac{1}{n}
{\sum_{\alpha,\beta}}^{'}(\delta
q^{\alpha\beta})^{2}<0$$
To conclude, we introduced the representation for the Potts glass model based on the operators of the quadrupole momenta. With the help of this representation the expansions for the free energy functionals of the Potts spin glass models with 3, 4 and 5 states up to the fourth order in $\delta q_{\alpha \beta }$ around the replica symmetric solution were obtained. The temperature dependence of the 1RSB order parameters in the vicinity of the point $T=T_c$ where the RS solution becomes unstable was derived. The crossover from continuous to jumpwise behavior with the growing of the number of states was traced analytically. The comparison was made of the free energy expansions for the Potts spin glass with that for other models and the similarity of such expansions was demonstrated.
Authors thank V.N. Ryzhov and N.M.Chtchelkatchev for helpful discussions and valuable comments.
This work was supported in part by the Russian Foundation for Basic Research (Grant No. 08-02-00781), and the Program of the Presidium of the Russian Academy of Sciences.
References {#references .unnumbered}
==========
[99]{}
Wu F 1982 [**]{} [**54**]{} 235 Elderfield D and Sherrington D 1983 [**]{} [**16**]{} L497; Elderfield D and Sherrington D 1983 [**]{} [**16**]{} L971 (1983); Elderfield D and Sherrington D 1983 [**]{} [**16**]{} L1169 Goldbart P and Elderfield D 1985 [**]{} [**18**]{} L229 Sherrington D and Kirkpatrick S 1975 [**]{} [**32**]{} 1972; Kirkpatrick S and Sherrington D 1978 [*B*]{} [**17**]{} 4384 Janis V and Klic A 2010 [*Preprint*]{} arXiv:1002.4576. Alvarez Banos B R et al 2010 [*J. Stat. Mech.*]{} P05002 ([*Preprint*]{} arXiv:1002.4288) Parisi G 1980 [**]{} [**13**]{} L115; Parisi G 1980 [**]{} [**13**]{} 1887 Schelkacheva T I, Tareyeva E E and Chtchelkatchev N M 2006 [*Physics Letters A*]{} [**358**]{} 222 Gross D J, Kanter I and Sompolinsky H 1985 [**]{} [**55**]{} 305 Nobre F D and Sherrington D 1993 [**]{} [**26**]{} 4539 De Santis E, Parisi G and Ritort F 1995 [**]{} [**28**]{} 3025 Gribova N V, Ryzhov V N and Tareyeva E E 2003 [*E*]{} [**68**]{} 067103 Crisanti A and De Dominicis C 2010 [**]{} [**43**]{} 055002 Lutchinskaia E A, Ryzhov V N and Tareyeva E E 1984 [**]{} [**17**]{} L665; Lutchinskaia E A, Ryzhov V N and Tareyeva E E 1986 [*Theor. Math. Phys.*]{} [**67**]{} 623; Lutchinskaia E A and Tareyeva E E 1995 [*B*]{} [**52**]{} 366 Schelkacheva T I, Tareyeva E E and Chtchelkatchev N M 2009 [*E*]{} [**79**]{} 021105; Schelkacheva T I, Tareyeva E E and Chtchelkatchev N M 2009 [*Theor. Math. Phys.*]{} [**160**]{} 1190 Lutchinskaia E A and Tareyeva E E 1991 [*Theor. Math. Phys.*]{} [**87**]{} 669; Lutchinskaia E A and Tareyeva E E 1992 [*Theor. Math. Phys.*]{} [**91**]{} 438 Lutchinskaia E A and Tareyeva E E 1992 [*Europhys. Lett.*]{} [**17**]{} 109 Lutchinskaia E A and Tareyeva E E 1987 [*Theor. Math. Phys.*]{} [**70**]{} 336 Vainberg M M and Trenogin V A 1974 [*Theory of branching of solutions of non-linear equations (Monographs and Textbooks on Pure and Applied Mathematics)*]{} (Leyden: Noordhoff International Publishing) Almeida J R L and Thouless D J 1978 [**]{} [**11**]{} 983
[^1]: It is worth to notice that our representation in terms of three matrices is not unique. For example, one can use the set: $A_k=(1,-1,-1,1)$, $B_k=(-1,1,-1,1)$, $C_k= (1,1,-1,-1).$ We use, however, the quadrupole–like operators in order to retain the physical meaning of generalized anisotropic models and to preserve the analogy between even–p and odd–p cases.
| 2023-11-07T01:26:57.367551 | https://example.com/article/3530 |
<html>
<head>
<style>
.reveal-if-active {
opacity: 0;
max-height: 0;
overflow: hidden;
font-size: 14px;
-webkit-transform: scale(0.8);
transform: scale(0.8);
-webkit-transition: 0.5s;
transition: 0.5s;
}
.reveal-if-active label {
margin: 0 0 3px 22px;
display: block;
font-size: smaller;
}
.reveal-if-active input[type=text] {
width: 300px;
}
input[id="choice-archive4"]:checked ~ .reveal-if-active {
opacity: 1;
max-height: 120px;
padding: 0px 0px;
-webkit-transform: scale(1);
transform: scale(1);
overflow: visible;
}
table {
margin: 14px auto;
opacity: 0;
}
table, th, td {
border-collapse: collapse;
}
th, td {
padding: 1px;
text-align: left;
font-family: "My Custom Font", Verdana, Tahoma;
font-size: 12px;
}
tr{
border-bottom: 1px solid #ccc;
border-top: 1px solid #ccc;
}
#title {
display: block;
text-align: center;
padding: 22px 0 0 0
}
.url{
display: block;
text-align: center;
}
#text_url{
width:333px;
font-size: 12.5px;
}
#select_label{
padding: 0px 270px 0 0;
text-align: center;
margin-bottom: 8px;
}
#choices{
text-align: center;
padding: 0px 20px 0px 0px;
margin-left: 133px;
}
#choices2{
text-align: left;
display: inline-block;
}
#perma_cc_api{
margin: -2px 93px 0px 21px;
}
#submitdiv{
text-align: center;
padding: 20px 242px 0 0;
margin: 0 0 0 38px;
}
input[type=submit] {
width: 5em;
height: 2em;
font-size: 12px;
background-color: gainsboro;
margin: 0px 0px 0px 13px;
}
#errors{
font-size: smaller;
color: brown;
padding: 6px 0px 3px 104px;
}
.img1{
width: 13px;
opacity: 0;
}
.img2{
width: 13px;
opacity: 0;
}
.img3{
width: 13px;
opacity: 0;
}
.img5{
width: 13px;
opacity: 0;
}
.img6{
width: 13px;
opacity: 0;
}
.img4{
width: 13px;
opacity: 0;
}
#apilink{
font-size: smaller;
padding-top: 39px;
}
</style>
</head>
<body>
<script src="https://ajax.googleapis.com/ajax/libs/jquery/1.7.1/jquery.min.js" type="text/javascript"></script>
<h3 id="title"> Preserve a web page in web archives </h3>
<div class="url">
<label for="text_url" id="label_url">URL</label>
<input type="text" id="text_url" required>
</div>
<div>
<p id="select_label">Select archives:</p>
<div id="choices">
<div id="choices2">
<input type="checkbox" id="choice-archive1" checked > Internet Archive <img src={{ url_for('static', filename = "ajax-loader.gif") }} class="img1" id="img1"> <br>
<input type="checkbox" id="choice-archive2" checked > Archive.is <img src={{ url_for('static', filename = "ajax-loader.gif") }} class="img2" id="img2"> <br>
<input type="checkbox" id="choice-archive3" > WebCite <img src={{ url_for('static', filename = "ajax-loader.gif") }} class="img3" id="img3"> <br>
<input type="checkbox" id="choice-archive5" > Archive.st <img src={{ url_for('static', filename = "ajax-loader.gif") }} class="img5" id="img5"> <br>
<input type="checkbox" id="choice-archive6" checked > Megalodon.jp <img src={{ url_for('static', filename = "ajax-loader.gif") }} class="img6" id="img6"> <br>
<input type="checkbox" id="choice-archive4" > Perma.cc <img src={{ url_for('static', filename = "ajax-loader.gif") }} class="img4" id="img4">
<div class="reveal-if-active">
<label for="perma_cc_api">Permaa.cc requires <a href="https://perma.cc/settings/tools" target="_blank"> an API Key </a></label>
<input type="text" id="perma_cc_api">
</div>
</div>
</div>
</div>
<div id="submitdiv">
<input type="submit" value="Submit" onClick="push_archive();">
<input type="submit" value="Reset" onClick="reset();">
<div id ="errors"></div>
</div>
<table id="results" width="600">
<thead>
<tr>
<th scope="col" width="130">Archive</th>
<th scope="col" width="450">Link to the archived page</th>
</tr>
</thead>
</table>
<div id="apilink"><a href="/api" target="_blank">Archive Now API</a></div>
<script type="text/javascript">
document.getElementById('perma_cc_api').value = localStorage.getItem("permaccapikey");
if (localStorage.getItem("check_archive_1") !== null){
if (localStorage.getItem("check_archive_1") == 'true'){
document.getElementById('choice-archive1').checked = true
}else{
document.getElementById('choice-archive1').checked = false
}
}
if (localStorage.getItem("check_archive_2") !== null){
if (localStorage.getItem("check_archive_2") == 'true'){
document.getElementById('choice-archive2').checked = true
}else{
document.getElementById('choice-archive2').checked = false
}
}
if (localStorage.getItem("check_archive_3") !== null){
if (localStorage.getItem("check_archive_3") == 'true'){
document.getElementById('choice-archive3').checked = true
}else{
document.getElementById('choice-archive3').checked = false
}
}
if (localStorage.getItem("check_archive_5") !== null){
if (localStorage.getItem("check_archive_5") == 'true'){
document.getElementById('choice-archive5').checked = true
}else{
document.getElementById('choice-archive5').checked = false
}
}
if (localStorage.getItem("check_archive_6") !== null){
if (localStorage.getItem("check_archive_6") == 'true'){
document.getElementById('choice-archive6').checked = true
}else{
document.getElementById('choice-archive6').checked = false
}
}
if (localStorage.getItem("check_archive_4") !== null){
if (localStorage.getItem("check_archive_4") == 'true'){
document.getElementById('choice-archive4').checked = true
}else{
document.getElementById('choice-archive4').checked = false
}
}
function reset() {
window.location.reload();
}
function push_archive() {
document.getElementById('errors').innerHTML="";
localStorage.setItem("check_archive_1", false);
localStorage.setItem("check_archive_2", false);
localStorage.setItem("check_archive_3", false);
localStorage.setItem("check_archive_5", false);
localStorage.setItem("check_archive_6", false);
localStorage.setItem("check_archive_4", false);
var arr = []
var table = document.getElementById('results');
for (var r = 1, n = table.rows.length; r < n; r++) {
if(table.rows[r].cells[0].innerHTML.indexOf("https://archive.org") !== -1){
arr.push("ia");
}
if(table.rows[r].cells[0].innerHTML.indexOf("https://archive.is") !== -1){
arr.push("is");
}
if(table.rows[r].cells[0].innerHTML.indexOf("https://archive.st") !== -1){
arr.push("st");
}
if(table.rows[r].cells[0].innerHTML.indexOf("https://megalodon.jp") !== -1){
arr.push("mg");
}
if(table.rows[r].cells[0].innerHTML.indexOf("https://www.webcitation.org") !== -1){
arr.push("wc");
}
if(table.rows[r].cells[0].innerHTML.indexOf("https://perma.cc") !== -1){
arr.push("cc");
}
}
function validateURL(textval) {
var urlregex = /^(https?|ftp):\/\/(((([a-z]|\d|-|\.|_|~|[\u00A0-\uD7FF\uF900-\uFDCF\uFDF0-\uFFEF])|(%[\da-f]{2})|[!\$&'\(\)\*\+,;=]|:)*@)?(((\d|[1-9]\d|1\d\d|2[0-4]\d|25[0-5])\.(\d|[1-9]\d|1\d\d|2[0-4]\d|25[0-5])\.(\d|[1-9]\d|1\d\d|2[0-4]\d|25[0-5])\.(\d|[1-9]\d|1\d\d|2[0-4]\d|25[0-5]))|((([a-z]|\d|[\u00A0-\uD7FF\uF900-\uFDCF\uFDF0-\uFFEF])|(([a-z]|\d|[\u00A0-\uD7FF\uF900-\uFDCF\uFDF0-\uFFEF])([a-z]|\d|-|\.|_|~|[\u00A0-\uD7FF\uF900-\uFDCF\uFDF0-\uFFEF])*([a-z]|\d|[\u00A0-\uD7FF\uF900-\uFDCF\uFDF0-\uFFEF])))\.)+(([a-z]|[\u00A0-\uD7FF\uF900-\uFDCF\uFDF0-\uFFEF])|(([a-z]|[\u00A0-\uD7FF\uF900-\uFDCF\uFDF0-\uFFEF])([a-z]|\d|-|\.|_|~|[\u00A0-\uD7FF\uF900-\uFDCF\uFDF0-\uFFEF])*([a-z]|[\u00A0-\uD7FF\uF900-\uFDCF\uFDF0-\uFFEF])))\.?)(:\d*)?)(\/((([a-z]|\d|-|\.|_|~|[\u00A0-\uD7FF\uF900-\uFDCF\uFDF0-\uFFEF])|(%[\da-f]{2})|[!\$&'\(\)\*\+,;=]|:|@)+(\/(([a-z]|\d|-|\.|_|~|[\u00A0-\uD7FF\uF900-\uFDCF\uFDF0-\uFFEF])|(%[\da-f]{2})|[!\$&'\(\)\*\+,;=]|:|@)*)*)?)?(\?((([a-z]|\d|-|\.|_|~|[\u00A0-\uD7FF\uF900-\uFDCF\uFDF0-\uFFEF])|(%[\da-f]{2})|[!\$&'\(\)\*\+,;=]|:|@)|[\uE000-\uF8FF]|\/|\?)*)?(\#((([a-z]|\d|-|\.|_|~|[\u00A0-\uD7FF\uF900-\uFDCF\uFDF0-\uFFEF])|(%[\da-f]{2})|[!\$&'\(\)\*\+,;=]|:|@)|\/|\?)*)?$/i;
return urlregex.test(textval);
}
if (validateURL(document.getElementById('text_url').value) == false){
document.getElementById('text_url').focus();
document.getElementById('errors').innerHTML="*Enter a correct URL*";
return;
}
if (document.getElementById('choice-archive4').checked == true){ // perma.cc
if(document.getElementById('perma_cc_api').value.trim() == ""){
document.getElementById('perma_cc_api').focus();
document.getElementById('errors').innerHTML="*Enter your Perma.cc API Key*";
return;
}
}
var selected_archives = 0;
if (document.getElementById('choice-archive1').checked == true){
selected_archives = selected_archives + 1;
if(arr.indexOf("ia") == -1){
document.getElementById('img1').style.opacity = 1
$.ajax({
type: "GET",
url: "ia/"+document.getElementById('text_url').value,
success: function(json) {
if (validateURL(json['results'][0]) == true){
var table=document.getElementById("results");
var row=table.insertRow(-1);
var cell1=row.insertCell(0);
var cell2=row.insertCell(1);
cell1.innerHTML='<a href="https://archive.org" target="_blank"> Internet Archive </a>'
cell2.innerHTML='<a href="'+json['results'][0]+'" target="_blank"> '+json['results'][0]+' </a>'
document.getElementById('results').style.opacity = 1
document.getElementById('img1').style.opacity = 0
}
},
complete: function(){
document.getElementById('img1').style.opacity = 0
}
});
}
localStorage.setItem("check_archive_1", true);
}
if (document.getElementById('choice-archive2').checked == true){
selected_archives = selected_archives + 1;
if(arr.indexOf("is") == -1){
document.getElementById('img2').style.opacity = 1
$.ajax({
type: "GET",
url: "is/"+document.getElementById('text_url').value,
success: function(json) {
if (validateURL(json['results'][0]) == true){
var table=document.getElementById("results");
var row=table.insertRow(-1);
var cell1=row.insertCell(0);
var cell2=row.insertCell(1);
cell1.innerHTML='<a href="https://archive.is" target="_blank"> Archive.is </a>'
cell2.innerHTML='<a href="'+json['results'][0]+'" target="_blank"> '+json['results'][0]+' </a>'
document.getElementById('results').style.opacity = 1
document.getElementById('img2').style.opacity = 0
}
},
complete: function(){
document.getElementById('img2').style.opacity = 0
}
});
}
localStorage.setItem("check_archive_2", true);
}
if (document.getElementById('choice-archive3').checked == true){
selected_archives = selected_archives + 1;
if(arr.indexOf("wc") == -1){
document.getElementById('img3').style.opacity = 1
$.ajax({
type: "GET",
url: "wc/"+document.getElementById('text_url').value,
success: function(json) {
if (validateURL(json['results'][0]) == true){
var table=document.getElementById("results");
var row=table.insertRow(-1);
var cell1=row.insertCell(0);
var cell2=row.insertCell(1);
cell1.innerHTML='<a href="https://www.webcitation.org" target="_blank"> WebCite </a>'
cell2.innerHTML='<a href="'+json['results'][0]+'" target="_blank"> '+json['results'][0]+' </a>'
document.getElementById('results').style.opacity = 1
document.getElementById('img3').style.opacity = 0
}
},
complete: function(){
document.getElementById('img3').style.opacity = 0
}
});
}
localStorage.setItem("check_archive_3", true);
}
if (document.getElementById('choice-archive5').checked == true){
selected_archives = selected_archives + 1;
if(arr.indexOf("st") == -1){
document.getElementById('img5').style.opacity = 1
$.ajax({
type: "GET",
url: "st/"+document.getElementById('text_url').value,
success: function(json) {
if (validateURL(json['results'][0]) == true){
var table=document.getElementById("results");
var row=table.insertRow(-1);
var cell1=row.insertCell(0);
var cell2=row.insertCell(1);
cell1.innerHTML='<a href="https://archive.st/" target="_blank"> Archive.st </a>'
cell2.innerHTML='<a href="'+json['results'][0]+'" target="_blank"> '+json['results'][0]+' </a>'
document.getElementById('results').style.opacity = 1
document.getElementById('img5').style.opacity = 0
}
},
complete: function(){
document.getElementById('img5').style.opacity = 0
}
});
}
localStorage.setItem("check_archive_5", true);
}
if (document.getElementById('choice-archive6').checked == true){
selected_archives = selected_archives + 1;
if(arr.indexOf("mg") == -1){
document.getElementById('img6').style.opacity = 1
$.ajax({
type: "GET",
url: "mg/"+document.getElementById('text_url').value,
success: function(json) {
if (validateURL(json['results'][0]) == true){
var table=document.getElementById("results");
var row=table.insertRow(-1);
var cell1=row.insertCell(0);
var cell2=row.insertCell(1);
cell1.innerHTML='<a href="https://megalodon.jp" target="_blank"> Megalodon.jp </a>'
cell2.innerHTML='<a href="'+json['results'][0]+'" target="_blank"> '+json['results'][0]+' </a>'
document.getElementById('results').style.opacity = 1
document.getElementById('img6').style.opacity = 0
}
},
complete: function(){
document.getElementById('img6').style.opacity = 0
}
});
}
localStorage.setItem("check_archive_6", true);
}
if (document.getElementById('choice-archive4').checked == true){
selected_archives = selected_archives + 1;
if(arr.indexOf("cc") == -1){
document.getElementById('img4').style.opacity = 1
$.ajax({
type: "GET",
url: "cc/"+document.getElementById('text_url').value+'?cc_api_key='+document.getElementById('perma_cc_api').value,
success: function(json) {
if (validateURL(json['results'][0]) == true){
var table=document.getElementById("results");
var row=table.insertRow(-1);
var cell1=row.insertCell(0);
var cell2=row.insertCell(1);
cell1.innerHTML='<a href="https://perma.cc" target="_blank"> Perma.cc </a>'
cell2.innerHTML='<a href="'+json['results'][0]+'" target="_blank"> '+json['results'][0]+' </a>'
document.getElementById('results').style.opacity = 1
document.getElementById('img4').style.opacity = 0
}
},
complete: function(){
document.getElementById('img4').style.opacity = 0
}
});
}
localStorage.setItem("permaccapikey", document.getElementById('perma_cc_api').value);
localStorage.setItem("check_archive_4", true);
}
if (selected_archives == 0){
document.getElementById('errors').innerHTML="*Select at least one archive*";
return;
}
}
</script>
</body>
</html>
| 2024-04-02T01:26:57.367551 | https://example.com/article/8787 |
Burglars in prison paid £20 a week to quiz you about your valuables: Convicts asking unsuspecting families if they want to save money on house insurance
Inmates read script which sees them ask for names and postcodes
Those at HMP Oakwood and Drake Hall ask about possessions value
Prisoners say they are ringing up on behalf of a market research firm
Inmates at both jails do not reveal they are calling from behind bars
Prisoners are earning £20 a week phoning householders and quizzing them about their valuables.
Burglars and other criminals are asking unsuspecting families if they would like to save money on their home insurance.
The inmates get paid to read from a script which includes asking potential customers their names and postcodes.
'Contained': A G4S spokesman said that the incident at HMP Oakwood, near Wolverhampton, had been contained
They also inquire about the total value of their possessions – including details of any worth large sums.
The scheme, which was launched two months ago, has been introduced at HMP Oakwood, near Wolverhampton, and Drake Hall women’s prison in Eccleshall, Staffordshire.
The prisoners say they are ringing up on behalf of a market research company and do not reveal they are calling from behind bars.
Last night a horrified source said: ‘This may have put hundreds of homes’ security at risk.’
In March 2009, Drake Hall (left) was re-designated from a semi-open women's jail to a closed prison. Labour MP Keith Vaz (right) said: 'This is not work prisoners should be involved in'
A Ministry of Justice spokesman confirmed the controversial scheme was operating.
Unsuspecting families are being asked by burglars and other criminals if they would like to save money on their home insurance (file picture)
He said: ‘We do not want prisoners sitting idle in their cells when they should be working towards their rehabilitation.
‘We prepare offenders for work inside prison so they can get a job after release – this reduces the chances that they will reoffend in the future, meaning lower crime and fewer victims.
‘All prisoners working in call centres are risk-assessed and stringent security measures are in place, with calls supervised and recorded.’
The spokesman said the full address details were not made available to the prisoners and they did not take a pen into the call room so they could not write any details down.
He also stressed that the computers the prisoners use do not have access to the internet. The inmates wear telephone headsets as they sit at the specially prepared jail call centres. They phone homeowners using a script.
A source said: ‘When they get through, they are told to ask, “Would you like to save some money?” Most customers say yes.
‘They then ask to confirm names and postcode, enough to identify where they live, and if they have valuable items.’
Last night Peter Cuthbertson, of the Centre for Crime Prevention, said: ‘Trusting criminals with people’s financial details is incredibly naive.
HOW DOES THE SYSTEM WORK?
Prisoners walk into a prepared call centre inside the jail and are given a headseat. They sit down at an 'isolated' computer which has no internet access - and they are not allowed to bring in any writing implements. They use a script to call homeowners and ask them to confirm their names and postcodes - but are not given the person's full address. They also inquire about the total value of their possessions and ask homeowners if they would like to save money on their home insurance.
'Burglars will know who to target when they are released.’
Labour MP Keith Vaz said: ‘This is not work prisoners should be involved in.’
In March 2009, Drake Hall was re-designated from a semi-open women’s jail to a closed prison.
Last month, HMP Oakwood was one of two privately-run prisons the Ministry of Justice had ‘serious concerns’ over. The jail, run by G4S, was given the lowest performance rating of one.
| 2024-02-12T01:26:57.367551 | https://example.com/article/5387 |
These Flower head pins are very popular as you can use them to do the following: Sewing, Pinweavering and Quilting. You can also iron over the Flat Flower- shaped head with caution. Color- NickelSize/ Type - 54mmCard- 60pcs in each pack | 2024-01-03T01:26:57.367551 | https://example.com/article/9331 |
Dural penetration by interspinous process segmental spine instrumentation: case report.
A case of pseudo-meningocele complicating posterior spine fusion with interspinous process segmental spinal instrumentation is presented. The L rod rotated postoperatively 90 degrees and penetrated the dura. The presenting symptoms were pain accompanied by fixed lumbar and hip flexion contractures that improved dramatically after removal of the penetrating part. The unusual course as well as the treatment and recommendation for prevention are discussed. | 2024-02-09T01:26:57.367551 | https://example.com/article/1231 |
Summary: To get the reading into stopping people from bringing up their past.
Down through the years we’ve discovered that many people try to keep you the way that you were. They would say that “A leopard cannot change it spots” and “You can’t teach an old dog new tricks.” And they believe that there is a slight chance for us changing. It reminds me a of my favorite comic strip, Tumble Weeds. There is a little character in the strip by the name of Limpin’ Lizard who is not known for his intellect. In one instance Limpin’ Lizard was in a communications workshop and the facilitator was explaining the Tom-Tom as the new way to communicate. At the end of the lesson he asked were there any questions and Limpin’ Lizard raised his hand and asked. Why did they decide to call it a Tom-Tom? Why didn’t they call it a George-George or a Ralph-Ralph? Limpin’ Lizard, I tell you was not known for his intellect and they felt that he would never change because he was what he was. In another comic episode, Limpin’ Lizard spotted a young lady that he took a fancy to ( to the young people that means that he liked her). So old Limp went to her house, knocked on the door and said Bon Suey Madamazelly, that’s French for “howdy Mam.” And he must’ve impressed the young lady so she responds “O’ Limpin’ Lizard, how suave, have you ever been abroad?” now this is when Limpin Lizard murdered the moment he said, “Shucks no, once a fella always a fella.” I tell you they had reason to believe that Limpin’ Lizard would never change.
But that is not true with those of us who were once sinners. The bible lets us know that “while we were yet sinners, Christ died for us.” In other word Christ took our worst case scenario to the cross with Him, then presented us before the Father as faultless.
In the first ten verses of Ephesians 2, Paul has discussed the salvation of sinners in general, but now he turns to the work of Christ for Gentiles in particular. Most of the converts in the Ephesians church were Gentiles, and they knew that much of God’s program in the Old Testament involved the Jews. For centuries, the “circumcision” (Jews) had looked down on the “un circumcision” (Gentiles) with an attitude that God had never intended them to display. The fact that a Jew had received the physical mark of the covenant was no proof he was a man of faith. Only those who trusted Christ have received a spiritual circumcision made without hands as found in Colossians 2:11 “In whom you are circumcised with a circumcision made without hands.”
So, let’s peek into our past that our now can be appreciated.
First of all Paul tells us to . . .
1. NOT FORGET THAT WE WERE ONCE THE OUTSIDERS. (v.11)
“Don’t forget that you Gentiles used to be outsiders. You were called “Uncircumcised heathens” by the Jews, who were proud of their circumcision, even though it affected only their bodies and not their hearts.”
-N. L. T.-
The King James’ text says, Wherefore remember. The design of this evidently is to excite a sense of gratitude in the bosoms of the Ephesians for that mercy which had called them from the errors and sins of their former lives to the privileges as Christians. It is a good thing for Christians to "remember" what they were. No faculty of the mind can be better employed to produce humility, repentance, gratitude, and love, than the memory. If we would think back to the way we were, it would cause us to thank the Lord for saving us. It is well to recall the recollection of our former sins; to dwell upon our hardness of heart, our alienation from God, and our unbelief. But most importantly we should remember how sinful we really were, remember it just long enough to feel it in our hearts. Then we can truly appreciate what the Lord has done in our lives. | 2024-04-21T01:26:57.367551 | https://example.com/article/2887 |
Monday, October 13, 2008
"Bi-gender" people express their masculine and feminine sides of their personality. Someone who is "bi-gendered" will express facets of both genders and/or changes their role depending on a situation and context.
Bi-gender is actually the natural condition of being a bit of both sexes, I am bi-gendered as I have a condition called Klinefelter's Syndrom, I have XXY chromosomes instead of either XX (female) or XY (male), it is the third scientifically recognized sex, the forth is Hermaphrodite which basically doesn't exist, the odds of a full Herm is like I think it was 1 in 4.7 billion or something and the chances of an actual Herm surviving is slim to noon due to several reasons.
Some one can't just label them selves as bi-gender as it's a natural condition which has really only come to light in the past decade or so.
A good movie based on the condition would be 1 in 2000 which I believe came out in either 2003 or 2005, can't recall for sure.
BTW I'm also trans to a mostly female side considering in my case my brain developed as female and that is for the most part how I consider my self, so mise well look the part right? :) | 2023-09-16T01:26:57.367551 | https://example.com/article/7041 |
Seeking help with VA
July 17, 2013Letters to the Editor
In response to “Mistreated by VA” published July 6: The similarities are very much the same as my husband’s have been for the past 23 years.
My husband’s situation is a long, complicated and unresolved mess. I’ll give a short version: a back injury in training and three surgeries over a period of several years. Without divulging too much of my husband’s personal information, just recently he has contacted Sen. Pat Toomey’s office and has received a response. He has tried in the past contacting members of Congress to no avail.
We are praying that the right person will get hold of his information and take an interest to get his mess fixed.
I am sure letter writer Kevin Prokop and many other servicemen and women are tired of hearing “contact your congressman.” My husband and I are also tired of hearing this. The news does not encourage us either, with the reports of backlog, paperwork, cutbacks etc., in the Veterans Affairs Department.
Still, I still strongly urge Prokop and other service members to contact their congressman. Maybe just the right person will show interest and notice a big problem — and get it resolved. | 2023-12-12T01:26:57.367551 | https://example.com/article/5426 |
---
id: 5900f4761000cf542c50ff88
challengeType: 5
title: 'Problem 265: Binary Circles'
videoUrl: ''
localeTitle: 问题265：二进制圆圈
---
## Description
<section id="description"> 2N二进制数字可以放在一个圆圈中，这样所有N位顺时针子序列都是不同的。 <p>对于N = 3，两个这样的圆形布置是可能的，忽略旋转： </p><p>对于第一种布置，顺时针顺序的3位子序列是：000,001,010,101,011,111,110和100。 </p><p>通过将以全零的子序列开始的二进制数字连接为最高有效位并顺时针进行，可以将每个循环排列编码为数字。因此，N = 3的两种排列表示为23和29：00010111 2 = 23 00011101 2 = 29 </p><p>调用S（N）唯一数值表示的总和，我们可以看到S（3）= 23 + 29 = 52。 </p><p>找到S（5）。 </p></section>
## Instructions
<section id="instructions">
</section>
## Tests
<section id='tests'>
```yml
tests:
- text: <code>euler265()</code>应该返回209110240768。
testString: assert.strictEqual(euler265(), 209110240768);
```
</section>
## Challenge Seed
<section id='challengeSeed'>
<div id='js-seed'>
```js
function euler265() {
// Good luck!
return true;
}
euler265();
```
</div>
</section>
## Solution
<section id='solution'>
```js
// solution required
```
/section>
| 2023-08-30T01:26:57.367551 | https://example.com/article/4185 |
Q:
Sonata + Fos_user - How to display only entities related to the user?
I have users who are venue managers. I want them to be able to manage their places and events that are happening in these places.
I created fos_user_user and there I built relations to places:
<entity name="Application\Sonata\UserBundle\Entity\User" table="fos_user_user">
<id name="id" column="id" type="integer">
<generator strategy="AUTO" />
</id>
<many-to-many field="places" target-entity="EchoBundle\Entity\Place">
<join-table name="users_places">
<join-columns>
<join-column name="user_id" referenced-column-name="id" />
</join-columns>
<inverse-join-columns>
<join-column name="place_id" referenced-column-name="id" />
</inverse-join-columns>
</join-table>
</many-to-many>
</entity>
So now, I can manage users and add places that they manage. It works fine.
Questions:
How can I filter so once they log in they only see their own places?
How can I allow them to only add events to their own places? Currently when you add an event you have a full list of places to select from.
How can I filter all events so that they only see events related to places they manage?
I looked at "CUSTOMIZING THE QUERY USED TO GENERATE THE LIST" in the Sonata documentation but don't know how to use it. I tried to add security queries found in answers on StackOverflow from 4 years ago but it didn't work.
A:
In your Admin class you can override createQuery (you should check and fix example below to meet your app model) ;)
This solution will cover question 1 and 3.
public function createQuery($context = 'list')
{
$query = parent::createQuery($context);
$aliases = $query->getRootAliases();
$query
->leftJoin($aliases[0] . '.users_places', 'users_places')
->andWhere($query->expr()->eq('users_places.user_id', ':user') )
->setParameter('user', $this->getConfigurationPool()->getContainer()->get('security.token_storage')->getToken()->getUser());
return $query;
}
Question 2:
If you are using sonata formMapper and configureFormFields method, you can pass Custom Query Builder in field definition.
$formMapper
->add('events', 'sonata_type_model', [
'label' => 'Events',
'placeholder' => 'Select ...',
'required' => true,
'query' => $blQueryBuilder,
]);
| 2024-03-14T01:26:57.367551 | https://example.com/article/7533 |
[Preparation and identification of immune serum against recombinant fusion protein of rhoptry 2 and major surface protein 1 from Toxoplasma gondii].
To prepare and identify immune serum against the recombinant fusion protein of rhoptry 2 (ROP2) and major surface protein 1(P30) from Toxoplasma gondii. The constructed recombinant plasmid of pET28b/ROP2-P30 was transformed to a bacterium BL21-Codon Plus (DE3)-RIL strain and was expressed under IPTG induction. Cells were lysed by multiple rounds of sonication to obtain supernatant and inclusion body respectively. The inclusion body was washed with 2 mol/L and 4 mol/L urea to remove the nonspecific protein. The washed products dissolved in 8 mol/L urea were received by SDS-PAGE. Two rabbits were immunized with the fusion protein rROP2-P30 and sera from the rabbits were collected. Immune diffusion test, indirect ELISA and Western-blot were used to detect antibody titer and specificity of the immune serum against rROP2-P30. Immune diffusion test demonstrated that specific immune serum were obtained. Indirect ELISA confirmed that the antibody titer in the serum reached 1:12 800 and the rROP2-P30 was recognized by specific IgG in this serum by Western-blot analysis. Specific immune serum against the recombinant fusion protein rROP2-P30 has been prepared. | 2024-07-02T01:26:57.367551 | https://example.com/article/7082 |
Perfect for use in on all types of fork lifts. Keeps everything in plain sight when you need it and out of the way when you don't. Attach a Device holder for your device to the face plate and you are all set to go.
Click the "View Options" button to select the Mount that meets your requirements | 2023-10-02T01:26:57.367551 | https://example.com/article/6899 |
Rose on the lawless country with the king as head of state
Chatwadee Rose Amornpat is a Thai activist living abroad who has been accused and charged with lese majeste. She has sent the following article to PPT and we reproduce it as received, with a couple of links added:
With all the turmoil which is being waged in Thailand now, it is interesting to observe that one so-called “revered” institution, namely, the monarch and/or the monarchy, has not come out to stop the chaos and the daily arrests of unarmed and peaceful democracy activists and alleged lese majeste violators.
Thai royalists and the royal household often surreptitiously inform the local and foreign media that the king has no political power to do anything, but a quick glance at the current constitution reveals the opposite results.
Though the junta chief. Gen. Prayuth Chan-Ocha stated last May 22, 2014 after he successfully seized power from a democratically-elected government, that the constitution was then abrogated except “all the constitutional articles relating to the monarch and lese majeste law.” That is to say, the laws concerning the power of the king, his welfare and his protection through the barbaric and unjust lese majeste law would be left intact and enforceable.
It is very odd indeed. To me, now we do not have just one dictator but two dictators in the same country!
This is symbiotic relationships between the monarchy and the military, which have been going on for the past six decades, while the poor people of Thailand continue to suffer and their quality of life worsen. One can still see many poor children of Thailand selling garlands to drivers of cars on the busy and smoggy streets in Bangkok every morning. Such poor children should have been in school, not selling garlands or flowers to help their family! It pains me to see such a sight.
Symbiotic relationships, according to “thefreedictionary.com,” are a special type of interaction between species. Sometimes beneficial, sometimes harmful, these relationships are essential to many organisms and ecosystems, and they provide a balance that can only be achieved by working together.
Indeed I compare the Thai monarchy and military to low-class animals which are the lowest of the lows as they have done nothing good for the Thai people. The monarchy has always stayed intact while the general changes every 5 or 10 years during the past 19 coup d’ tates. Because of this fact, one can only see that with all the troubles happening in Thailand, we can only blame it at the top, the monarchy. This is the main character which never changes.
Here’s a look at the current laws relating to the monarchy which Gen. Prayuth left intact, though he abrogated the constitution when he seized power from the former PM Yingluck Sinawatra.
Section 3 of the Thai constitution states:
The sovereign power belongs to the Thai people. The King as Head of State shall exercise such power through the National Assembly, the Council of Ministers and the Courts in accordance with the provisions of this Constitution.
The above Section is not known to Westerners or even most Thais. It is like saying the car belongs to the people but only the king can drive the car. Or, the people own the gun but only the
king can pull the trigger. In both cases, the people have to do the maintenance and upkeep of the car and gun.
Section 8 of the Thai constitution states:
The King shall be enthroned in a position of revered worship and shall not be violated. No person shall expose the King to any sort of accusation or action.
This section means the Thai King is like God and no one can sue the King even if he commits robbery, blatant lies, mayhem or murders. Article 112 of the Thai Criminal Code states:
No one can criticize the king and any member of his family, even if such criticisms are based on the truth.
This is called “lese majeste” and it carries a jail term of 3 to 15 years for each offense.
Section 10 states:
The King holds the position of Supreme Commander of the Thai Armed Forces.
Section 11 is:
The King has the prerogative to create titles and confer decorations.
Last Friday, June 19, 2015, Thai police arrested 14 students who had been protesting against the ruling junta, in defiance of a ban on public gatherings. These are young university students who are brave and full of democratic spirit. They just want nothing except the rights to express themselves freely on issues effecting their lives and future. They are now confined to a filthy and crowded Thai jail in Bangkok.
I urge the leaders of the civilized world and all the human rights organizations to put pressures on the Thai junta to release these students unconditionally.
The students took part in peaceful rallies calling for an end to military rule under the National Council for Peace and Order (NCPO). The army commander-in-chief, Gen. Udomdej Seetabutr, publicly accused the 14 student activists of being backed by anti-government groups and claimed their actions could lead to disturbances and violence.
Additionally, Gen. Udomdej Seetabutr, who was hand-picked by the top royals to head the Army’s top brass, has indicated that a charge of lese majeste may be leveled on them, because these students may have gotten supports from anti-monarchy elements as well.
It is against the international norm that the Thai Army is designed to protect only the monarchy as opposed to protecting the country!
I can’t help but to draw the attention to another group of students who were arrested and charged with lese majeste law violations in 2014. These brave young people are university students with bright future but cut short by the military and the monarchy. Two were arrested and sentenced in jail while the rest of them had to leave their university study and are now in hiding or flee the country.
Patiwat Saraiyaem, 23, and Pornthip Munkong, 26, involved in producing a play called “The Wolf Bride” about a fictional monarch and his adviser. It was performed at the prestigeous Thammasat University in 2013 to mark the anniversary of a successful 1973 anti-dictatorship uprising led by students.
Their bail requests were repeatedly turned down by a Bangkok court. Both had pleaded guilty, a common practice in lese majeste cases in December 2014. Because failure to do so only means a 100% guilty verdict and a long jail term.
In announcing the verdict, a Bangkok Criminal Court judge said the play contained content that insulted and defamed the monarchy and was shown in front of a large number of spectators.
Keep in mind all Thai judges are approved and appointed by the king. Thus there is no chance of acquittal in any lese majeste case. Now all the lese majeste cases are being handled by the military court.
Now not even people who are suffering from mental illness are spared from lese majeste charge.
A man was sentenced to more than three years in jail last week under lese majeste law, a controversial royal defamation law, despite having a history of mental illness. Tanet Nonthakot, 45, from northeastern Phetchabun province, is the second person in the last few months suffering from mental health condition to be convicted under this barbaric law.
Not even a very friendly and mild-mannered editor of the prestigious on-line newspaper, Thai E-News, Somsak Pakdeedej, 36, who was sentenced to 5 years in prison for allowing an article deemed lese majeste to be published three years ago.
Since the coup on May 22, 2014, in addition to the hundreds and hundreds of intellectuals and democracy activists who are now serving long jail term, scores of free spirited academics have fled the country and are now living in exile in neighboring Laos and Cambodia. Thanks to the Internet, they are now waging a daily and weekly war in cyberspace against the military junta as well as the monarchy who is alleged to be the mastermind of all the messes including the 10 coups in Thailand.
Three distinguished individuals with whom I greatly admire and who have sacrificed all their life for equality, justice and democracy for Thailand:
Suda Rangkupan, in her 40’s, a Fullbright scholar and holder of a doctorate degree and university professor, chose to flee the country when she received a call from the military junta to report to their central command for the so-called “attitude adjustment.”
Dr. Rangkupan is a very courageous woman who has cared for the poor much more than herself. She has a long history of fighting for justice and equality. She is a champion for the poor. A woman with her impressive credentials, she could have easily enjoyed her life as a university professor, just as the rest of the complacent academics in universities throughout Thailand now.
Because of her democracy activism, after 13 years as a professor at Chulalornkorn University, she was forced to resign or fired by her department chief. Soon after that, a call from the junta, she informed me that she chose to flee the country and continues with her noble struggle for justice. She had been at the forefront of Thailand’s democracy movements. She is now waging a cyber war against the military junta and the monarchy. Her Facebook page (https://www.facebook.com/sudarang ) receives thousands of views from her supporters, Thai and foreigners alike, from all over the world each day. Here’s Dr. Suda Rangkupan’s impressive resume.
Surachai Dangwathananusorn, aka, Surachai Sae Dang, in his mid 70’s. Surachai is considered a legendary democracy activist since the era of student uprising at Thammasart university in 1973 during the era of Gen. Thanom Kitikachon. He spent a total of 22 years in Thai jail for the alleged “encouraging uprisings against the military regimes.” He, too, chose to flee the country, because the military junta filed a lese majeste charge on him after the coup by Gen. Prayuth Chan-Ocha. He is considered the Nelson Mandela of Thailand. His weekly broadcast in YouTube is heard by his supporters all over the world.
Phisanu Phomsorn, aks “Anti” who was charged with lese majeste law violation for giving several speeches at Red Shirt’s political events. He, too, fled the country and is now living in exile at the border around Vietnam and Cambodia. His weekly program on Thai royalty and politics has been very popular among Thai audience in Northeast Thailand.
It is now becoming clearer and clearer that Thailand is being raped and governed by 2 types of thieves-in-uniforms which share a symbiotic relationship.
Under the so-called “Article 44” which gives unlimited power to the junta, they can just about do anything they so please from search anyone house or body without a warrant or jailing anyone on any minor charges. Now many lower ranking soldiers are behaving like hooligans extorting money from street vendors and retailers in up-country and big cities in open daylight with impunity.
The first type of thieves-in-uniform is the monarchy with the king dubbed as head of state who often wears decorative pins and trappings and occasionally wears uniforms similar to characters in ancient, Ramayana play, with head gears ancient hat for religious ceremonies. This type is only concerned about their stability and their vast wealth under the control of his investment arms, “Crown Property Bureau.” They tend to prolong their continued status qua and privileges and entitlement for generations to come.
The second thieves-in-uniform is the military, the generals, who benefit from their collusion with the monarchy for decades. Each top general has benefited from the yearly military budget and the allocation for purchases of arms. Each year, the budget gets increased by 10-20 percents even though Thailand has no wars or conflicts with her neighbors.
The top brass stands to benefit millions and millions of dollars or bahts in terms of commission.
This is a known fact among officials in the Thai Armed Forces.
Unless Thai people unite and demand the reorganization of the two institutions from the group floor and up, the chance of realizing a true democracy is probably next to nothing.
Last but not least, I hope that world leaders from the United States, the UK, Australia, New Zealand, Japan, France, Germany and all the civilized nations which have commercial contacts or educational exchanges with Thailand continue to put pressures and bring about the human rights issue and the need to abolish their despicable lese majeste law once and for all. | 2024-04-21T01:26:57.367551 | https://example.com/article/5772 |
title: Chapter 6
ID: 10
type: folder
compile: 2
setGoal: 1000
| 2023-11-26T01:26:57.367551 | https://example.com/article/1821 |
#!/usr/bin/env bash
# A script to run ocamlmerlin within the container. This exists because Emacs
# needs an executable name, and doesn't allow you to say `run-in-docker
# ocamlmerlin`.
script=$(basename "$0")
DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )"
"$DIR/run-in-docker" esy "$script" "${@}" <&0
| 2023-10-18T01:26:57.367551 | https://example.com/article/5901 |
The Medical Scientist Training Program at Baylor College of Medicine goal matches that of the NIH: to prepare its graduates to function independently in both basic research and clinical investigation. The program, while emphasizing continuity between clinical and basic science curricula, is designed for students who will become future leaders in academic medicine and biomedical research. The national and international applicant pool hails from throughout the United States and abroad with only 5-10% of applicants being selected to matriculate. Thus, our highly competitive program enables a selection from a truly exceptional pool of students. [unreadable] [unreadable] The program currently has 85 students enrolled with 39 in medical school and 46 in graduate school. In this academic year, 24 will be partially or fully supported by the NIH MSTP training grant, and institutional and private endowment funds are used to support other students. A new curriculum was implemented in 1995, effectively allowing students to complete medical school in three years. Thirteen graduate programs are available at Baylor College of Medicine including a joint program in bioengineering at Rice University and interdepartmental programs at Baylor, Rice University and the University of Houston. [unreadable] [unreadable] James R. Lupski, MD, PhD, FAAP, FACMG, FAAAS, is the MSTP Director. He is supported by three Co-Directors: Jeffrey M. Rosen, PhD, Martin M. Matzuk, MD, PhD, and Sharon Plon, MD, PhD. Three of the four program directors received their training through NIH MSTPs (Dr. Lupski at NYU, Dr. Matzuk at Washington University, and [unreadable] Dr. Plon at Harvard). (Full CVs are in Appendix 1) In addition to ensuring that the program meets the needs of all students, each director is assigned students as a mentor throughout their training, participates in various extracurricular educational and social activities, and works closely with students and faculty during the interview/selection process. [unreadable] [unreadable] Located in the heart of the Texas Medical Center, the world's largest medical center, Baylor College of Medicine offers outstanding training facilities in both clinical and laboratory settings. Graduates are accepted into highly competitive residency programs and many receive academic appointments at leading health care institutions across the United States. [unreadable] [unreadable] | 2023-09-19T01:26:57.367551 | https://example.com/article/7691 |
Q:
How to load a referenced assembly within another referenced assembly?
I have a scenario where my client is a WPF App let's say W , There are 3 assemblies say A, B,C which are not to be exposed to W , hence the 3 assemblies are referenced in a concrete factory , say F which creates the required instance and provides to W.
I am using reflection in F to generate the 3 objects.
Below is the code
return (IClass)Activator.CreateInstance(Assembly.Load("A").GetType("A.AClass"),
BindingFlags.NonPublic | BindingFlags.Instance, null, null, null);
The code works if I reference A in W , but doesn't work (System.IO.FileNotFound Exception) when A is referenced in F (I cross checked with Assembly.GetExecutingAssembly().GetReferencedAssemblies() and A is not loaded).
I read that an Assembly may not be loaded if it is not required. But how does my WPF client then load the assembly. If this is the expected behavior , I would go with finding the current Directory path and appending A.dll to it.
Any thoughts around ?
A:
I suspect the files are not getting copied to the AppDomain.CurrentDomain.BaseDirectory when the solution is built. Without a reference you would need to manually ensure the required DLLs are being copied to the right location.
See BasconSah comment above for a suggestion.
| 2023-11-07T01:26:57.367551 | https://example.com/article/2891 |
4 Steps To Help Family Members Who Are Addicts
Dealing with a friend or family member who is struggling with addiction can be a painful, difficult experience. If you are concerned about the effects of an addiction on someone in your life, read through this article for tips and information that will be helpful to you as you try to help them overcome this struggle. Though these steps may not always be easy to take, they are essential for someone who needs support as they try to change their addicted lifestyle.
Step 1: Gain a Better Understanding of Addiction
The mistake that many people make when trying to help an addict is not taking the time to learn about addiction and its effects on the body. Many people believe that addiction stems from a lack of willpower. This reinforces the mistaken belief that if a person wants to quit badly enough, they will be able to. Unfortunately, true addiction is much more complicated than a lack of motivation.
Addiction is actually a chronic brain disorder that hinders a person's ability to control their needs and desires for whatever they have become addicted to. Doctors have used brain imaging to determine that substance abuse can result in serious alterations to the brain areas that affect learning, memory, behavior control and decision making. This is what leads to the compulsive and destructive actions associated with addiction. The good news is that, although addicts will always have to deal with the possibility of relapse, this brain disorder is treatable.
It's also helpful to keep in mind that many factors may have come into play when the individual first began his or her decent into addiction. Studies have shown that genetic factors can play a role in addiction. Additionally, issues with childhood traumas, social environment and mental health can also contribute to addiction.
Step 2: Confront Your Loved One
Perhaps the most difficult step is confronting addicts about their addiction. However, this step is absolutely critical for laying the groundwork for a successful recovery. First and foremost, choose an appropriate time to talk to the person dealing with the addiction. Those dealing with addicts who are substance abusers should talk to them when they are sober. When confronting the addict, emphasize your support and steer clear of any language that will make the addict feel judged or guilty. Remain positive yet firm; let them know that though you won't do anything to enable their addiction, you will be there to support them throughout their recovery.
Some people are not as willing to accept or acknowledge their problem. If you feel this may be an issue, consider having at least a few friends or family members there for support. You may even want to ask an addiction therapist or counselor to attend the meeting as well. This is commonly referred to as an intervention. For some individuals, it may take a few confrontations of this nature before they agree to overcoming their addiction.
Step 3: Talk to a Doctor
If you haven't done so already, this is a good time to reach out to a physician. Addicts should, at some point, talk to a doctor about the depth of their addiction as well as their goals and expectations for recovery. A doctor will be able to provide valuable information about the physical effects of withdrawal that people might experience as they pull away from their addiction. For example, addicts going through withdrawal often develop serious side effects like nausea, vomiting, confusion or shakiness. Make sure that anyone who will be helping with the withdrawal process is well aware of what symptoms are normal and when to seek medical attention.
Step 4: Utilize Addiction Recovery Resources
There are several resources for addicts that can be especially helpful during their recovery. Here are some of the key options available:
Support groups: Joining a support group can be a great motivator and aid for addicts going through recovery.
Rehabilitation centers: Addicts who may need constant supervision and assistance throughout their withdrawal may want to check themselves into a rehab center for addiction. The Substance Abuse and Mental Health Services Administration provides a helpful tool for finding substance abuse treatment centers all over the U.S. at this website: http://findtreatment.samhsa.gov/TreatmentLocator/faces/quickSearch.jspx.
Tips to Keep in Mind
Now that you know more about addiction and how to help your loved one, it's important to have these tips in mind as you prepare to help him or her overcome addiction:
Be wary of lies and deceit. Addicts may lie or manipulate people to satisfy their addictions. Remember that addicts feel whatever it is they are addicted to is critical to their survival, so even if they have to hurt family and friends, they may lie in order to obtain it. Encourage your loved one to show his or her commitment to quit through actions and not words.
Use financial support wisely. It can be dangerous to give an addict cash or checks. Use your financial support for medical bills, rehab treatments, therapies, etc. Be sure your checks or cash go directly to the facilities providing this support.
Set realistic goals. Don't expect an addict to recover overnight, or even in a matter of weeks. It can take months or years to achieve a full recovery, and even more work after that to prevent relapse. Setting realistic goals will make the process easier for both you and the person facing addiction.
Be positive in your support. Criticism and judgment won't be helpful for an addict, especially when coming from a trusted friend or family member. Keep your encouragement and advice positive, and reinforce your love and support through the process of recovery. | 2024-07-15T01:26:57.367551 | https://example.com/article/5298 |
View Geocache Log
Found as one of many as we cached the Bear Bait trail today, myself along with Zuma, K-Spud and Sweetlife, signing as HIKSZ. We had perfect jacket weather all day, beginning with overcast and ending with glowing golden light filtering through the leaves at sunset. We quickly settled into a rhythm, each taking their turn jumping out for the find while the rest tried to spot the cache before the car was even stopped and often succeeding. Backroads caching is my favorite, and I enjoyed the views out the window as we passed from hardwoods to pines, pines to marsh and back again, very rarely running into other vehicles, and then parked hunters for the most part. We even had a chance to take a break and hunt a couple waterfall caches, breaking away at our sixty-sixth find on the trail for a couple hours of hiking to a couple remote ones. TFTC WabenoTrail, and for all the caches. Be cool Honeybunny!
This is the original cache type consisting, at a bare minimum, a container and a log book. Normally you'll find a tupperware container, ammo box, or bucket filled with goodies, or smaller container ("micro cache") too small to contain items except for a log book. The coordinates listed on the traditional cache page is the exact location for the cache. | 2024-03-21T01:26:57.367551 | https://example.com/article/9477 |
In vitro and in vivo effects of rifampin on Staphylococcus epidermidis graft infections.
Rifampin, bound in high concentrations to prosthetic grafts, has been proposed for the treatment of vascular graft infections. The optimum antibiotic concentration and duration of treatment for infected grafts is not known. This study compared the in vitro and in vivo efficacy of varying concentrations of rifampin against three different strains of slime producing Staphylococcus epidermidis (RP62A, KC2, and KB1) bound to the knitted Dacron at high and low concentrations at 10(4) and 10(7) CFU/cm2 of prosthetic. Time kill experiments were performed at 4, 18, and 42 hr, in which each Dacron bound bacterial strain was exposed in vitro to 4X, 64X, 100X, and 1,000X minimum inhibitory concentration (MIC) of rifampin. In vitro, the Dacron bound laboratory strain RP62A was implanted subcutaneously in the backs of male Swiss-Webster mice and exposed to 4X, 100X, and 1,000X the MIC of rifampin for similar time periods. In addition, systemic vancomycin (10 mg/kg) was assessed for synergy and prevention of rifampin resistance. In vitro, all concentrations of rifampin showed near total killing (< 1 log) of all bacterial strains at low initial concentrations (10(4) CFU/cm2) but not high (10(7) CFU/cm2) to 42 hr. Importantly, resistance was shown to develop in all three strains of S. epidermidis with high initial bacterial biofilm concentrations. In vivo, rifampin concentrations between 4X MIC and 100X MIC achieved a balance between optimal killing and prevention of resistance. Systemic vancomycin slightly improved bacterial clearance but did not alter the development of rifampin resistance at high local concentrations. Caution is advised with the use of antibiotic bonded grafts because resistance may develop even with the addition of systemic antibiotics. | 2024-02-19T01:26:57.367551 | https://example.com/article/1675 |
Q:
google geo coding with json and php
My php file:
<?php
if (isset($_POST['query'])) {
$query = $_POST['query'];
$url='http://maps.googleapis.com/maps/api/geocode/json?address=';
$callback="&callback=?";
$sensor='&sensor=false';
$result= $url.$query.$sensor.$callback;
curl_init ($result);
}
?>
My html file:
<html>
<head>
<script type="text/javascript" src="http://ajax.googleapis.com/ajax/libs/jquery/1.4.2/jquery.min.js"></script>
<script type="text/javascript">
$(document).ready(function(){
// when the user clicks the button
$("button").click(function(){
$.getJSON("geo.php",function(json){
$('#results').append('<p>Latitude : ' + json.results[9].geometry.location.lat+ '</p>');
$('#results').append('<p>Longitude: ' + json.results[9].geometry.location.lng+ '</p>');
});
// get the json file
});
});
</script>
</head>
<body>
<input type="text" id="query" /><button>Get Coordinates</button>
<div id="results"></div>
</body>
</html>
I get this error when i used firbug:
json is null
$('#results').append('<p>...s[9].geometry.location.lat+ '</p>');
Google's json response:
{
"status": "OK",
"results": [ {
"types": [ "street_address" ],
"formatted_address": "1600 Amphitheatre Pkwy, Mountain View, CA 94043, USA",
"address_components": [ {
"long_name": "1600",
"short_name": "1600",
"types": [ "street_number" ]
}, {
"long_name": "Amphitheatre Pkwy",
"short_name": "Amphitheatre Pkwy",
"types": [ "route" ]
}, {
"long_name": "Mountain View",
"short_name": "Mountain View",
"types": [ "locality", "political" ]
}, {
"long_name": "California",
"short_name": "CA",
"types": [ "administrative_area_level_1", "political" ]
}, {
"long_name": "United States",
"short_name": "US",
"types": [ "country", "political" ]
}, {
"long_name": "94043",
"short_name": "94043",
"types": [ "postal_code" ]
} ],
"geometry": {
"location": {
"lat": 37.4219720,
"lng": -122.0841430
},
"location_type": "ROOFTOP",
"viewport": {
"southwest": {
"lat": 37.4188244,
"lng": -122.0872906
},
"northeast": {
"lat": 37.4251196,
"lng": -122.0809954
}
}
}
} ]
}
EDITED:
<?php
if (isset($_POST['query'])) {
$query = $_POST['query'];
$url='http://maps.googleapis.com/maps/api/geocode/json?address=';
$callback="&callback=?";
$sensor='&sensor=false';
$result= $url.$query.$sensor.$callback;
$resp = file_get_contents($result);
header('Cache-Control: no-cache, must-revalidate');
header('Expires: Mon, 26 Jul 1997 05:00:00 GMT');
header('Content-type: application/json');
echo $resp;
}
A:
Are you sure you are capturing the result?
I'm assuming this:
Your PHP code is called via Ajax from the html.
So basically you are constructing the URL for the query and using CURL to fetch the json result.
It seems you are not actually capturing or returning the results to the ajax call
You might want to do this:
$process = curl_init($url);
//init curl connection
curl_setopt($process, CURLOPT_HEADER, 0);
curl_setopt($process, CURLOPT_POST, 1);
curl_setopt($process, CURLOPT_RETURNTRANSFER,1);
curl_setopt($process,CURLOPT_CONNECTTIMEOUT,1);
$resp = curl_exec($process);
//your content
curl_close($process);
//$resp contains your response
header('Cache-Control: no-cache, must-revalidate');
header('Expires: Mon, 26 Jul 1997 05:00:00 GMT');
header('Content-type: application/json');
echo $resp;
| 2024-03-29T01:26:57.367551 | https://example.com/article/5713 |
To view the educator resources, you must log in as an educator. Register now >>
As an educator, you have access to the resources listed below. These resources were developed by scientists and educators and can be used to test students on the video material. Note that students do not have access to these questions and answers.
iBioEducation Short Clips are 2-10 min videos highlighting a specific topic, experiment, or principle. The short clips in this section have accompanying questions of varying levels.
Assessments for iBioEducation Discovery Talks
The Famous Discoveries section include a video by a scientist describing their personal experience of the events leading to this discovery, a research paper, as well as a series of questions and suggested answers.iBioEducation Discovery Talks >>
Course Materials for Cell Biology Flipped Course
In the Cell Biology Flipped Course section, you will have access to all the course materials developed in collaboration with Dr. Jon Scholey at UC Davis, and used for a flipped course in Spring 2013.See Cell Biology Flipped Course >>
Other Educator Resources
iBioEducation Science Standards and Core Concepts Index
Examples that align the iBiology talks with Vision and Change, AP Biology, and Next Generation Science Standards Core Concepts and Competencies | 2023-09-26T01:26:57.367551 | https://example.com/article/4861 |
Dear Colleagues:
There are some advocates who claim to be abolitionists but who support welfare reform or who support violence.
In this Commentary, I explain why welfare reform and violence cannot fit into the abolitionist approach.
Gary L. Francione
©2009 Gary L. Francione | 2024-03-15T01:26:57.367551 | https://example.com/article/1621 |
Community Grade
(1 User)
Your Grade
There are worse ideas for a film than “Let’s provide a vehicle for Melissa Leo’s default pained expression.” Leo, the Oscar-winning actor from The Fighter, Frozen River, and Treme is this generation’s Dianne Wiest, a nuanced performer whose perpetually sad eyes and careworn face suggest endless past traumas, along with a capacity for suffering not yet fulfilled. Francine, the first fiction film from husband-and-wife documentary team Brian M. Cassidy and Melanie Shatzky (The Patron Saints), makes excellent use of her Wiest-like ability to generate sympathy with a look, even when her characters are making poor or dangerous decisions. That said, Francine is so minimalist that it has to rely almost entirely on Leo for solidity, and it would be a far stronger film if it supported and framed her more effectively.
Francine could easily pass for a belated sequel to Kelly Reichardt’s Wendy And Lucy; it centers on a similar impoverished, disconnected woman in trouble with the law, and prone to seek comfort in animals rather than people. As the film begins, Leo is leaving jail after an unknown sentence for a never-revealed crime. It’s unclear how she became so emotionally damaged, but her cringing approach to life suggests a history of abuse. She gets joy out of sunlight, wind, and swimming, but she approaches people awkwardly if at all, and pulls away from human connection without revealing why. She’s most at home with animals, but as she builds a menagerie of strays, her tiny small-town home disintegrates around her.
From the beginning, Leo’s tentative, defeated air suggests the film will be an exercise in waiting for the other shoe to drop. Because her goals are so minimal and her attempts to meet them are so timid, the story never suggests it will be about how she crawls up toward respectability, so much as how soon and how badly she falls. Cassidy and Shatzky follow her through barely connected snippets—a shot of her waiting tables gives way to humiliating, painful sex in a bathroom; a metal band performing in a vacant lot gives her a moment of thrashing freedom—but it’s all observational, with little dialogue and practically no connective tissue. Until one lyrical sequence near the end, the cuts tend to be artlessly abrupt and jarring, emphasizing disconnection, particularly as she moves from loud arcades and bars to quiet stable stalls and silent rooms.
The lack of continuity, development, or background puts a lot of pressure on Leo to make every moment significant, and to make her character appealing even when she’s navigating her wrecked house, flinging dog food everywhere as if feeding chickens in a yard. And to some degree, she pulls it off. Her situation attracts some degree of empathy, and her hesitant, vulnerable performance attracts another. But it’s hard to be absorbed into such a fragmented, impressionistic world, where the questions about what her actions mean, who she is, and who she’s becoming perpetually outweigh the information at hand. Leo alone is practically enough for a movie, but Francine is such a small, short, incomplete one that it’s easy to want more—or less of this, and more of her in a more fully realized role. | 2024-01-13T01:26:57.367551 | https://example.com/article/9610 |
Q:
Best way to detect installed version in a PHP web application
We have a custom PHP/MySQL web application that gets updated and copied (using SFTP) to multiple servers regularly.
For some of those updates, database or filesystem changes are needed. Typically this requires manually running an update.php script that checks and updates everything.
I'd like to be able to have the application check whether there are new updates to decrease manual activity.
What would be the best way to do this?
A:
You might be over thinking it?
Record the latest version in a file, record the current version in a script that you're uploading.
You can add as much complexity as you want to the version parsing.
Simplest possible example:
define('CURRENT_VERSION', 2);
$lastVersion = (int)file_get_contents('VERSION');
if (CURRENT_VERSION > $lastVersion) {
if (update()) {
file_put_contents('VERSION', CURRENT_VERSION);
}
}
You could also detect the presence of an update by the fact that update.php exists. If it's there, run it and delete it. Obviously add your own error checking and fallback if update fails.
if (file_exists('update.php')) {
require('update.php');
unlink('update.php');
}
| 2024-07-07T01:26:57.367551 | https://example.com/article/3252 |
Gedaliah Nadel
Rabbi Gedaliah Nadel (1923–2004) is known as one of the heads of Kollel Chazon Ish and was the leading decider of Jewish Law in the Chazon Ish neighborhood of Bene Barak. Rabbi Nadel is celebrated as an expert in all facets of Torah and Talmudic knowledge.
Biography
Gedaliah Nadel was born in Šiauliai, Lithuania to Rabbi Reuven Keshel. In 1936, his family immigrated to Palestine and settled in the Balfouria Colony in what is now Northern Israel. After briefly studying in Yeshiva in Tel Aviv, he switched to the Lomzha Yeshiva in Petach Tikvah. There, he studied under Rabbi Elazar Menachem Shach. As a child, Gedaliah was known for his diligence and devotion to studying Torah, reportedly studying for up to eighteen hours consecutively. Afterwards Gedaliah studied under Rabbi Avrohom Yeshaya Karelitz, known as Chazon Ish, who molded him into his principal student. Rabbi Nadel said that the Chazon Ish garnered knowledge in medical science by reading medical journals.
Rabbi Nadel married the daughter of Rabbi Eliyahu Weiner, a student of the Chofetz Chaim. Following the wedding, they lived in Jerusalem for a short while, before moving to Bene Barak. There he lived an ascetic life. When the Chazon Ish realized his dream of establish a special community of religious devotees, he chose Rabbi Gedaliah Nadel to lead the community. As leader of the community, such leaders as Rabbi Yaakov Yisrael Kanievsky, and his son Rabbi Chaim Kanievsky, turned to Rabbi Nadel for decisions. Rabbi Nadel's house was used as a gathering place for the Rabbinic personalities of his day, such as Rabbi Shach and Rabbi Y.Y. Kanievsky. Rabbi Nadel also briefly served as the Rosh Yeshiva of the Vishnitz Hasidic yeshiva. Rabbi Nadel would learn Maimonides' Guide for the Perplexed immediately after praying Shacharis vattikin.
Rabbi Nadel asserted that certain concepts mentioned in the Torah and Talmud that were not of a legalistic nature were not necessarily to be taken literally. His writings were analyzed and taken into account during the Slifkin affair. Rabbi Natan Slifkin himself subscribes to views espoused by Rabbi Nadel, including acceptance of evolution and belief "that the world was billions of years old." Rabbi Nadel maintained that it is acceptable to believe that the Zohar was not written by Rabbi Shimon bar Yochai and that it had a late authorship.
Rabbi Nadel's teachings have continued to exert influence on Hareidi leaders.
Works
Many of Rabbi Nadel's lectures were collected by his various students and two volumes were published under the title "Shiurei Reb Gedaliah." "B'Torato Shel Rav Gedaliah," a compilation of Rav Gedaliah's teachings prepared from audio recordings, was published by Rabbi Yitzchak Sheilat, one of Rabbi Nadel's main students, and at the personal request of Rabbi Nadel. In 5772, Rabbi Sheilat published a new edition of the book.
References
External links
B'Torato Shel Rav Gedaliah (from http://seri-levi.com/2008/11/13/mito/)
Rabbi Nathan Slifkin, Part I of Partial Translation of B'Toraso Shel R' Gedaliah
Rabbi Nathan Slifkin, Part II of Partial Translation of B'Toraso Shel R' Gedaliah
Category:1923 births
Category:2004 deaths
Category:Israeli Orthodox rabbis
Category:People from Bnei Brak
Category:People from Šiauliai | 2023-12-23T01:26:57.367551 | https://example.com/article/9274 |
Q:
Specify a time zone to use as the reference time zone
Based on the first two answers, the question was unclear as originally posted, thus I am completely rewriting it:
The following question is concerned only with how and what data is stored, and is no way shape or form concerned with converting data upon retrieval. As such, converting at SELECT to the desired time zone is not an appropriate answer.
When inserting a value into a timestamp with time zone field, it is retrieved (and thus presumably stored) with the timestamp converted to the local time zone of the database at the time it was inserted.
That is so say, a timestamp inserted as 2012-01-01 00:00:00+00:00 is retrieved as 2011-12-31 19:00:00-05, where the local time zone of the database at the time of the insert was -05. Timestamps that were inserted during daylight savings time, when the database was at -04, are returned using the -04 time zone.
What I want is for all timestamps to use an arbitrary time zone when stored (and thus all be retrieved without any additional work as having that time zone). That is to say, were the server orbiting the planet, all times would be at +00:00 (arbitrary time zone), instead of -12:00 to +12:00.
Can I insert into a timestamp with time zone column such that all timestamps are stored relative to an arbitrary time zone? If so, how?
Original follows.
When inserting a value into a timestamp with time zone field, it is being converted to the server's current time zone.
Example: If I insert a value specifying a time zone of -1, retrieving it will give back the time at -5 (the time zone of the server at the time it was inserted).
Is it possible to specify that it should be stored using an arbitrary time zone?
Note: This question is not how to convert the returned time to another time zone, this is specific to how the time is stored.
A:
You have to save the time zone offset in addition to the timestamp.
As @Milen already explained (and linked to the manual): a timestamp only saves a point in time (as abstract value). The time zone modifier is not saved, it only serves to adjust the timestamp relative to UTC.
Consider the following demo:
-- DROP TABLE tbl;
CREATE TEMP TABLE tbl (id int, myts timestamptz, mytz interval);
INSERT INTO tbl VALUES
(1, now() , EXTRACT (timezone from now()) * interval '1s')
,(2, '2012-01-01 00:00-05', interval '-5h')
,(3, '2012-01-01 00:00+04', interval '4h')
,(4, '2012-11-11 20:30+03', interval '3h');
SELECT *
,(myts AT TIME ZONE mytz)::text
|| CASE WHEN mytz > '0:0' THEN '+' ELSE '' END
|| to_char(mytz, 'FMHH24:mi') AS timestamp_at_origin
FROM tbl;
Run it locally to see. Pay special attention to the details of the AT TIME ZONE construct, and how I extract the time zone from the (local!) timestamp with time zone.
now() returns timestamp with time zone or timestamptz for short.
EXTRACT (timezone from now()) * interval '1s'
timestamp_at_origin displays the timestamp with time zone as seen at its origin. If I understood your question, then that is what you are looking for.
You could further improve formatting.
You may be interested in this related question which sheds some light on the ambiguities and pitfalls of time zones.
| 2024-03-02T01:26:57.367551 | https://example.com/article/3447 |
882 F.Supp. 770 (1995)
Sandra A. SMITH, Individually and as the Natural Mother and Legal Custodian of Marc W. Brown, a Minor Child Deceased, and Benjamen R. Brock, a Minor Child, Deceased, Cory Brock, a Minor Child, Deceased, and as Legal Guardian and Natural Mother of Jeramie D. Watkins, a Minor Child; Patti S. Gasper and Kenneth Gasper, Individually and as the Natural Parents and Legal Custodians of Nick J. Gasper, Minor Child, Deceased, and as Legal Custodians and Natural Parents of Luke D. Gasper, a Minor Child, Plaintiffs,
v.
FORD MOTOR COMPANY, Defendant.
No. 1:93-CV-143.
United States District Court, N.D. Indiana, Fort Wayne Division.
March 30, 1995.
*771 Sherrill W. Colvin, Cynthia Rockwell, John O. Feighner, Haller and Colvin, Fort Wayne, IN, Daniel J. Sigler, Zwick and Sigler, Decatur, IN, for Sandra A. Smith.
W. Scott Montross, Townsend Hovde and Montross, Indianapolis, IN, for Patti S. Gasper, Kenneth Gasper.
Mark A. Garvin, Richard E. Steinbronn, Barnes and Thornburg, Fort Wayne, IN, for Ford Motor Co.
ORDER
WILLIAM C. LEE, District Judge.
This matter is before the court on Defendant Ford Motor Company's Motion to Exclude Anticipated Testimony of James Grinolds. Ford filed the motion on January 24, 1995; plaintiffs responded on February 28, 1995; and Ford replied on March 10, 1995. The court held a Federal Rule of Evidence 104(a) hearing on March 17, 1995, and thereafter took the matter under advisement. For the following reasons, Ford's motion is granted in part and denied in part.
BACKGROUND
This product liability case arises out a vehicle fire involving a 1985 Ford F-150 Supercab with dual fuel tanks. On March 20, 1993, Patti Gasper, Sandy Smith, and the Smith and Gasper children were driving south on I-69. Their vehicle caught on fire, as a result of which the four children who were riding in the bed of the truck were killed.
Plaintiffs have identified James Grinolds as an expert witness whom they expect to call to testify at the trial of this matter. Ford moves the court to exclude from the trial (1) any opinion testimony by James Grinolds concerning the design of the fuel, exhaust, and other systems of the Ford F-150 because he lacks the necessary knowledge, skill, experience, training, or education to *772 provide testimony on such issues; and (2) any testimony by Grinolds concerning the cause of the fire in the Ford F-150 because such testimony does not constitute scientific knowledge that will assist the trier of fact to understand a fact in issue.
1. James Grinolds' Testimony Concerning the Adequacy of the Design of the Ford F-150 Is Inadmissible.
The Federal Rules of Evidence provide the following:
Rule 702 Testimony by Experts
If scientific, technical, or other specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact in issue, a witness qualified as an expert by knowledge, skill, experience, training or education, may testify thereto in the form of an opinion or otherwise.
F.R.E. 702. Rule 702 requires (1) that the witnesses qualify, by reason of training, education or experience, to state the proffered opinion, and (2) that the opinion will be helpful to the trier of fact. United States v. Carr, 965 F.2d 408, 412 (7th Cir.1992). "A district judge should assure himself, before admitting expert testimony, that the expert knows whereof he speaks." Bammerlin v. Navistar Int'l Transp. Corp., 30 F.3d 898, 901 (7th Cir.1994) (citing Daubert v. Merrell Dow Pharmaceuticals, Inc., ___ U.S. ___, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993)).
James Grinolds worked in the auto collision repair industry from approximately 1963 to 1983. While in the auto collision repair industry, Grinolds worked for a Chrysler dealership body shop and a General Motors dealership body shop and ultimately owned his own body shop. In this capacity, Grinolds was involved with repairing the structural and cosmetic components of vehicles, including but not limited to fuel systems, electrical systems, and power trains.
Grinolds formed his fire and accident investigation business in Minot, North Dakota, in approximately 1986. From 1986 to 1988, Grinolds attended the North Dakota School of Science, taking four class hours per week in fire science. His studies focused on structure and automobile fires, proper building construction, properties of fuels, and fire management. As a result of his efforts, Grinolds received certificates in fire management and fire science from the North Dakota School of Science. Since beginning his fire and accident investigation business, Grinolds has been employed by an automobile manufacturer, insurance companies, law enforcement agencies, and law firms to investigate fires and determine their cause. Grinolds previously has been involved in the investigation of a Ford F-Series vehicle fire.
Although James Grinolds is qualified to render certain expert opinions in this case, he is not qualified to testify about the design of the fuel or electrical systems of the Ford F-150 truck.[1] Grinolds has no education, experience, or training in the design of motor vehicle fuel or electrical systems; his background and experience are in the area of automobile body repair and fire and accident investigation. Determining the cause of an automobile accident or vehicle fire is very different from determining the safety and feasibility of fuel and electrical system designs for automobiles. Further, expertise in the areas of automobile body repair and fire and accident investigation does not translate into expertise in the design of fuel and electrical systems. Thus, Grinolds is not qualified, "by reason of training, education or experience, to state the proffered opinion" about the design of the Ford F-150 fuel and/or electrical systems. F.R.E. 702. See Perkins v. Volkswagen of America, Inc., 596 F.2d 681 (5th Cir.1979) (holding that a mechanical engineer with no experience in designing automobiles was permitted to express opinions on general mechanical engineering principles, but was prohibited from testifying as an expert in the area of automotive design); Hoban v. Grumman Corp., 717 F.Supp. 1129 (E.D.Va.1989) (finding that a licensed professional engineer with degrees *773 in electrical and mechanical engineering lacked the requisite "knowledge, skill, experience, training or education" to render opinions concerning aerodynamics or jet engine design); Tokio Marine & Fire Ins. Co. v. Grove Mfg. Co., 762 F.Supp. 1016, 1017-18 (D.P.R.1991) (deciding that the proposed witness's work as a civil engineer in the construction field did not qualify him as an expert concerning the design and manufacture of cranes).
2. James Grinolds' Testimony Concerning the Origin and Cause of the Fire in the Ford F-150 Is Admissible.[2]
James Grinolds stated the following during his deposition:
1. That the origin of the fire was under the bed of the vehicle in the area of the mid-ship fuel tank.
2. That the fire in the Ford F-150 was caused by one of the three following possibilities:
a. A short circuit in the wiring harness in the left frame rail of the F-150 which caused a breach in the nylon 12 fuel line (also located in the left frame rail) and ignited the fuel;
b. A short circuit in the wiring harness in the left frame rail caused by a gasoline leak from a selector valve, o-ring, or crack in the fuel line, which then breaches the nylon 12 fuel line and ignites the fuel; or
c. Heat from the exhaust system breached the nylon 12 fuel line and ignited the fuel.
Grinolds Deposition, pp. 109, 123, 158. "As to the first opinion, while Ford will not stipulate its agreement with Grinolds, it agrees that Grinolds is qualified under Rule 702 to render the opinion and that in reaching the opinion Grinolds relies upon facts or data reasonably relied upon by experts in the pertinent filed." Ford's Memorandum, p. 12. As to the second opinion (and its subparts), Ford contends that it is inadmissible. Ford claims that Grinolds' methodology in arriving at his opinions about the cause of the fire was "flawed and unscientific," thus rendering his causation opinions inadmissible. Ford's Reply Memorandum in Support of Its Motion to Exclude Portions of the Anticipated Testimony of James Grinolds ("Ford's Reply Memorandum"), p. 6.
Grinolds testified that when examining a vehicle to determine the cause and origin of a fire, he makes a determination by gathering as much data as he can about events leading up to the fire, identifying and interviewing eyewitnesses to the fire, and reviewing the evidence that remains. Additionally, if necessary, he will look at another vehicle that is intact in order to make a comparison. Grinolds had an opportunity to examine the Gasper vehicle, and in doing so, he looked at the burn patterns in the vehicle, the condition of the engine compartment, the condition of all of the remains of the vehicle, the wiring harnesses and components in the engine compartment, and all of the sheet metal. In arriving at his opinions, Grinolds also considered the descriptions given by the driver of the truck, one of the passengers in the truck, and another eyewitness to the fire. Also contributing to Grinolds' opinions is his knowledge about F-Series trucks and their components, along with his background, training, and experience in the area of fire and accident investigation.
Ford does not contest Grinolds' proposed testimony about the place of origin of the fire. Grinolds testified that the fire originated in the area of the midshift tank at a point close to the selector valve. Grinolds further testified that in arriving at this conclusion, he relied on the remains of the truck, descriptions of the fire, and the physical properties of the components in that area of the truck. He stated that the burn patterns indicate there was greater heat there for a longer period of time; and given the description and the physical properties of the components, that was only place it could have originated.
*774 Ford does contest Grinolds' proposed testimony about possible causes of the fire. Grinolds testified that he relied upon the following factors in arriving at his conclusions about the possible causes of the fire: descriptions of the fire; the condition of the vehicle; the characteristics of the components of the truck; the location of the truck's components; possible heat sources in the truck's components; testimony or statements of witnesses; and the physical properties of the materials that make up the fuel system and electrical system. Thus, Grinolds used essentially the same method as he did in arriving at his conclusion about the origin of the fire, a conclusion (and method) about which Ford does not contest Grinolds' ability to testify.
This court finds that Grinolds' proposed testimony satisfies the requirements of Federal Rule of Evidence 702. In Daubert, the Supreme Court found that before expert testimony is admitted, the trial court judge, acting as a "gatekeeper," must make a "preliminary assessment of whether the reasoning or methodology underlying the testimony is scientifically valid...." Daubert, ___ U.S. at ____, 113 S.Ct. at 2796. In other words, the judge must determine whether the expert's opinion has "a reliable basis in the knowledge and experience of ... [the expert's] discipline." This task requires that the district court rule out "`subjective belief or unsupported speculation.'" O'Conner v. Commonwealth Edison Co., 13 F.3d 1090, 1106 (7th Cir.1994) (quoting Daubert, ___ U.S. at ___, 113 S.Ct. at 2795). The Daubert Court noted:
Of course, it would be unreasonable to conclude that the subject of scientific testimony must be "known" to a certainty; arguably, there are no certainties in science.
Daubert, ___ U.S. at ___, 113 S.Ct. at 2795. From all indications, Grinolds has adhered to a reliable method of fire and accident investigation. To the extent that the rule in Daubert pertaining to "scientific knowledge" applies to this situation,[3] Grinolds' proposed testimony appears to have a reliable basis in his knowledge and experience and is not merely subjective belief or speculation.
Ford claims that Grinolds began his investigation with the conclusion of the existence of a product defect, and "reasoned from an end result in order to hypothesize *775 what needed to be known but was not." Ford's Reply Memorandum, p. 8. Ford states that two days after Grinolds first learned of the Gasper/Smith accident through a newspaper article, he sent a letter to attorney Dan Sigler, which discussed his view of the various defects in the Ford F Series truck. Ford asserts that "Grinolds set forth these opinions even though he had not inspected the vehicle, reviewed any of the investigative materials pertaining to the fire or spoken with any of the witnesses." Ford's Reply Memorandum, p. 7.
The court finds that when Grinolds sent his initial letter to Sigler, he did not intend it to represent his opinions on the cause and origin of the fire. Grinolds testified at the hearing that he did not intend the letter to be opinions on the cause and origin of the Gasper truck fire. He stated that the purpose of the letters was to try to give a general description of some of the components of the F-Series trucks. This testimony is consistent with the content of the letter. The March 24, 1993, letter states "RE: FORD F SERIES TRUCK FIRES" at the top of the first page. This indicates that Grinolds was writing generally about Ford F-Series trucks, and was not rendering a conclusion about the Gasper/Smith accident. The content of the letter further supports the assertion that Grinolds was giving Sigler information about F-Series trucks based on his knowledge and experience with them. Thus, Grinolds' letter to Sigler is not evidence of a flawed methodology on the part of Grinolds.
At the hearing, Ford attempted to identify "proper accident reconstruction methodology" and have Grinolds agree as to these principles on the witness stand.[4] With respect to these "principles," Ford claims that Grinolds' opinions about causation are inadmissible because he included possibilities that are not in fact possible. Ford also asserts that Grinolds' opinions about causation are inadmissible because he has failed to rule out other possible causes of the fire. However, according to Grinolds' own testimony, his "methodology" involved his taking into account the following factors in arriving at a conclusion: descriptions of the fire; the condition of the vehicle; knowing the characteristics of the components of the truck; the location of the truck's components; possible heat sources in the truck's components; testimony or statements of witnesses; and the physical properties of the materials that make up the fuel system and electrical system. Ford does not claim that Grinolds failed to take any of these factors into account. "[I]t would be unreasonable to conclude that the subject of scientific testimony must be `known' to a certainty; arguably, there are no certainties in science." Daubert, ___ U.S. at ___, 113 S.Ct. at 2795. See also Porter v. Whitehall Laboratories, Inc., 9 F.3d 607, 613 (7th Cir.1993) ("The testimony need not be known to a certainty.").
Ford, in arguing with the possible causes of the fire that Grinolds may or may not have included in his opinions, is in fact arguing with Grinolds' conclusions, not his methodology in arriving at certain opinions. Ford can more appropriately attack Grinolds' inclusion or exclusion of certain possibilities upon cross-examination at trial. As the Supreme Court noted in Daubert,
Vigorous cross-examination, presentation of contrary evidence, and careful instruction on the burden of proof are the traditional and appropriate means of attacking shaky but admissible evidence.... Additionally, in the event the trial court concludes that the scintilla of evidence presented supporting a position is insufficient to allow a reasonable juror to conclude that the position more likely than not is true, the court remains free to direct a judgment *776 ... and likewise to grant summary judgment.
___ U.S. at ___, 113 S.Ct. at 2798.
Ford contends that Grinolds' opinions are based upon speculation and assumption. Ford claims that Grinolds found no physical evidence and did no testing to support either his theory that an electrical short circuit or heat from the exhaust may have breached the fuel line and ignited the fuel. At the hearing, Ford questioned Grinolds about whether he had conducted testing himself or was aware of any testing that had shown that the exhaust system can produce sufficient heat to melt nylon 12 fuel lines. While Grinolds admitted that he had never conducted such testing, he stated that he was aware of exhaust temperature tests to demonstrate the melting of nylon 12 fuel lines that had been conducted. Thus, to the extent that this specific allegation by Ford is a methodology issue, Ford has failed to show that Grinolds was not aware of any testing to prove the point upon which he relies with regard to the exhaust system producing the heat that may have breached the fuel line. Additionally, as mentioned earlier, see supra note 3, the testing factor is not as relevant to the methodology issue in this case, which does not present a "novel" scientific technique.
Furthermore, in response to Ford's argument that Grinolds did not find physical evidence to support his theory, any paucity of evidence with regard to the remains of the truck was a direct result of the fire. Grinolds necessarily relied on many additional factors in arriving at his opinions: descriptions of the fire; the characteristics of the components of the truck; the location of the truck's components; possible heat sources in the truck's components; testimony or statements of witnesses; and the physical properties of the materials that make up the fuel system and electrical system.
Plaintiffs have satisfied the requirements of Rule 702 by demonstrating that James Grinolds' opinions have "a reliable basis in the knowledge and experience of his discipline." Daubert, ___ U.S. at ___, 113 S.Ct. at 2796. Ford's alleged problems with Grinolds' testimony about the cause of the fire in the Ford F-150 are most appropriately raised on cross-examination.
CONCLUSION
For the foregoing reasons, Ford's Motion to Exclude Portions of the Anticipated Testimony of James Grinolds in GRANTED IN PART and DENIED IN PART. James Grinolds' testimony concerning the adequacy of the design of the Ford F-150 is inadmissible; and James Grinolds' testimony concerning the origin and cause of the fire in the Ford F-150 is admissible.
NOTES
[1] Indeed, even Ford concedes that "Grinolds' studies in arson investigation and traffic accident reconstruction may qualify him, under Rule 702, to offer opinions concerning origin and cause of automobile accidents, and automobile fires in particular, but they do not qualify him as an expert in the area of automotive design." Ford's Memorandum of Law in Support of Motion to Exclude Portions of the Anticipated Testimony of James Grinolds ("Ford's Memorandum"), p. 5.
[2] The initial inquiry, whether Grinolds is qualified to render such opinions, is not at issue in this case. As mentioned earlier, see supra note 1, even Ford concedes that Grinolds may be qualified to offer opinions concerning the origin and cause of automobile accidents and fires and does not contest this point.
[3] The issue in Daubert was whether serious birth defects were caused by the mothers' prenatal ingestion of Bendectin, a prescription drug marketed by Merrell Dow Pharmaceuticals, Inc.; therefore, the Daubert case discusses F.R.E. 702 in terms of "scientific" evidence. As has been noted by some courts, Daubert's focus appears to be on "novel scientific evidence" and the "junk science" problem. See, e.g., Iacobelli Const. Inc. v. County of Monroe, 32 F.3d 19 (2d Cir.1994) (holding that affidavits submitted by a geotechnical consultant and an underground-construction consultant summarizing and analyzing site conditions, contract documents, and project results did not present the kind of "junk science" problem that Daubert meant to address); Lappe v. American Honda Motor Co., Inc., 857 F.Supp. 222 (N.D.N.Y.1994) ("Daubert only prescribes judicial intervention for expert testimony approaching the outer boundaries of traditional scientific and technological knowledge.") However, as the Daubert Court itself stated, "we do not read the requirements of Rule 702 to apply specially or exclusively to unconventional evidence." Daubert, ___ U.S. at ___ n. 11, 113 S.Ct. at 2796 n. 11.
The Court provided a non-exclusive list of four factors, none of which alone is determinative, that would be useful in determining the soundness of the scientific methodology from which the proffered conclusions are derived: (1) whether the theory or technique can be or has been tested; (2) whether the theory or technique has been subjected to peer review and publication; (3) the known or potential rate of error or the existence of standards; and (4) whether the theory or technique has been generally accepted. Id. at ___-___, 113 S.Ct. at 2796-97. However, before discussing these factors, the Court stated,
"We are confident that federal judges possess the capacity to undertake this review. Many factors will bear on the inquiry, and we do not presume to set out a definitive checklist or test."
Id. at ___, 113 S.Ct. at 2796.
In the present case, because of the nature of the expert witness, the four factors laid out in Daubert are not as readily applicable. In cases where a novel scientific theory or technique is presented, these four factors are effective means of determining whether such a theory or technique "will assist the trier of fact to understand or determine a fact in issue." The participation of Grinolds is not based on novel scientific evidence or testimony. In this case, he will participate as a fire and accident investigator and convey opinions relating to the Ford F-150 fire. His opinions are based on facts, an investigation, and traditional automobile body repair and fire and accident investigation expertise.
[4] Ford identifies the following principles as "scientific methodologies":
(1) One should review all the evidence before forming a conclusion;
(2) One should make certain that everything that is considered as a possibility is in fact possible;
(3) Everything included as possible causes should be consistent with the known evidence;
(4) One should not exclude a possibility unless the facts require its exclusion; and
(5) One should not assume things, particularly those things which are capable of being proven one way or the other.
Hearing Transcript, pp. 81-84.
| 2024-06-03T01:26:57.367551 | https://example.com/article/2199 |
This is an easy quick and delicious recipe for lunch or snack especially when zucchini is in season.
Ingredients:
2 big or 4 medium zucchini grated
1 tsp salt
1/4 cup all-purpose flour
1 egg beaten
1 garlic crushed
1/4 cup scallions finely diced
1/2 tsp cumin powder (optional)
1/2 tsp crushed red pepper (optional)
1/2 tsp salt
1/8 tsp black pepper
4 tbs olive oil for frying
Pita bread and lemon wedges for serving
Method:
Place the shredded zucchini in a colander set over a bowl and sprinkle the zucchini lightly with salt. Allow the zucchini to stand for 10 minutes. Using your hands, squeeze out as much liquid from the zucchini as possible.
Transfer the zucchini to a large bowl.
Add the flour, eggs, scallions, garlic, salt, pepper, cumin and red pepper (if used) to the bowl, stirring until the mixture is combined. Line a plate with paper towels.
Add the olive oil to large sauté pan set over medium heat. Once the oil is hot, scoop 1 tablespoon mounds of the zucchini mixture into the pan, (or 2 tbs depends how big in diameter you want the fritters) pressing them lightly into rounds and spacing them at least 2 inches apart. Cook the zucchini fritters for 2 to 3 minutes, then flip them once and cook an additional 2 minutes until golden brown and cooked throughout. Transfer the zucchini fritters to the paper towel-lined plate.
Repeat the scooping and cooking process with the remaining zucchini mixture.
in a big bowl, mix together ajwain, salt, flour and ghee. Rub the flour well in between the palms to incorporate the ghee well.
Add water little by little and mix the flour to dough. Cover and set aside until the filling is ready.
In a sauté pan, heat the 2 tbs of ghee butter, then add cumin until begins to sizzle.
Add ginger, green chilli and saute until it begins to smell good. Add ground beef an brown for 5 minutes or until all liquid absorbed.
5. Add smashed potatoes, peas, chili powder, garam masala, crushed coriande, mango powder and salt. Mix everything well and saute for 3 to 5 mins. Set aside to cool.
How to assemble Samosa:
Make 8 portions of the dough and roll to balls.
2. Grease the rolling area and then flatten a ball. Begin to roll to a oblong or oval shape. Cut it to two. If the edges are too thick, gently roll it to thin down.
4. Smear water over the straight edge and join the edges to make a cone. Press gently to seal the cone.
5. Fill the cone with potato masala and smear water on the edges. Bring the edges together and make a pleat on one side. Bring back the pleat and seal it. Make sure the samosa has been sealed well.
6. Finish making all the samosas.
How to deep fry samosa
Heat oil in a pan until medium hot. A piece of dough dropped in the hot oil must rise gently not immediately. This is the right temperature.
Add the samosas to the oil and deep fry them until golden. When they are half fried you can increase the flame to little high and fry until crusty.
Gözleme is a traditional savory Turkish flatbread, made of hand-rolled dough that are filled with various toppings, sealed, and cooked over a griddle. The name derives from the Turkish word göz, meaning “compartment”, in reference to the pocket of dough.
Ingredients:
For the dough:
2 1/4 cup all purpose flour
1/2 tsp salt
1/2 cup plain yogurt room temperature
1/2 cup warm water
For Filling
1 tbs vegetable oil
1 small onion chopped
1/2 lb ground beef
1 tsp ground cinnamon
1/2 tsp turmeric
1 tsp crushed red pepper شطة مجروشة
2 tsp ground cumin
salt and pepper to taste
1/2 cup green bell pepper chopped
1/2 red bell pepper chopped
1/2 yellow bell pepper chopped
1/2 cup fresh parsley chopped
1/2 cup feta cheese crumbled
1 medium tomato seeded and chopped
lemon wedges and olives to serve
Method:
For the dough:
In a large bowl combine the flour and salt. Make a well in the center and stir in the yogurt and water. Continue mixing with a spoon until everything is well combined. If the dough is too dry adjust by adding a bit more water, start with a tablespoon at a time.
Flour your work surface and knead the dough for about 3 minutes, or until smooth and elastic. Cover with a clean tea towel and set aside.
To make the filling, in a medium skillet, heat the oil over medium heat, sauté onion until translucent.
Add the ground beef to the skillet and break it up with a wooden spoon. Cook for about 4 to 5 minutes, until browned. Stir in the bell peppers, parsley, crushed red pepper, cinnamon, turmeric, cumin, salt and pepper. Add the chopped tomatoes and stir everything together and cook for another 2 minutes.
Transfer the meat mixture to a wide dish and let cool slightly.
Meanwhile, divide the dough into 4 equal balls. Roll each piece into a rectangle that’s about 12×10 inches.
Spoon about 3 tbsp of the lamb mixture in the middle of the rectangle and spread it out a bit. Add some of the crumbled feta. Fold over the edges of the dough to seal. Repeat with remaining dough pieces.
Heat a large nonstick griddle over medium heat. Depending on the size of your griddle, add 1 or 2 of the gozleme and cook on each side for about 3 to 4 minutes or until golden brown and crisp.
Cut each gozleme in half diagonally and serve with lemon wedges. Enjoy!
Zaatar bread … مطبق الزعتر . Is a Palestinian pastry mostly made in spring, which is the season of collecting fresh wild thyme. It is a flat bread stuffed with sautéed fresh zaatar leaves, onions and sumac and I like to add crumbled white cheese to mine. When baking these pastries my house smells phenomenal.
مطبق الزعتر الاخضر الفلسطيني
2 1/2 all purpose flour
1/2 cup whole weat flour
1/4 cup olive oil
1 tbs rapid rise yeast
1 tsp honey
2 tbs powdered milk
1 tsp salt
1 1/2 lukewarm water
More olive oil for brushing
For filling
3 cups fresh green thyme (zaatar)
1 medium size onion diced
2 tbs sumac
1/2 tsp black pepper
1/2 tsp salt
4 tbs olive oil
White cheese crumbled (optional)
Method:
To make the dough: in the mixer bowl add flour, salt, powdered milk and mix to combine.Make a well in the middle and stir in the yeast and honey and 1/4 cup of the luke warm water.
Let rest 5 minutes until the yeast mixture is frothy. Add the rest of the water with the olive oil and with the paddle hook attached to the mixer, mix for a minute.
Replace the paddle hook with the dough hook and start kneading until a soft dough ball is obtained and the sides of the bowl are clean.
Brush some olive oil on top of the dough and cover to rest for about 20 minutes.
Meanwhile start preparing the Zaatar filling.
Take off the Zaatar out of the branches and wash it for 3 times by changing the water each time until the water comes out clean.
Get rid of excess water by squeezing the zaatar leaves between your hands.
In a medium sized frying pan and on medium heat on the stove add olive oil and sauté the onions until translucent.
Stir in salt, black pepper, sumac, and Zaatar .
Sauté for a minute and transfer the Zaatar mixture to a plate to cool down.
In a big tray that is brushed with olive oil, divide the dough into 8 equal sized balls.
Brush a little olive oil on top of a clean surface and start rolling out each ball into very thin squares.
Brush some olive oil on top of each square to help rolling out.
Spread a table spoon of the thyme ( Zaatar ) mixture in the middle and top it with a tablespoon of the cheese.
Fold dough the same way as in the picture.
Keep repeating the same procedure with all the balls.
Leave them aside to rest for 10 minutes. Meanwhile, preheat the oven to 400F.
With the roller pin flatten each square to make it bigger (as shown in picture).
In a oven tray covered with parchment paper arrange the Zaatar squares and brush the top with olive oil.
Bake in the middle rack of the preheated oven for about 10 minutes or until golden brown.
Arrange the Zaatar pastries in the serving plate and brush with more olive oil.
Swiss chard and lentils soup is a traditional Old Palestinian soup that is made with chard, potatoes, whole chickpeas and lots of lemon juice. What you get is a naturally gluten free vegan nutritious, delicious and satisfying meal for lunch or dinner any day of the week.
Ingredients:
1 cup lentils washed
Bunch of Swiss chard
1 small onion diced
2 small potatoes diced
1 cup whole cooked chickpeas drained
1 tsp dried coriander crushed
1/2 tsp cumin powder
1 tsp dried mint
1/2 tsp dried chili crushed ( optional)
2 tbs corn oil
3 cups vegetable broth
Juice of 2 lemons
1 tsp dried mint
Salt to taste
Method:
Wash theSwiss chard really good
Wash the lentils
Cut the stems of the chard leaves and set them aside
Chop the leaves roughly and the stems.
In a medium sized pot heat the 2 tbs of oil and sauté the onion until translucent.
Stir in the coriander, cumin and chili.
Stir in the lentils and potatoes.
Add the broth and let cook on a medium low heat for about 30 minutes.
Stir in the Swiss chard and chickpeas and let cook for about 20 minutes more. | 2023-12-17T01:26:57.367551 | https://example.com/article/9935 |
Game Theory, Body Scanners and Security Theater
I’m back… and regularly again. After a brief hiatus from blogging to build two of my portfolio companies, I am back to to being the scourge of the ultra right wing posse and the loony left.
Game theory is a personal interest of mine and is widely quoted and misunderstood so it’s hard to master. In a nutshell, it says people or parties act in their own self interest in the conduct of their activities, affairs and negotiations. That’s easy enough to understand but there is a second and perhaps more important aspect in the theory – people or parties will make rational informed decisions as they pursue their self interest.
Here lies the craziness behind all the security theater the country is undertaking. Al Qaeda’s rationale is something we do not fully comprehend or understand, even to this date.
Take for example the use of full body scanners. Scanners are now the “hot technology” du jour and is the nexus of debate between privacy advocates and security hawks. Privacy advocates say it’s a virtual body strip search. People concerned with health want limits on it’s image depth and penetration to body surface fearing radiation-type risks and destruction of your DNA. Security hawks want to put one in every gate in every airport worldwide.
Consider the following stories.
Time Magazine did a story on drug mules getting paid a lousy $3000 a trip to swallow 120 packets of cocaine wrapped in condoms. Each cocaine condom packet was 1 and 1/4 inch long. Considering the intestines are 27 feet long, that makes for a lot of cocaine and a quick $3000.
You can put a lot of cocaine or PETN in a drug mule or terrorist
Did the drug mule act in his best interest? He clearly thought so. Was it rational? Clearly not. Customs caught one of these mules because one of the condoms burst and the guy had a massive cocaine overdose. How many of these guys got through customs? I’ll bet you more than one and perhaps several dozen hapless mules got in and got paid their lousy $3000.
Saudi Arabian deputy minister of interior, Prince Mohammed bin Nayef, survived a bombing attack launched by an al Qaeda cell based in Yemen. The assassin used the same PETN explosive to attempt the assassination but he hid the explosive in his rectum with a remote trigger mechanism. There are conflicting reports later of PETN being in his underwear, but the payload location is irrelevant. It’s better to assume, PETN was up his rear end.
An article published in Sada al-Malahim, Al Qaeda’s online magazine titled “War is a Trick,” the group’s leader, Abu Basir al-Wuhayshi, advised would-be al-Qaida members to use small amounts of explosives to kill “apostates” and Western nationals, including on passenger aircraft and in airports.
What’s the game theory perspective?
Western nations do not understand the mind of the terrorist. These terrorists believe in becoming martyrs and going to heaven for blowing themselves up. If they get to the airport or any target for that matter, it’s game over. If some crackpot drug mule will swallow cocaine for $3000, what is stop some guy who wants a fast pass to heaven from shoving PETN up his rectum?
Guess what? Full body pat downs that even includes your crotch area and whole body scan systems WILL NOT catch something inside your body. All the security theater we will spend money on and all the privacy debates will not catch PETN mules. We are applying game theory rules as if we’re playing with ourselves. We are playing with terrorists who have different self interests and concepts of reason.
Does it really make a difference whether they blow themselves up in the crowded TSA lines or the airplane?
The point is there should be far more focus on identifying these bastards before they even get to cab headed to the airport. Sure there is the lone wolf phenomenon but it has proven to be far more rare than one would expect. None of the known thwarted terrorist attempts involved lone wolves. Intelligence simply has to get better before the airport check-in counter.
And in the airport, we just have to do what the Israelis do in Ben-Gurion airport. It will piss off privacy advocates and civil rights activists but if society wants safe air travel, there is a price to pay and an emasculated body scanner system is not enough. | 2023-10-23T01:26:57.367551 | https://example.com/article/3056 |
Image copyright Thinkstock Image caption Children in care are more vulnerable to mental health problems
Children in care are too often missing out on treatment for mental health problems despite being four times more likely to experience them, say MPs.
The Commons Education Committee says children fostered in England are sometimes denied treatment simply because they move placement too often.
They should instead be given priority for mental health support, the committee says in a report.
The government says it is investing £1.4bn in children's mental health.
Almost half of children in care have a diagnosable mental health disorder, the MPs heard, compared with about one in 10 children who are not in care.
But provision for children in care with mental health problems is poor in many parts of England, says the report.
A significant number of local authorities are failing to identify mental health issues when children enter care, it adds.
Poor provision
And in some areas children in care can be turned away because their conditions are not deemed severe enough to qualify for treatment.
The MPs heard some child and adolescent mental health services (Camhs) were unwilling to begin treatment if a child moved into a new foster placement, even if this was within the same local authority.
"Given children in care may have unstable family lives and are frequently moving foster or residential placement, this inflexibility puts vulnerable children in care at serious disadvantage in getting the support they deserve," said Neil Carmichael, who chairs the committee.
"This must change - we recommend children in care be given priority access to mental health assessments and never refused care based on their placement or severity of their condition."
Image copyright Other Image caption Committee chairman Neil Carmichael said vulnerable children in care were at a serious disadvantage in accessing mental health care
The committee heard evidence from one 16-year-old girl in foster care who had waited for more than two and a half years to be seen by her local Camhs.
In her evidence, given last December, the girl told the committee she had not been seen because she had moved placement 13 times.
Her foster carer told the committee the 16-year-old had been with her for 10 months and was still waiting for an appointment.
This was in spite of official guidance that states: "Looked-after children should never be refused a service, including for mental health, on the grounds of their placement."
Annie's evidence
Annie (not her real name), from Faversham, Kent, had been taken into care because she had not felt safe at home - but the experience was "traumatic" from the outset, she told MPs.
She said: "They took me out and they could not find me a placement at the start, so they expected me to go back to my family after they had taken me away, an hour later I said, 'I can't do that, I'm sorry. My life is at risk if I do that.'"
She had had 13 different placements in her first two and a half years in care, she said.
"I have locked everybody away," she said. "I went to so many [placements], I gave up giving myself to people."
Despite her need for counselling, the local Camhs would not give her an appointment because of her frequent moves.
"They will not move around with the child, which they should do," Christine Malone, Annie's foster carer for the past 10 months, said.
Annie missed six months of school and most of her GCSEs because of a placement on completely the opposite side of the county.
At school, she found it difficult to talk to teachers because of too much "baggage" and severe bullying.
She is now in college but told MPs: "It is like your work, college - you do not tell them everything about your personal life do you?
"It can be awkward as well, like with depression or self-harm, you are embarrassed and do not want people to know because you feel like they will judge you in some way.
"So you need Camhs to be there as well, or have one person and not loads of different carers to help you through school. I had Christine, lucky enough."
Image copyright Thinkstock Image caption Young people leaving care need more mental health support, say the MPs
'Traumatic experiences'
Similar incidents of Camhs refusing to see children without stable placements were described in several other pieces of evidence to the committee.
Too often, schools are left to try to help children with mental health problems with little support from other agencies, its report says..
It also calls for more mental health support for young people leaving care, who are more than five times more likely than their peers to attempt suicide.
The report urges:
better mental health assessments of children entering care
better mental health support for care leavers
better integration of services for children in care, with more focus on mental health
more mental health training for teachers, foster carers and children's home staff
a bigger voice for children in care on the services they receive
better research on the prevalence of mental illness in children
A government spokesman said: "Children in care have often lived through traumatic experiences, and it is vital they receive the support they need.
"That's why we are putting a record £1.4bn into children and young people's mental health, and investing in better links between these services and schools.
Image copyright FotoCuisinette Image caption The government says it is putting more money into children's mental health care
"This is backed up by £700m in reforming the social work profession, so staff are supported to make the right decisions for those in their care."
The Local Government Association's community wellbeing spokeswoman, Izzi Seccombe, said mental health services for children needed urgent improvement.
She said it was wrong for any child to have to wait a long time for treatment, but for the level of support needed, there had to be a joint approach with every organisation involved in a young person's life, such as schools, carers and health services, as well as councils.
If the government's vision to improve mental health services was to be realised, it needed to be backed with proper funding, she added.
Young Minds chief executive Sarah Brennan said the charity had long advocated giving children in care priority for mental health treatment.
They could have traumatic experiences but the impact on their mental health was too often ignored, she said.
"Yet these are the young people who too often go on to have the most severe and enduring mental health conditions throughout their lives," said Ms Brennan.
"If they receive the right help early on, as all the evidence shows, much of this long-term distress for them, and their friends and families, could be avoided." | 2024-06-24T01:26:57.367551 | https://example.com/article/2880 |
Health, Fitness and the Workplace
Get out of your seat
A recent report from the department for digital culture, media and sport, examined the impact that being inactive has on your health and wellbeing. The report reveals that workers can spend up to three quarters of their day sitting down, which contributes to a range of preventable health conditions including the two leading causes of workplace absence, namely back injuries and stress, depression or anxiety. In 2016/17, some 1.3 million workers suffered from work related ill-health, which equated to 25.7 million working days lost. This has been estimated to cost £522 per employee and up to £32 billion per year for UK businesses.
Moral of the story – encourage people to be more active and make your workplace more enjoyable and and your business will benefit from having a healthier, more productive workforce.
As fitness and health issues become increasingly more talked about, anything an employer can do to help is likely to be welcomed by employees. | 2024-05-29T01:26:57.367551 | https://example.com/article/7869 |
A Spanish man apparently raised by wolves says since leaving the pack he has been disappointed by humans. Have others reportedly reared by animals felt the same?
You would be forgiven for thinking stories of orphans growing up with wolves were the domain of The Jungle Book or the mythic founders of Rome, Romulus and Remus.
In reality, such tales might not be so far-fetched. In 1965, people living in Spain’s Morena mountains discovered Marcos Rodríguez Pantoja, who has been dubbed the Mowgli of Spain, who had apparently been living in a cave with a pack of wolves.
Pantoja’s mother died when he was young, and his father sold him to a local farmer who also died soon after. Pantoja, who was the subject of a documentary in 2010, estimates he was six when he retreated into the mountains, to live with the pack that was to become his family for the next 12 years. Now 72, he has revealed that in the years since, human life has been a disappointment. He has even attempted to go back to the mountains, but, as he told El Pais, “it’s not what it used to be” – the wolves have forgotten him. “If I call out they are going to respond, but are not going to approach me,” he says. “I smell like people, I wear cologne.”
How does his experience compare with that of other children apparently raised by wolves?
Shamdeo
Shamdeo was four years old when he was discovered in a forest in India in 1972. According to photographer Julia Fullerton-Batten, who recently undertook a project called Feral Children, he was playing with wolf cubs. “He had sharpened teeth, long hooked fingernails, matted hair and calluses on his palms, elbows and knees. He was fond of chicken-hunting, would eat earth and had a craving for blood.” The boy died in 1985 – it’s not clear if he ever fully adjusted to life with humans.
The wolf boy of Hesse
A story dating back to 1304 tells of a boy kidnapped by wolves and raised in the German region of Hesse. According to a translation by Karl Steel, an expert in critical animal studies and posthumanism, the boy’s meals were prepared by the doting wolves, which even, according to reports, made a small pit furnished with leaves to protect him from the cold. Like Pantoja, he apparently preferred living among the wolves.
The Chilean “Dog Boy”
ABC News reported in 2001 that the “Dog boy”, who has not been named, was abandoned by his 16-year-old mother and, after spending a period in care, fled to live among a pack of wild dogs in a cave. The boy would scavenge with the dogs for food and eat out of bins. It is said the boy wanted to remain with his new family so much that he even jumped into the Pacific Ocean to evade capture by the police. | 2024-01-19T01:26:57.367551 | https://example.com/article/3120 |
LAS VEGAS -- UNLV coach Marvin Menzies has added another player to his first recruiting class, signing JUCO player Jovan Mooring.
Mooring's signing was announced by the school Wednesday.
A 6-foot-2 guard from Country Club Hills, Illinois, Mooring averaged 26.6 points, 6.2 rebounds, 3.8 assists and 3 steals for South Suburban College as a sophomore.
He was named the NJCAA Division II basketball player of the year and earned MVP honors in the national tournament. | 2023-12-22T01:26:57.367551 | https://example.com/article/9125 |
Q:
Why is Java picking this overload over the other
Could anyone explain to me why Java is picking the second overload instead of the first?
public static void foo (int a, double b, double...c) {}
public static void foo (double...a) {}
public static void bar ()
{
// this is the second
foo(1);
}
I thought when I pass 1 as the argument, Java would have picked the first argument because int is more specific than double?
Thanks
A:
The second method is the only one that can match. You have only 1 argument. The first foo has at least two required: an int and a double, so that can't match.
The second foo matches because any number of numbers can match. Java will implicitly promote your int 1 to a double so it can match, with method invocation conversion.
| 2024-04-07T01:26:57.367551 | https://example.com/article/4532 |
[Descriptive study of factors facilitating and restricting the development of critical care nursing competence].
In the literature, few studies have examined the development of critical care nursing competence. Given that nursing practice is different in these units and that the nursing workforce is less and less experienced, the goal of this study is to describe the factors facilitating and restricting the development of critical care nursing competence. A semi-supervised interview conducted in the spirit of Chinn and Kramer (2008) was used to collect data from 10 nurses, categorized according to Benner's (1995) levels of competency. Results demonstrated that 27 factors grouped on the basis of individual, contextual and organizational perspectives facilitated or restricted the development of competence by nurses working in intensive care. Although some factors had already been documented in the literature, several factors listed in this study were clarified for the first time. | 2023-08-26T01:26:57.367551 | https://example.com/article/4785 |
What is ‘salvation’? Jesus, the Saviour, came to earth as a little babe (Luke 2:7) to bring us peace with God. Basically, ever since the fall of Adam, all people have been tainted with his sin. Essentially, sin is disobedience … Continue reading → | 2024-03-02T01:26:57.367551 | https://example.com/article/6280 |
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html>
<head>
<title>一点知识系统</title>
<meta http-equiv="keywords" content="keyword1,keyword2,keyword3"/>
<meta http-equiv="description" content="this is my page"/>
<meta http-equiv="content-type" content="text/html; charset=UTF-8"/>
<script type="text/javascript" src="../../head.js"></script>
</head>
<body>
<form id="my-form">
<div id="tools" class="datagrid-toolbar">
<a onclick="save();" class="easyui-linkbutton" iconCls="icon-save" plain="true">保存</a>
<a onclick="window.location.reload();" class="easyui-linkbutton" iconCls="icon-reload2" plain="true">刷新</a>
<a onclick="closeWindow();" class="easyui-linkbutton" iconCls="icon-close" plain="true">关闭</a>
</div>
<p id="errorMsg" style="color:red;margin-top:5px;"></p>
<table class="formTable" >
<tr>
<td class="tdc1">
动态标题：
</td>
<td align="left">
<input type="text" name="cynamictitle" id="cynamictitle" class="easyui-validatebox" data-options="required:true" maxlength="100" />
</td>
</tr>
<tr>
<td class="tdc1">
内容：
</td>
<td align="left">
<input type="text" name="cynamiccontent" id="cynamiccontent" class="easyui-validatebox" data-options="required:true" maxlength="100" />
</td>
</tr>
<tr>
<td class="pn-flabel pn-flabel-h">知识图片:</td>
<td class="pn-fcontent">
<input type="file" id="collegeimage"style="width:180px;" name="collegeimage"
style="width:150px;"
onchange="uploadFile();"/><br />
<span style="color: red;">* 建议280*280</span>
</td>
</tr>
<tr>
<td class="tdc1">
备注：
</td>
<td>
<textarea id="remark" name="remark"></textarea>
</td>
</tr>
</table>
</form>
</body>
</html>
<script>
function checkFormMe(form){
return $('#my-form').form('validate');
}
window.onload = function(){
for(i=0;i<=228;i++){
var icon = "<div onclick=\"checkIcon('icons"+i+"');\" class='icons"+i+"' style='border:1px solid #E0ECFF;width:18px;height:19px;float: left;margin-bottom: 3px;margin-left: 3px;cursor: pointer;'></div>";
$('#checkIcon').append(icon);
}
}
function checkIcon(ico){
$('#icon').removeClass();
$('#icon').addClass(ico);
$('#img').val(ico);
$('#checkIcons').dialog('close')
}
var imageurl = "";
function uploadFile(){
$.ajaxFileUpload ({
url:contextPath+"/system/uploadFile",//用于文件上传的服务器端请求地址
secureuri: false, //是否需要安全协议，一般设置为false
fileElementId: "collegeimage", //文件上传域的ID
dataType: 'json', //返回值类型 一般设置为json
data:{},
success: function (res){//服务器成功响应处理函数
if(res['flag'] == "success"){
$.messager.alert("提示","操作成功!");
imageurl = res['id'];
}else{
$.messager.alert("提示","操作失败!",'error');
}
},
error: function (data, status, e){//服务器响应失败处理函数
alert(e);
}
});
}
var id = getPram('id');
$('#parentId').val(id);
function save(){
if(checkFormMe()){
var arg = $('#my-form').serializeArray();
jQuery.post(contextPath+"/system/addCynamic?imageurl="+imageurl,arg,
function (data) {
var res =$.parseJSON(data.toString());
if(res['flag'] == "success"){
$.messager.alert("提示","操作成功!",'info',function(rt){
window.parent.query();
closeWindow();
});
}else{
$.messager.alert("提示","操作失败!",'error');
}
});
}
}
</script> | 2023-08-31T01:26:57.367551 | https://example.com/article/1824 |
Q:
foreach loop in mvc5 is not working
I am working on a view which shows the table data in a table format in mvc.
its shows an error that Model does not contain a definition for GetEnumerator.
MY VIEW
@model IEnumerable<TestComponents.Models.testorder>
<!DOCTYPE html>
<html>
<head>
</head>
<body>
<table id="myTable">
<tr class="header">
<th style="width:60%;">TEST NAME</th>
<th style="width:40%;">TEST ID</th>
</tr>
@foreach (var items in Model)
{
<tr>
<td>@items.searchtest</td>
<td>@items.searchtestid</td>
</tr>
}
</table>
MY CONTROLLER
public ActionResult test()
{
testorder to = new testorder();
DbHandle dh = new DbHandle();
to.searchtest = dh.searchtes('S', to);
to.searchtestid = dh.searchtesid('S', to);
return View("test", to);
}
MY DBHANDLE MODEL
public List<string> searchtes(char flag, testorder to)
{
List<string> items = new List<string>();
connection();
SqlCommand cmd = new SqlCommand("SPNAME", con);
cmd.CommandType = CommandType.StoredProcedure;
cmd.Parameters.AddWithValue("@PstrOperationFlag", flag);
cmd.Parameters.AddWithValue("@Pstrtestname", 'w');
con.Open();
SqlDataAdapter sd = new SqlDataAdapter(cmd);
DataTable dt = new DataTable();
sd.Fill(dt);
using (SqlDataReader sdr = cmd.ExecuteReader())
{
while (sdr.Read())
{
items.Add(sdr["testname"].ToString());
}
}
con.Close();
return items;
}
I don't know where I am getting wrong or why it doesn't getting retrieved from model to my view.
A:
Your Return model is not collection so that Your View Should be :
I have added new model with controller action and method check below code it will work for you.
Here was the problem in your code so I have modify and added into collection collection : to.searchtest =dh.searchtes('S', to);
I have modify this model to create below model because no need of this List< string > element now :
public partial class ItemModel
{
public virtual List<string> itemId { get; set; }
public virtual List<string> itemName { get; set; }
}
**So your new model should be :**
public partial class ItemModel
{
public virtual string itemId { get; set; }
public virtual string itemName { get; set; }
}
And Your **Controller Action** should be :
public ActionResult test()
{
ItemModel model = new ItemModel();
List<ItemModel> itemList = new List<ItemModel>();
testorder to = new testorder();
DbHandle dh = new DbHandle();
var itemmodel = dh.searchtes('S', to);
//Second way if not work first way
foreach (var itemvalue in itemmodel)
{
model.itemId = itemvalue.itemId;
model.itemName = itemvalue.itemName;
itemList.Add(model);
}
return View("test", itemList);
}
//your method should be
public List<ItemModel> searchtes(char flag, testorder to)
{
List<string> items = new List<string>();
connection();
SqlCommand cmd = new SqlCommand("SPNAME", con);
cmd.CommandType = CommandType.StoredProcedure;
cmd.Parameters.AddWithValue("@PstrOperationFlag", flag);
cmd.Parameters.AddWithValue("@Pstrtestname", 'w');
con.Open();
SqlDataAdapter sd = new SqlDataAdapter(cmd);
DataTable dt = new DataTable();
sd.Fill(dt);
ItemModel item = new ItemModel();
List<ItemModel> itemList = new List<ItemModel>();
//item.itemId.Add(sdr["testId"].ToString());
//item.itemName.Add(sdr["testname"].ToString());
using (SqlDataReader sdr = cmd.ExecuteReader())
{
while (sdr.Read())
{
item.itemId = sdr["testId"].ToString();
item.itemName = sdr["testname"].ToString();
itemList.Add(item);
}
}
con.Close();
return itemList;
}
And here we go for View Page :
@model IEnumerable<TestComponents.Models.ItemModel>
<!DOCTYPE html>
<html>
<head>
</head>
<body>
<table id="myTable">
<tr class="header">
<th style="width:60%;">TEST NAME</th>
<th style="width:40%;">TEST ID</th>
</tr>
@foreach (var items in Model)
{
<tr>
<td>@items.itemId</td>
<td>@items.itemName</td>
</tr>
}
</table>
Definitely this will help you !!
Cheers!!
| 2024-01-10T01:26:57.367551 | https://example.com/article/2819 |
Keynes pretty much admitted that his political goal as a conservative was saving capitalism at a time socialism was rife and the conditions existed for capitalism to fail. It worked, and that system was it place pretty much until the fall of the Berlin Wall, when capital figured it was safe again to assert itself decisively over labor.
Neoliberalism was already in place to do this through the efforts of Thatcher and Reagan. Now the world is entering the Second Great Depression, and the political outcome is uncertain for a variety of reasons. But one thing seem certain, the game is changing.
Bob, you are a clearly a racist imbecile. I would ask you to read "Traitor to his Class" and "The Plots Against the President" and so many others actual histories that document Wall Street's war against FDR but I sincerely doubt you are literate. Everyone I mention the actual histories for the benefit of everyone else.
The American Fascist movement funded by Wall St. become the "libertarian" movement and this is well documented.
Bob probably thinks Lincoln was a dictator based on the historical revisionist nonsense from the Neoconfederate racist DiLorenzo and wonder how longer before Bob starts posting that revisionist crap here.
It's time to put the foot down. Stop treating these racist so-called "Libertarian" neoconfederates with any respect. It's time to call this "free market" guise for resegregationist dixiecrats for what they are. Bob is one of them and so is everything from the Klan connected Mises Institute in Alabama. It's time to add them the official hate sites along with Storm Front.
The promotion of racism under the guise of economic "liberty" shouldn't be tolerated by this blog or any other respectable economics blog. These are not issues up for debate in polite company. | 2024-05-10T01:26:57.367551 | https://example.com/article/4526 |
Effect of organic amendments on degradation of atrazine.
Pesticide contamination of soil and ground water at or near the agricultural fields is a major problem world wide. The ability of several amendments like rice straw, manure, saw dust and charcoal were used to stimulate the degradation of atrazine in soil. Field soil fortified with pesticide at two concentration levels were amended separately with rice straw, farm yard manure, saw dust and charcoal at rates of 2.5% (w/w) and maintained at field capacity moisture regime and kept at ambient temperature 25 +/- 5 degrees C. The results indicate 89.5% degradation of atrazine in farm yard manure during 60-day period followed by rice straw, saw dust charcoal and recording 87.2% and 83.8%, 67.7%, respectively, as compared to unamended treatment where 63.3% degradation was observed. The FYM was found to be most effective in soil and enhances the degradation as compared to the other amendments. Although addition of organic manures has been an integral part of sustainable agriculture practices; the present findings give a new dimension of it's utilization for removal of persistent pesticides. | 2024-01-12T01:26:57.367551 | https://example.com/article/9087 |
Category Archives: Education
Childhood is a defining time in everyone’s life. The experiences had during this time can stay with and effect a person forever. This is especially true when it comes to connecting to the natural world. When I think of my own connection to nature, I see that it is rooted in the memories I have of hiking, sailing, and exploring during Summers in Maine. The younger someone is when they foster this connection, the deeper it is. However, modern children spend the majority of their time in doors. In the U.S., kids in public school have an average of 27 minutes of recess a day. Their time at home is typically spent inside playing video games and watching TV. Children are simply not spending enough time in the outdoors.
Source: Ars Technica
This lack of outdoor time is especially harmful to the youngest schoolchildren. Research is developing that shows 3-7 year olds need free play, especially in the outdoors, to develop skills like critical thinking, risk management, and grit. The typical kindergarten includes more reading and math time than play time though. Children don’t get the chance to explore the world with their senses in these schools, but are expected to inherently understand how to use their brains and sit still. This is not where their development is at though. An exploration into the senses is necessary to develop critically thought, and the more a child moves and the stronger they are, meaning they can better hold themselves still for long periods of time. Free play, especially in the outdoors, is not just a nice thing for kids to do but a necessity for their healthy development.
In many parts of Europe this is already understood. Forest preschools have become common. The children who attend these schools spend every day, all day, outside, no matter the weather. They may have wooden shelters to hide from the rain and snow, and good clothing to keep them warm. Their playground is the forest around them – they climb the trees, get muddy in the brooks, and learn through interacting with the ecosystem. In the older grades this concentration on outdoor play continues, with many public schools offering recesses of over an hour in playgrounds that incorporate natural features such as trees and logs. Clearly this is a superior model for getting kids close to nature.
The Fiddleheads Forest School in Seattle, WA – Source: New York Times
Some in the U.S. have realized the importance of the outdoors in childhood development. Forest playgrounds have begun to pop up around the country. Other types of alternative learning schools such as Waldorf and Montessori encourage both the free play and outdoor time children so desperately need. However, this type of learning is currently available only to those who can pay for it, and is confined to a few institutions. To truly make a difference this type of learning must be incorporated into the public school system. Our children are missing out on nature, and it is our job to make sure this deficit ends as soon as possible. | 2023-10-28T01:26:57.367551 | https://example.com/article/7510 |
Dear Monument Australia visitors, we are a self-funded, non-profit organisation, dedicated to recording monuments throughout Australia. Over time the costs of maintaining this website have risen substantially (in fact they are probably larger than those of many companies who exist for profit). In the past we have borne all the costs associated with maintaining the website but we are now having difficulties in paying the monthly expenses.
If Monument Australia is useful to you, please make a donation to keep this historical and educational resource available. Your help would be greatly appreciated.
The roll contains five panels with over 700 names being recorded with the centre panel containing the names of those who died in service or were killed in action during the conflict.
The honour board was presented to Roma Town Council and Bungil Shire Council by Alderman Crawford and erected in the Roma Town Hall. At the beginning of the war a small honour board was erected in the Town Hall but it was found to be inadequate because of the number of men who had enlisted from the district. The roll was unveiled on the 11th November 1921 by Colonel Bourne.Western Star and Roma Advertiser, (Toowoomba)
16th November 1921 | 2023-12-16T01:26:57.367551 | https://example.com/article/7824 |
Q:
How to REPLACE a string in SQL Query
I have a table set up with the following rows:
Id_Customers | Name
-------------------------------------------------------
KG001TEI5 | Dave Sullivan
RU024GHE4 | Rose Red White
MK842LGS9 | AKBLC DHEIFL
I want to REPLACE Name value with the following rule:
Replace ABC with DEF and DEF with ABC
Example: Original Name AKBLC DHEIFL after replacing DKELF AHBICL
This is how I approach it:
CREATE FUNCTION [dbo].[replace_name](@inputString VARCHAR(200))
RETURNS VARCHAR(200)
AS
BEGIN
DECLARE @increment INT= 1;
DECLARE @tmpString VARCHAR(1);
WHILE @increment <= DATALENGTH(@inputString)
BEGIN
SET @tmpString = (SUBSTRING(@inputString, @increment, 1));
IF(@tmpString LIKE '%[ABCDEF]%')
BEGIN
IF(@tmpString ='A')
SET @inputString = STUFF(@inputString, @increment, 1, 'D');
ELSE IF(@tmpString ='B')
SET @inputString = STUFF(@inputString, @increment, 1, 'E');
ELSE IF(@tmpString ='C')
SET @inputString = STUFF(@inputString, @increment, 1, 'F');
ELSE IF(@tmpString ='D')
SET @inputString = STUFF(@inputString, @increment, 1, 'A');
ELSE IF(@tmpString ='E')
SET @inputString = STUFF(@inputString, @increment, 1, 'B');
ELSE IF(@tmpString ='F')
SET @inputString = STUFF(@inputString, @increment, 1, 'C');
END;
SET @increment = @increment + 1;
END;
RETURN @inputString;
END;
GO
Is there a better way to create it?
I'm not sure if this is good with million rows.
How to create a query to fill the requirements?
A:
If you use SQL Server 2017+, you may use TRANSLATE():
Table:
CREATE TABLE Data (
Id_Customers varchar(9),
Name varchar(50)
)
INSERT INTO Data (Id_Customers, Name)
VALUES
('KG001TEI5', 'Dave Sullivan'),
('RU024GHE4', 'Rose Red White'),
('MK842LGS9', 'AKBLC DHEIFL')
Statement:
UPDATE Data
SET Name = TRANSLATE(Name, 'ABCDEF', 'DEFABC')
Result:
Id_Customers Name
KG001TEI5 ADvB SullivDn
RU024GHE4 RosB RBA WhitB
MK842LGS9 DKELF AHBICL
For earlier versions, you need to use nested REPLACEs. Choose a character, which is not present in your data (? in the example):
UPDATE Data
SET Name = REPLACE(
REPLACE(
REPLACE(
REPLACE(
REPLACE(
REPLACE(
REPLACE(
REPLACE(
REPLACE(Name, 'F', '?'),
'C', 'F'),
'?', 'C'),
'E', '?'),
'B', 'E'),
'?', 'B'),
'A', '?'),
'D', 'A'),
'?', 'D')
| 2023-08-28T01:26:57.367551 | https://example.com/article/8077 |
The Dana-Farber Cancer Institute is an independent federally designated Comprehensive Cancer Center affiliated with the Harvard Medical School. The aims of the Institute can be summarized as follows: a) to develop and maintain clinical and laboratory research programs of high quality related to biology, cause and prevention, diagnosis and treatment of cancer; b) to provide total care for cancer patients on an interdisciplinary basis and in association with neighboring institutions; c) to serve as the major cancer center in the Harvard Medical School area and to play a leadership role for cancer research and patient care in New England providing expertise in cancer diagnosis and treatment; d) to add to the total oncology resources of the medical community in terms of available staff, facilities and expertise; e) to participate in the teaching and training of medical students, house staff, fellows and graduate students and in the continuing education of practicing physicians and other health professionals. | 2023-08-10T01:26:57.367551 | https://example.com/article/6190 |
Q:
SQL not in to Linq
I have a (Oracle)SQL query that I want to translate to Linq to use on a NHibernate context.
What the query does is to get me all the Functions that has been deleted from a given Module (ie. they don't exist in the given module but are linked to a predecessor Module)
"SELECT * FROM Function WHERE ModID = " + module.PredecessorID + "
AND FncName NOT IN
(SELECT FncName FROM Function WHERE ModId = " + module.ModID + " )"
Here is a image to explain the data model
I'm looking for an expression with the "dot notation", something like this:
DBContext.GetAll<Function>()
.Where(x => x.Module.ModID == module.PredecessorID)...
A:
You could do this in two steps (for readability).
var excludedNames = DbContext.GetAll<Function>()
.Where(x => x.ModeID == module.ModID)
.Select(x => x.FncName);
var result = DBContext.GetAll<Function>()
.Where(x => x.Module.ModID == module.PredecessorID &&
!excludedNames.Contains(x.FncName));
| 2024-02-15T01:26:57.367551 | https://example.com/article/1788 |
Front pattern of invasion in squamous cell carcinoma of the penis: new prognostic factor for predicting risk of lymph node metastases.
Treatment of penile carcinoma is based on excision of the primary tumor and regional lymphadenectomies. Lymph node metastasis is the most important prognostic factor, yet it cannot be reliably evaluated by imaging studies or clinical examinations; thus, new prognostic factors such as the pattern of invasion must be investigated. The data of 112 patients with penile carcinoma who underwent amputation and lymphadenectomy were analyzed. The clinicopathologic variables were studied, in addition to the front pattern of invasion according to the standards established by the modified criteria of Anneroth and Bryne first used as a part of a prognostic score in squamous cell carcinoma of the oral cavity. The pattern of invasion expresses the tumor-host tissue relationship, demonstrating the infiltrative tumor characteristics. This pattern was defined as infiltrating (invasion in blocks of small solid strands of cell tumors that broadly infiltrated the organ's stroma) and pushing infiltration (tumor cells have invaded large cell blocks with well-defined tumor-host interfaces). The variables that were significantly associated with the presence of lymph node metastases on univariate analysis were clinical N stage (P = 0.011), lymphatic permeation (P <0.001), venous embolization (P = 0.025), and the pattern of invasion (P <0.001). In the multivariate model, the lymphatic permeation, clinical N stage, and invasion type were independent risk factors for lymph node metastases. The follow-up varied from 0.1 to 452.8 months (mean 76.9, median 27.3). Lymphatic permeation, clinical N stage, and standard infiltrating invasion were independent prognostic variables for the risk of lymph node metastasis. | 2024-02-22T01:26:57.367551 | https://example.com/article/4278 |
Intel: if we can’t defeat them, we’ll join them
Intel lost a recent opportunity to save itself from the mobile disaster, but there’s still hope for them. Not as makers, though: the recent deal made with ARM will allow other microprocessor designers to take advantage of Intel’s resources in the production process.
That could be a good way to leverage the technology and experience Intel has accumulated through all this years, but it’s also another sign -we didn’t need much more of these- of an Intel that threw in the towel long ago in the mobile space.
AMD made a similar move when GlobalFoundries spun off, and it went pretty well for them. We’ll see what happens with this new strategy from Intel. | 2024-01-28T01:26:57.367551 | https://example.com/article/2305 |
South Africa’s High Court has ruled that Jacob Zuma’s appointment of a state prosecutor to decide whether to reinstate corruption charges against him was not valid.
The court also ruled that President Zuma should not make a new appointment. It said deputy president Cyril Ramaphosa should appoint a new public prosecutor within 60 days.
The ruling is a stinging rebuke for Mr Zuma, who is due to step down in 2019 after more than a decade in power. The President’s office said he would appeal the decision.
The National Prosecuting Authority (NPA), whose head prosecutor Shaun Abrahams was removed from his post by the High Court ruling, was not available to comment.
The rand was buoyed by the court ruling, with analysts saying markets saw Mr Zuma’s influence curtailed.
Judge president Dunstan Mlambo said at the Pretoria High Court the appointment of Mr Abrahams as the national director of public prosecutions was “invalid and set aside”.
In his ruling, judge Mlambo said: “In our view, President Zuma would be clearly conflicted in having to appoint a national director of public prosecutions, given the background ... and particularly the ever present spectre of the many criminal charges against him that have not gone away.”
The charges against Mr Zuma relate to a 30 billion rand (£1.5bn) government arms deal arranged in the late 1990s and have amplified calls for Mr Zuma to step down before his term as President ends in 2019.
In October the Supreme Court of Appeal upheld an earlier decision by a lower court that the nearly 800 corruption charges filed against Mr Zuma before he became President be reinstated.
It then fell to Mr Abrahams, appointed by Mr Zuma as chief state prosecutor in 2015, to decide whether or not the NPA would pursue a case against Mr Zuma.
Mr Abrahams gave Mr Zuma until the end of November to make representations to prevent the charges being brought against him. Mr Abrahams has not commented on whether Mr Zuma presented any submissions of what decision he has taken on the matter.
The case in the Pretoria High Court had been brought by a trio of civil society organisations seeking an order declaring the previous state prosecutor’s removal was invalid.
The African National Congress (ANC), which is set elect Mr Zuma’s successor next week, said it would allow the parties involved to reflect on the judgment and its implications as well as decide whether to appeal or not.
Calls for Zuma resignation in Pretoria Show all 14 1 /14 Calls for Zuma resignation in Pretoria Calls for Zuma resignation in Pretoria Calls for Zuma resignation in Pretoria Riot police guard the entrance to a courthouse in Pretoria AP Calls for Zuma resignation in Pretoria Calls for Zuma resignation in Pretoria A protester holds a butternut squash painted with an image of South African President Jacob Zuma AP Calls for Zuma resignation in Pretoria Calls for Zuma resignation in Pretoria South African opposition party Economic Freedom Fighters' (EFF) supporters shout as they demonstrate for the release of a report into corruption allegations against South African president on front of the High Court in, Pretoria AFP/Getty Calls for Zuma resignation in Pretoria Calls for Zuma resignation in Pretoria Riot police guard the entrance to a courthouse in Pretoria AP Calls for Zuma resignation in Pretoria Calls for Zuma resignation in Pretoria An Economic Freedom Fighters (EFF) member burning tyres in Pretoria city centre, South Africa EPA Calls for Zuma resignation in Pretoria Calls for Zuma resignation in Pretoria Protestors looting a shop at a BP garage in Pretoria city centre, South Africa Reuters Calls for Zuma resignation in Pretoria Calls for Zuma resignation in Pretoria A protestor runs for cover as police officers fire stun grenades during a march to call for the removal of President Jacob Zuma Reuters Calls for Zuma resignation in Pretoria Calls for Zuma resignation in Pretoria Police officers stand guard during a march by the members of the Economic Freedom Fighters (EFF) Reuters Calls for Zuma resignation in Pretoria Calls for Zuma resignation in Pretoria A member of the Economic Freedom Fighters (EFF) carries a tire around his neck as they call for the removal of President Jacob Zuma outside the Union buildings in Pretoria Reuters Calls for Zuma resignation in Pretoria Calls for Zuma resignation in Pretoria Members of the Economic Freedom Fighters (EFF) hanging on the tree during an anti-government march outside the Union Building in Pretoria AP Calls for Zuma resignation in Pretoria Calls for Zuma resignation in Pretoria Members of the Economic Freedom Fighters (EFF) march during an anti-government protest outside the Union Building in Pretoria AP Calls for Zuma resignation in Pretoria Calls for Zuma resignation in Pretoria A metro police officer, runs past a burning barricade during an anti-government march in Pretoria AP Calls for Zuma resignation in Pretoria Calls for Zuma resignation in Pretoria A man carries a sledgehammer as South African opposition Economic Freedom Fighters (EFF) party members and supporters demonstrate in Pretoria AFP/Getty Calls for Zuma resignation in Pretoria Calls for Zuma resignation in Pretoria Economic Freedom Fighters supporters gather by a burning garbage bin as they march in downtown Pretoria AP
Ralph Mathekga, an independent analyst, said the court’s decision “will most likely be appealed if you look at government’s track record. It will surely end up in the Constitutional Court”.
“It will spilt the ANC even further. In a way it does empower Ramaphosa. But remember, there is faction in the ANC that is very suspicious of the courts and this could embolden Zuma and his allies’ anti-establishment campaign.” | 2023-12-14T01:26:57.367551 | https://example.com/article/4245 |
Living News
DENVER (AP) — Some governors in the U.S. West say a new Trump administration directive threatens to undermine a hard-won compromise aimed at saving a beleaguered bird scattered across their region. The directive, issued in…
BELLEVILLE, Ill. (AP) — Firefighters had to rescue a southwestern Illinois man from deep mud after he became stuck while trying to reach his pet parrot. Belleville firefighters were called Sunday afternoon to the city’s…
LONDON (AP) — A British fashion model has been convicted of murdering a more successful rival after a social media-fueled dispute. A jury found George Koh guilty on Monday of stabbing Harry Uzoka through the…
BUCHAREST, Romania (AP) — Israel’s embassy on Monday protested to the Romanian government after four Israeli tourists were allegedly dragged out of a taxi and beaten by riot police in Bucharest during a violent anti-corruption…
SEATTLE (AP) — U.S. investigators are working to find out how an airline employee stole an empty commercial airplane from the main airport serving Seattle and crashed into a small island after being chased by…
For our Shredded Barbecued Beef, we cut a chuck roast into quarters. The smaller pieces of beef absorbed more smoke flavor and cooked much faster. After cooking the meat in a disposable roasting pan on…
It’s hard to believe that the well-loved combination of cake, cream and strawberries could be anything other than perfection, but soggy cake, squishy cream, or bland, underripe berries can easily ruin this heavenly trio. We…
A recent study found people spend more time in the doctor’s waiting room than getting there or talking to the doctor. Lean Plate Club blogger Sally Squires said a local endocrinologist is looking for ways to make that time educational, and he’s not alone.
Living spotlight
DENVER (AP) — Some governors in the U.S. West say a new Trump administration directive threatens to undermine a hard-won compromise aimed at saving a beleaguered bird scattered across their region. The directive, issued in… | 2024-07-20T01:26:57.367551 | https://example.com/article/1097 |
Error Spoolsv.exe Windows Xp
Contents
Proffitt Forum moderator / June 16, 2006 1:42 AM PDT In reply to: Thanks Bob Any discussion would be off topic but it's fairly known that the quality of drivers is Facebook Twitter LinkedIn Google+ Link Public clipboards featuring this slide × No public clipboards found for this slide × Save the most important slides with Clipping Clipping is a handy BetaFlux 443,758 views 9:34 Call of Duty 1 Error Win7 Buffer overrun detected! ~problem fixed~ - Duration: 0:46. Please try again now or at a later time. navigate here
ERROR The requested URL could not be retrieved The following error was encountered while trying to retrieve the URL: http://0.0.0.1/ Connection to 0.0.0.1 failed. Sign in to make your opinion count. It's a Microsoft FixIt tool and I had to use the "Full" version instead of the "Light". Spoolsv.exe Error Windows 10 When the error occurs, you may receive an error message similar to the one given below. “Spoolsv.exe - Application Error" “The instruction at "0x77fcc2c0" referenced memory at "0x00000000". “The memory could
TheQwertystudios 46,190 views 5:13 How to Fix 0xc000007b Application Error - Duration: 5:55. Click here to Download Hitmanpro_x32bit.exe Click here to Download Hitmanpro_x64bit.exe Restart the computer once. To solve it, you should download spoolsv.exe from a trusted websites or copyit from other computer which use the same operating system, and then paste it under“system32”. http://www.instant-registry-fixes.org/how-to-fix-spoolsvexe-application-error/ QUESTION: Can somebody explain how this error occurs?
Ask a Question Make a donation All our tools, guides and support is completely free of cost. Kb312052 Loading... Show more Language: English Content location: United States Restricted Mode: Off History Help Loading... Post a reply Discussion is locked Flag Permalink You are posting a reply to: SPOOLSV.EXE Application Error The posting of advertisements, profanity, or personal attacks is prohibited.
Did you call it in?Bob Flag Permalink Reply This was helpful (0) Collapse - Thanks Bob by akayarea / June 16, 2006 1:26 AM PDT In reply to: Try the XP check over here by Dave Konkel [Moderator] / November 20, 2005 10:08 PM PST In reply to: Looking for a technical explanation of the spoolsv error According to this mskb entry. -- Dave K. It allows you to print in the background and is responsible for how Windows manages printing tasks on your computer. There delete all the files with the extensions " *.SPL " and " *.SHD " 9. Spoolsv.exe Application Error Windows Xp Sp3
Ads Join us on Facebook Recent Discount Coupons eScan Internet security 15% Promo eScan coupon code to get 15% discount on eScan Antivirus home user Zemana Promo - Get 30% Off desertsun02 7,891,125 views 5:32 runtime error program c windows system32 spoolsv.exe - Duration: 1:51. I has something to do with printers and spoolers.I use a stand-alone PC with Windows 2000 Pro and a HP and Brother printer.At startup after the login I get the message:Spoolsv.exe his comment is here Add to Want to watch this again later?
In Services, find and right click with your mouse on "Print Spooler" service and then click "Stop" 6.
Find a one more file named as spoolss.dll, do the same permission steps on it as well. Reply Joëlle YI August 3, 2016 at 9:22 am Thanks for the fix. Reply Daniel Ortega April 12, 2014 at 6:17 pm Thanks a lot !!!!!!!!!! 😀 my error was that a virus was modifying the path of the archieve this can make you Spoolsv.exe Download That did not work.
Click Start -> Control Panel Select Classic View - > Administrative tools Select Services option In the Service Management console, look for Spooler Services, right-click on it and Select “Stop” Open THANKS! And then double click "Services" 5. weblink Spoolsv.exe is a system and hidden file and the main component of printing services.
Click on Owner tab. Reply Inder November 3, 2014 at 1:50 pm Hi Just used your printer spooler repair tool. Reply Martin December 24, 2013 at 8:57 pm Im back again, i just wanted to say thank you very much, Been suffering with this shit for several months and now its We are not responsible for any data loss / business loss, device corruption or any other type of loss due to the use of any software or command mentioned in our
Only with an error-free computer, can youenjoy every day with the computer. How to Fix Print Spooler Error 0x800706b9: Not enough resources are available Issue How to fix print spooler error 1053 ? (Free Repair Tool) (Solved) Print Spooler Error 1068 - The Reply Roberto February 13, 2016 at 4:46 am i tried all steps, but in step 2 the .reg file installation failed. do i need windows disk to repair this?
Sign in to add this video to a playlist. Step 2.1: Download print spooler service default registry file (spooler.reg), as per your Operating System. Reply Arun April 8, 2015 at 8:07 pm Thank you, worked for me. Even $1 can a make a huge difference for us in our effort to continue to help others while keeping this site free: « How to Backup & Restore Chrome Bookmarks | 2024-05-29T01:26:57.367551 | https://example.com/article/3424 |
Injured Keys retires, new Mom Azarenka advances at Miami Open
Victoria Azarenka, playing her first Miami Open as a mother, battled through a tough opening set against 15th-ranked American Madison Keys on a chilly Thursday night, but their second-round match ended abruptly when Keys retired with a left thigh injury.
Azarenka won 7-6 (7-5), 2-0, calling it “the worst way to win a match”. The former world No. 1, whose ranking dropped to 186 during her maternity leave, wore her son Leo’s name on her sneakers.
She said motherhood has given her new perspective on tennis and life.
“I think the harmony and the balance within is definitely different, because your perspective definitely changes on what's important, because tennis, in the end of the day, is a job,” she said. You do this job for a certain period of time. And when you're a parent, this is for the rest of your life.”
Be the first to know.
No one covers what is happening in our community better than we do. And with a digital subscription, you'll never miss a local story.
In an earlier match, Simona Halep, the current No. 1, rallied from a set down to beat 98th-ranked Oceane Dodin of France 3-6, 6-3, 7-5 and will face Agnieszka Radwanska.
“I played bad, but, you know, you have these days, and when you don't give up, is the most important thing,” Halep said. “I think that's why I won today, because I didn't give up. I didn't play well, but, you know, in the end, if you win it, you take only the positives.”
Halep reached her third Grand Slam final at the Australian Open in January but is still awaiting her first trophy. She reached the Miami Open semifinals in 2015.
She said she has learned to be more patient and that has resulted in more three-set victories.
“Before maybe previous years I couldn't win the matches in the third set, so I'm really happy that I changed this to myself, and I feel stronger,” she said. “So, when I go to the third set, I don't panic. I'm relaxed and I'm just fighting till the end, and I'm sure that I have a big chance to win the match, so I just go for it.”
Kerber is 19-4 this year and has reached at least the quarterfinals of all five tournaments she's played after struggling through a difficult 2017 season. In 2016, she won the Australian Open and the U.S. Open. She was eager to win on Thursday after a humbling 6-0, 6-2 loss to Daria Kasatkina at Indian Wells, Calif., last week.
“After Indian Wells, it was really important to come here and to win the first match,” Kerber said. “The first matches are always tough, especially because the conditions are completely different than in Indian Wells. To be honest, I just forget about the last match in Indian Wells. I mean, it was not my day. Here the tournament starts from zero.”
She took the same approach to 2018 after a forgettable 2017.
“New Year's, that was I think the moment where I knew, okay, completely the 2017 is over and we start a new year,” Kerber said. “I think that was the moment. It was not on court. It was just outside. But that was the moment that 2017 is over and a completely new year starts. I felt right away very good.” | 2024-07-09T01:26:57.367551 | https://example.com/article/9733 |
Effect of testosterone undecanoate hormone on sperm and its level in the hemolymph of male mud spiny lobster, Panulirus polyphagus.
The present study aimed to determine the effect of testosterone undecanoate hormone on sperm quality (sperm viability) and sperm quantity (sperm counts) and its levels in the hemolymph of male mud spiny lobster, Panulirus polyphagus. Male P. polyphagus was injected laterally in fifth abdominal segment of pure hormone, Testosterone Undecanoate (TU) and ethanol at days 1, 8, 15, 22 and 29. Hemolymph of P. polyphagus was taken every two weeks and checked with Enzyme-linked immunosorbent assay (ELISA) to measure hormone levels. The mean sperm quality and quantity were increased due to increase the TU dose and TU levels also increase. The sperm quality, quantity and hormone levels were relevance each others. These findings indicate that TU injection should be evaluated as a practical way of improving sperm quality and quantity in commercial operations. | 2024-02-17T01:26:57.367551 | https://example.com/article/9882 |
By a 3-to-1 margin, Hong Kongers do not want their government to hand over NSA whistle-blower Edward Snowden if Washington demands his extradition.
Whether it is because they support Mr. Snowden’s free-speech and privacy agendas, or because they are upset by his claims that the US National Security Agency has been hacking into Hong Kong’s computer network, 49.9 percent of people asked in a poll published here Sunday said he should not be surrendered. Another 17.6 percent said he should be turned over. A third of respondents had not made up their minds.
“Nobody welcomes a fugitive, but now he is here we have to safeguard his rights,” said Freddy Chu, a young privacy activist, as he brandished a photo of Snowden at a small rally in support of the American in central Hong Kong on Saturday.
Snowden is believed to be in hiding somewhere in Hong Kong, from where he divulged his identity to The Guardian newspaper a week ago. Since then, in an interview with the South China Morning Post, he has accused the NSA of hacking into the backbone of Hong Kong’s Internet system.
“He is welcome to Hong Kong,” said another demonstrator, James Hon, as he helped hold up a banner belonging to the League in Defense of Hong Kong’s Freedoms. “He is upholding our core values – freedom of expression and privacy. He is a brother.”
Snowden’s presence here puts the former British colony in a difficult spot, potentially subject to pressure from both Washington and Beijing. For the time being, the United States has not lodged an extradition request and Chinese officials have not tipped their hand about what they think should happen to Snowden. But many Hong Kongers are uncomfortable.
That may explain the low turnout at Saturday’s demonstration outside the US consulate. Persistent rain did not help, but few people here see Snowden’s fate as very important to their own lives.
Still, his suggestions that the NSA has been hacking in Hong Kong have won him a measure of sympathy. “When we learned that they had hacked into our Internet hub at the Chinese University of Hong Kong, all of us with a computer felt we had been targeted,” said Yves Azemar, a French expatriate dealer in rare books.
Get the Monitor Stories you care about delivered to your inbox. By signing up, you agree to our Privacy Policy
Snowden’s presence here also draws international attention to Hong Kong’s unique status as a “special administrative region” of China, where the rules are very different from the mainland.
“This is a golden opportunity for Hong Kong to explain to the world … that we still enjoy judicial autonomy,” says Alan Leong, a legislator and head of the pro-democracy Civic Party. “It’s a chance to say how proud we are that Snowden chose Hong Kong as a refuge.” | 2023-12-07T01:26:57.367551 | https://example.com/article/7849 |
A shortage of mangoes in Western Australia's far north has forced Perth charities to refund hundreds of people who purchased boxes of the fruit as part of a fundraiser.
The Dogs Refuge Home and the Cat Haven have begun refunding people who pre-purchased a 10-kilogram box of mangoes for $40.
The mangoes were supposed to come directly from Kununurra growers, with part proceeds from every purchase donated to the charity.
But the fundraiser was cancelled due to a lack of available mangoes.
WA Department of Agriculture and Food industry officer Peter Johnson said it was a "strange" year for mango production.
"We've pretty much had two crops this year - there was an early one back in September, early October, and then there's going to be a bit of late fruit coming in mid-November - and not a lot in between… | 2024-03-01T01:26:57.367551 | https://example.com/article/6851 |
"The person said the subject was feeling suicidal and depressed, but when officers got there and conducted their investigation, it was revealed no suicidal remarks were made," the source said. | 2024-01-15T01:26:57.367551 | https://example.com/article/9958 |
Brian McMahon (rowing)
Brian McMahon (born July 24, 1961) is a Canadian rower, who was the coxswain of the Canadian men's eights team that won the gold medal at the 1984 Summer Olympics in Los Angeles, California. The rowing team was inducted into the BC Sports Hall of Fame in 1985, and the Canadian Olympic Hall of Fame in 2003.
References
Canadian Olympic Committee
Category:1961 births
Category:Canadian male rowers
Category:Coxswains (rowing)
Category:Living people
Category:Olympic gold medalists for Canada
Category:Olympic medalists in rowing
Category:Olympic rowers of Canada
Category:Rowers at the 1984 Summer Olympics
Category:Rowers at the 1988 Summer Olympics
Category:Rowers from Toronto
Category:Medalists at the 1984 Summer Olympics | 2024-06-23T01:26:57.367551 | https://example.com/article/9556 |
Fantasy Baseball Player Rater 2012 - Relief Pitcher
The 2012 Fantasy Baseball Player Rater is now posted for relief pitchers. Take a look at the top 20 fantasy baseball relief pitchers for 2012. Please feel free to comment and we can adjust the list accordingly. The list is fluid and open to change throughout the off season based on injuries, trades, free agent signings, line up situations, etc. | 2024-03-08T01:26:57.367551 | https://example.com/article/4467 |
Q:
How to check if a string is a valid date
I have a string: "31-02-2010" and want to check whether or not it is a valid date.
What is the best way to do it?
I need a method which which returns true if the string is a valid date and false if it is not.
A:
require 'date'
begin
Date.parse("31-02-2010")
rescue ArgumentError
# handle invalid date
end
A:
Here is a simple one liner:
DateTime.parse date rescue nil
I probably wouldn't recommend doing exactly this in every situation in real life as you force the caller to check for nil, eg. particularly when formatting. If you return a default date|error it may be friendlier.
A:
d, m, y = date_string.split '-'
Date.valid_date? y.to_i, m.to_i, d.to_i
| 2024-01-13T01:26:57.367551 | https://example.com/article/9661 |
---
abstract: 'Even though classic theories and models of discrete choice pose man as a rational being, it has been shown extensively that people persistently violate rationality in their actual choices. Recent models of decision-making take these violations often (partially) into account, however, a unified framework has not been established. Here we propose such a framework, inspired by the Ising model from statistical physics, and show that representing choice problems as a graph, together with a simple choice process, allows us to explain both rational decisions as well as violations of rationality.'
author:
- 'Joost Kruis$^{1\ast}$, Gunter Maris$^2$, Maarten Marsman$^1$, Dylan Molenaar$^1$, Maria Bolsinova$^2$ & Han L. J. van der Maas$^1$'
bibliography:
- 'physicsofchoice.bib'
title: 'The physics of (ir)rational choice'
---
University of Amsterdam
ACT-Next by ACT
(Ir)rational choice {#irrational-choice .unnumbered}
===================
Over the last century, our perception of peoples choice behaviour as being rational, consistent, and profit-maximising, shifted to one where we now describe human decision-making as biased, uncertain, and settling for a satisfactory outcome. The feature thought to determine the appeal of choice alternatives, first changed from expected value[@huygens1657; @laplace1814], to expected utility[@bernoulli1954; @neumann1955], and then to the prospect of the alternative[@kahneman1979]. Instead of mulling over a decision until it was clear what the best option was, people where shown to stop when an acceptable alternative had been found, a consequence of what was termed bounded rationality[@simon1955]. Rather than an unwavering regularity in their choices, people behave rather inconsistently, for example because of the endowment effect, which entails that people tend to value an alternative higher as soon as they have chosen it, compared to when they have not[@thaler1980].
Two competing classes of decision-making models are random- and fixed utility models, the first assuming that choice is deterministic but with the appeal of the alternatives as a random variable that changes over time[@thurstone1927; @marschak1960; @mcfadden1974], whereas the second posits that the appeal of alternatives is fixed, but the decision process is of probabilistic nature[@luce1959; @luce1965]. These models of choice often assume choices to be rational in nature by endowing them with properties such as the independence of an alternative’s appeal from irrelevant alternatives, and the existence of a, either weak or strong, transitivity in the ordering of choice appeal[@thurstone1927; @luce1959]. Although these are statistically desirable properties, it has been shown that they are often violated. Most well known of these violations are the similarity, attraction and compromise effect[@debreu1960; @restle1961; @becker1963; @luce1965; @krantz1967; @rumelhart1971; @tversky1972; @luce1977; @simonson1989; @simonson1992; @tversky1993].
With the increased understanding of the complexity of human decision-making, both the theories and models describing choice behaviour have become increasingly complex as well[@tversky1972; @roe2001]. Although these models take violations of rationality into account, this is either the case for one of the violations and not the other, or only for violations taking place under specific conditions. Developing a framework that integrates these (partial) theories and explains the key choice phenomena is thus a worthwhile effort.
In this article we propose a simple choice framework, that allows us to explain both rational and irrational choices. We achieve this by framing choices as a combination of cues and alternatives and the relation between them, and introduce a simple decision process that stops when the choice conditions are met for the first time, as such implementing bounded rationality. Using only this decision rule, known violations of rationality, and other choice phenomena such as the endowment effect, follow naturally from the model.
Choice Setup {#choice-setup .unnumbered}
============
Choices are framed as combinations of cues and alternatives. Cues represent the conditions of the choice, e.g. ‘buy a book’ or ‘select a present’, and alternatives describe the possible choices. The relations between these alternatives and cues, and between alternatives themselves, are what govern the initial conditions of a choice. An appropriate representation of this framework is a graph, the vertices of which correspond to alternatives and cues, and the edges describing the relationship between the vertices. Figure \[fig:01\] shows two examples of possible choice graphs.
(8,3) node(y2) [$C$]{} (5,0) node\[draw, line width=5pt\](x21) [$A_1$]{} (8,0) node\[draw, line width=3pt\](x22) [$A_2$]{} (11,0) node\[draw, line width=.5pt,dashed\](x23) [$A_3$]{}; (y2) – (x21); (y2) – (x22); (y2) – (x23); (x22.south east) to\[bend right=40\] (x23.south west);
To formalise a choice graph with $n$ vertices as a probability distribution, we let $\boldsymbol{\sigma} = [\sigma_1, \sigma_2, \dots, \sigma_n]$ be the state vector of the $n$ vertices, where $\sigma_i \in \{0,1\}$ denotes if vertex $i$ is active $(\sigma_i = 1)$ or inactive $(\sigma_i = 0)$, and endow the states with the following distribution;
$$\boldsymbol{\sigma} \sim p(\boldsymbol{\sigma}) = \frac{1}{Z_{\beta}} \exp{\left( \beta \left[\sum_i \sigma_i \, \mu_i + \sum_{\langle i,j \rangle} \sigma_i \sigma_j \, \omega_{ij} \right]\right)} \, .
\label{eq:01}$$
Equation \[eq:01\] can be recognised as an Ising model[@lenz1920; @ising1925] originally introduced in statistical physics to model the ferromagnetic behaviour of magnetic dipole moments of atomic spins on a lattice. Where $\mu_i$ is the force of the external magnetic field interacting with spin $i$, $\omega_{ij}$ the pairwise interaction between spins $i$ and $j$, and $\beta$ the inverse temperature of the system. $Z_{\beta}$ is the appropriate, temperature dependent, normalising constant to make the probabilities sum to one, also known as the partition function. While this application was later rejected, due to incompatibility with quantum mechanics, in favour of the Heisenberg model, it has remained highly popular and one of the most studied models in modern statistical physics[@niss2005]. Also outside the field of physics the Ising model has attracted attention as a way to model the joint distribution of binary variables as a function of main effects and pairwise interactions[@marsman2015; @epskamp2016; @kruis2016; @marsman2018].
In the current context $\mu_i$ can be perceived as the base appeal of cue/alternative $i$, whereas $\omega_{ij}$ specifies the strength of the relation between cues and/or alternatives $i$ and $j$. In Figure \[fig:01\] it is shown how differences in $\mu$ and $\omega$ are represented in the choice graphs. In equation \[eq:01\] can be seen that the inverse temperature $\beta$ scales the magnitude of both the base appeal and pairwise interactions. While the quantity of $\beta$ is known in thermodynamics as a function of the temperature and the Boltzmann constant, in the current application it is not identifiable. It can however be interpreted, in the context of decision-making, as a measure of choice consistency, as with lower $\beta$ choice behaviour becomes increasingly random, until $\beta = 0$, where the choice is purely random. It is assumed that before a person is faced with a choice, the internal state of the decision maker, the resting state, is in accordance with the Ising model from equation \[eq:01\].
When a person is confronted with a choice, the appropriate cue vertex $(c)$ gets activated $(\sigma_c = 1)$, and remains active during the choice process. Second, to represent that a single alternative needs to be chosen, lateral inhibition, a neural concept where an exited neuron inhibits its direct neighbours, is introduced between the alternatives that are part of the choice[@feldman1982]. In the framework this is represented by lowering the interaction between each pair of alternatives with a constant $(\varepsilon)$, such that otherwise independent alternatives become negatively related. The consequence of this inhibition is that a choice structure will naturally tend to move to a state in which, apart from the cue vertex, only one alternative will remain activated[@yuille1989]. Furthermore, the stronger the lateral inhibition, the faster the system will tend to the state with only one alternative remaining active. Factors such as time-pressure, importance of a choice (e.g. buying a house vs. buying a book), and general (in)decisiveness might be possible contributors to variations in both $\beta$ and $\varepsilon$. Imposing the condition that, next to the cue vertex, only one additional alternative vertex can remain active $(\sum_{i=1}^n \sigma_i = \sigma_+ = 2)$, on the resting state distribution, represent the choice condition where a single alternative needs to be chosen. Enforcing these conditions on the Ising model we obtain conditional choice probabilities for each alternative.
$$\begin{aligned}
p(\sigma_{a\neq c} = 1 \mid \sigma_c = 1,\sum_{i=1}^n \sigma_i = \sigma_+ = 2) &= \frac{\exp{\left(\beta\left[\mu_a + \, \omega_{ca}\right]\right)}}{\sum_{j \neq c} \exp{\left(\beta\left[\mu_j + \, \omega_{cj}\right]\right)}} \, .
\label{eq:02}\end{aligned}$$
Equation \[eq:02\] can be recognised as a multinomial logistic regression model, and shows that under the imposed conditions, the pairwise interactions between alternatives, either from pre-existing conditions or lateral inhibition, vanish, and the probability of choosing alternative $a$ is only dependent on its base appeal ($\mu_a$) and relation with the cue ($\omega_{ca}$). The conditional distribution thus preserves both independence of irrelevant alternatives, and stochastic transitivity, that is, the model has rational decision-making as its ground state.
Choice Process {#choice-process .unnumbered}
==============
Why irrational choice behaviour does occur is explained by the choice process that describes how a person moves from an instance of the Ising model in equation \[eq:01\], to an instance of the appropriately conditioned Ising model as seen in equation \[eq:02\]. We model the choice process, one of sequential elimination of alternatives, as a Markov process with the conditioned choice probabilities as their unique invariant distribution. For expository purposes we consider one of the simplest examples of such a process, a Metropolis algorithm with single-spin-flip dynamics[@newman1999].
Let $\boldsymbol{\sigma}$ denote the current state of the choice structure, then at every step in the decision process one alternative is randomly selected $(\sigma_s)$ and its value flipped $(\sigma^*_s = (\sigma_s-1)^2)$. The resulting configuration is our proposal state, denoted by $\boldsymbol{\sigma}^*$, which we accept with the appropriate probability.
$$\begin{aligned}
p(\boldsymbol{\sigma} \rightarrow \boldsymbol{\sigma}^*) &= \min\left(1,\frac{\exp{\left( \beta \left[\sum_i \sigma^*_i \, \mu_i + \sum_{\langle i,j \rangle} \sigma^*_i \sigma^*_j \, \omega^{\varepsilon}_{ij} \right]\right)}}{\exp{\left( \beta \left[\sum_i \sigma_i \, \mu_i + \sum_{\langle i,j \rangle} \sigma_i \sigma_j \, \omega^{\varepsilon}_{ij} \right]\right)}}\right) \, ,
\label{eq:03}\end{aligned}$$
where $\boldsymbol{\Omega}^{\varepsilon}$ is the lateral inhibition adjusted pairwise interaction matrix, in which the relations among the alternatives are lowered by a constant $(\varepsilon)$. Using these acceptance probabilities ensures that in the long run states are in accordance with the conditioned Ising model.
From the choice probabilities in equation \[eq:02\] it is clear that if we let the Markov chain run until convergence, the effect of lateral inhibition and any other (in)direct interactions between choice alternatives will have worn out, and only the direct interactions between cues and alternatives matter. As such, in this case it is guaranteed that choices are rational. However, if a person stops when the conditions (i.e., $\sigma_c = 1$ and $\sigma_+ = 2$) hold for the first time, the same is not true.
By introducing the rule that a person stops whenever the conditions of the choice problem are met for the first time, the model implements the idea of bounded rationality, which states that, as a function of limited cognitive capacity and time, people will search for an alternative that is satisfactory, instead of the most appealing one[@simon1955]. The major contribution of this mechanism is that it predict the occurrence of irrational choices, as we explain next.
Results {#results .unnumbered}
=======
For each of the three classic violations of rationality, known as the similarity-, attraction-, and compromise-effect, we provide a simple example and illustrate how the proposed choice structure allows us to explain the observed deviations from rational behaviour as a function of interactions between the alternatives, and the rule that a choice will be made as soon as the choice conditions hold for the first time. As the decision is the result of a process that has Markov properties, we can, as we show in the methods section, for small $n$, analytically derive the expected response frequencies, by establishing a stochastic matrix containing the transition probabilities $p(\boldsymbol{\sigma} \rightarrow \boldsymbol{\sigma}^*)$ for all $2^n$ possible configurations of the system.
It is important to realise that for each of the upcoming examples more than one possible way exists in which a choice can be structured, such that we obtain expected response frequencies that align with empirical observations. One reason for this is the interchangeability in, for example, the relations between the inverse temperature ($\beta$), and the main effects ($\boldsymbol{\mu}$) and pairwise interactions $(\boldsymbol{\Omega})$, or the relation between the cue and and alternative ($\omega_{ca}$) and the base appeal of the alternative $(\mu_a)$. Furthermore, as some readers may have noticed, while traditionally the Ising model $\sigma$ takes values in $\{-1,1\}$, we use $\sigma \in \{0,1\}$ in the current paper. While there exists a linear transformation of the parameters such that they are mathematically indistinguishable, and we would thus obtain the same response probabilities, their interpretation in a choice context would be radically different.
We do therefore not propose that the choice structures we present in the next part are to be taken as our theory of the particular decision. Rather, our goal is to demonstrate the utility of the proposed modelling framework, that derived from the Ising model, is capable of predicting both rational decision-making behaviour, and deviations of this rationality.
Similarity {#similarity .unnumbered}
----------
The classic example for irrational choice behaviour known as the similarity effect, was given by Debreu[@debreu1960], and goes as follows;
> “Let the set $U$ have the following three elements:
>
> - $D_C$, a recording of the Debussy quartet by the $C$ quartet,
>
> - $B_F$, a recording of the eighth symphony of Beethoven by the $B$ orchestra conducted by $F$,
>
> - $B_K$, a recording of the eighth symphony of Beethoven by the $B$ orchestra conducted by $K$,
>
> The subject will be presented with a subset of $U$, will be asked to choose an element in that subset, and will listen to the recording he has chosen. When presented with $\{D_C, B_F\}$ he chooses $D_C$ with probability ${{}^{3}\!/_{5}}$. When presented with $\{B_F, B_K\}$ he chooses $B_F$ with probability ${{}^{1}\!/_{2}}$. When presented with $\{D_C, B_K\}$ he chooses $D_C$ with probability ${{}^{3}\!/_{5}}$. What happens if he is presented with $\{D_C, B_F, B_K\}$? According to *rationality* (original text: the axiom), he must choose $D_C$ with probability ${{}^{3}\!/_{7}}$. Thus if he can choose between $D_C$ and $B_F$, he would rather have Debussy. However if he can choose between $D_C$, $B_F$, and $B_K$, while being indifferent between $B_F$ and $B_K$, he would rather have Beethoven.”
It is immediately clear that the assumptions underlying rational decision-making, would for the this scenario result in expected choice probabilities that are counter-intuitive. Specifically, in this case one would expect that when presented with $\{D_C, B_F, B_K\}$, $D_C$ would be chosen with probability ${{}^{3}\!/_{5}}$, and the remaining ${{}^{2}\!/_{5}}$ would be split evenly among $B_F$ and $B_K$. An intuition that has since been proven right in comparable experiments[@restle1961; @rumelhart1971; @tversky1972].
One choice structure[^1] that explains the occurrence of the similarity effect, does this by introducing a strong negative association between the two Beethoven recordings, as in shown in Figure \[fig:02\]. The negative relation between $B_F$ and $B_K$ has no influence on choice probabilities for any of the possible two-element subsets, as such the slightly larger base appeal of $D_C$ will results in choosing $D_C$ with probability ${{}^{3}\!/_{5}}$, when presented with $\{D_C, B_F\}$ or $\{D_C, B_K\}$. While still present when presented with $\{B_F, B_K\}$, the negative relation works equally in both ways, such that one chooses $B_F$ and $B_K$ both with probability ${{}^{1}\!/_{2}}$. While the conditional distribution of the model from equation \[eq:02\] predicts that $D_C$ will be chosen with probability ${{}^{3}\!/_{7}}$, when a choice has to be made from all three alternatives together, the rule that one stops as soon as the choice conditions hold for the first time, will actually predict that when presented with $\{D_C, B_F, B_K\}$, $D_C$ is chosen with probability ${{}^{3}\!/_{5}}$, whereas $B_F$ and $B_K$ are both chosen with probability ${{}^{1}\!/_{5}}$. Hence, the explanation of ‘irrational’ (yet intuitive) behaviour rest on the presence of a negative relation between the Beethoven recordings and the stopping rule.
(0,3) node(y) [$R$]{} (-3,0) node\[draw, line width=3pt\](x1) [$B_F$]{} (0,0) node\[draw, line width=3pt\](x2) [$B_K$]{} (3,0) node\[draw, line width=3.5pt\](x3) [$D_C$]{}; (y) – (x1); (y) – (x2); (y) – (x3); (x1.south east) to\[bend right=40\] (x2.south west);
One could argue that the approach of reverse engineering a network structure until one obtains the desired probabilities is a weakness of the framework. We do however believe that this is not a vice but rather a virtue, as it allows the researcher to frame a single process into multiple theoretically distinct structures. And while the expected probability distribution for each of these can be the same, interventions or manipulations will naturally result in distinct predictions. For example, if you were just presented with $\{D_C, B_F, B_K\}$ and chose Beethoven recordings by $K$. If you are then given the opportunity to choose one other recording from the remaining set $\{D_C, B_F\}$, you would probably never choose $B_F$ in this case. This is also what would happen in the current choice structure, as the model now takes the fact that you have $B_K$ already into account ($\sigma_{B_K} = 1$), essentially turning it into an extra cue, such that the negative relation between the two Beethoven recordings makes it unlikely that you will choose the remaining Beethoven recording.
Attraction {#attraction .unnumbered}
----------
Whereas the occurrence of the similarity effect has been explained by later models, for example, Tversky’s elimination-by-aspects model[@tversky1972], the attraction effect poses a bigger challenge. We consider a simple choice setup from Simonson & Tversky[@simonson1992], in which one group was offered a reward $(R)$ and had to choose between a nice pen $(P_+)$, which was chosen 36% of the time, or 6\$ $(\$)$, chosen 64% of the time. A second group was offered the same choice, now with a second, less attractive, pen $(P_-)$, added to the choice alternatives. As expected the less attractive pen was only chosen 2% of the time, however the proportion of people choosing the more attractive pen rose to 46%.
(-.5,3) node(3a)\[fill=white\] [**a**]{} (3,3) node(y2) [$R$]{} (0,0) node\[draw, line width=4.25pt\](x21) [$\$$]{} (3,0) node\[draw, line width=3.675pt\](x22) [$P_+$]{} (6,0) node\[draw, line width=.825pt\](x23) [$P_-$]{} (8.5,3) node(3b)\[fill=white\] [**b**]{} (12,3) node(y) [$R$]{} (9,0) node\[draw, line width=1.75pt\](x1) [$\$$]{} (12,0) node\[draw, line width=1.175pt\](x2) [$P_+$]{} (15,0) node\[draw, dashed, line width=1.675pt\](x3) [$P_-$]{}; (y) – (x1); (y) – (x2); (y) – (x3); (y2) – (x21); (y2) – (x22); (y2) – (x23); (x21.south east) to\[bend right=40\] (x23.south west); (x21.south east) to\[bend right=40\] (x22.south west); (x2.south east) to\[bend right=40\] (x3.south west);
A key assumption of rational choice is that of regularity, meaning that introducing a new option cannot make the choice probabilities of the other alternatives increase. It is clear that assumption this is violated in the current example, as adding a less appealing pen boosted the attraction of the nice pen, thereby increasing its choice probability. Figure \[fig:03\] shows two possible choice graphs for this experiment, both these structures result in expected response frequencies similar to those found in the experiment, however, each of these explain the occurrence of the attraction effect in a different way. While in Figure \[fig:01\]a the explanation of the attraction effect rests on the presence of a negative association between the money and both of the pens, in Figure \[fig:01\]b the effect is explained by a positive association between both of the pens. The framework presented in this paper thus provides two ways in which the mere addition of a less appealing alternative, can boost the selection of otherwise less frequently chosen alternatives.
As mentioned before, the fact that we can obtain the same results from different structures should be seen as virtue of the model, as it allows us to compare the different theories that underlie each of the different structures. For example, both choice graphs in Figure \[fig:03\] predict that adding the less attractive pen to the set of available alternatives boosts the selection probability of the nice pen with respect to the money. Now, if we would ask someone, who just chose the nice pen, to make a subsequent choice between the money and the less attractive pen, essentially everyone would choose the money. This is however not the case for the one of the choice graphs in Figure \[fig:03\]. While in the case of graph \[fig:03\]a the money would now be chosen more than 90% of the time, for graph \[fig:03\]b 33% of the people would still choose the less attractive pen. This illustrates that while both of these graphs explain our current observations, the new predictions that they make are different, where one of the options is obviously much less plausible. The fact that the new framework allows us to make these kinds of predictions, makes theories in this framework thus falsifiable.
Compromise {#compromise .unnumbered}
----------
The third violation, known as the compromise effect, can be demonstrated by another example from Tversky and Simonson[@tversky1993]. Here one group of subjects was offered a choice between a medium quality camera, with a regular price $(M)$, or a less expensive, but lower quality camera $(L)$. The second group was given the same choice with a high quality, but expensive, camera added to the set of alternatives $(H)$. In the first group the subjects chose both the low and medium quality/price cameras with equal probability. In the second group the medium camera was still chosen approximately half of the time, and the remainder split among the high and low quality/price cameras. As the medium and low camera’s are equally appealing in the first group, and the low and high quality camera are equally appealing in the second group, it is rational to judge the medium and high camera also to be equally attractive. As this clearly is not the case, like the attraction effect, adding a new alternative violates the regularity assumption.
(7,3) node(y2) [$C$]{} (4,0) node\[draw, line width=7.5pt\](x21) [$L$]{} (7,0) node\[draw, line width=7.5pt\](x22) [$M$]{} (10,0) node\[draw, line width=7.5pt\](x23) [$H$]{}; (y2) – (x21); (y2) – (x22); (y2) – (x23); (x21.south east) to\[bend right=40\] (x23.south west);
A possible explanation for this finding might be that (dis)advantages of the medium camera with respect to the high and low quality cameras are rather small, whereas the (dis)advantages between the high and low quality camera are much larger. As such, adding the high quality/price camera frames the medium quality/price camera as the compromise. As is shown in Figure \[fig:04\], our explanation of the compromise effect, once again, rests on the presence of a negative relation between two alternatives.
Endowment {#endowment .unnumbered}
---------
Besides implementing bounded rationality and explaining the occurrence of well known violations of rationality, the model also gives a surprising explanation for the endowment effect[@thaler1980]. The endowment effect describes the tendency of people to value an object higher if they possess it compared to when they do not.
To illustrate the endowment effect we consider a variation on the Debreu example, where all three recordings are equally attractive. Suppose you are given one of the Beethoven recordings and are presented with the option to exchange it for the Debussy recording. While someone else is given the Debussy recording with the option to exchange it for the Beethoven recording. As both recordings are equally attractive, the choice axiom predict that you would exchange your current Beethoven recording for the Debussy about half the time. The endowment effect says however that people are unlikely to switch, a prediction that has been experimentally verified[@kahneman1990].
Explanations offered for this effect are choice-supportive bias, the finding that people tend to attribute more positive features to the object they chose compared to the object they did not choose[@mather2000], as this aligns their attitudes to be consistent with the choice and minimises cognitive dissonance. A second mechanism could be the idea of loss aversion, the tendency that people value preventing the loss of an object more than gaining an object of the same value[@kahneman1984]. In both cases this would translate to an increase in base appeal of the alternative as soon as it would come into your possession. As such, while two objects can be equally attractive before a choice is made, as soon as the choice is made the chosen object will have an increase in base attractiveness compared to the non chosen object.
With the current model we also obtain a new explanation, that does not depend on changes in the values of the choice problem, but instead ties in to the choice process itself. If we frame the choice as a choice between the Debussy of the Beethoven, having been given the Debussy makes that all conditions of the choice problem are met. That is the choice process starts in a configuration which is equivalent to one where a choice has already been made. As such there is no need to iterate at all, which would explain why people tend to stay more often with their initial choice.
Discussion {#discussion .unnumbered}
==========
This article discusses the key phenomena discovered in discrete choice behaviour, and proposes an elegant model, derived from the Ising model from statistical mechanics, with simple dynamics to explain these. The proposed model explains a number of deviations from rationality, and exhibits key theoretical principles from the field of decision making in a natural way, and allows the construction of new choice situations that can predict new types of irrational behaviour.
It is remarkable that the theoretically distinct violations of rationality, that are well established within the literature, can be explained by a simple mechanism in the choice model, namely the introduction of a (negative) relation between alternatives and the rule to stop as soon as the choice conditions hold for the first time. One might therefore argue that we are not dealing with three different effects that cause a violation of rationality, but just one mechanism that emerges in multiple situations.
While we introduced the model here in all its simplicity, it is clear that many extensions and alternations are possible. For example, it might be appropriate to introduce a non-symmetric relation between two alternatives in some situations. Also, the decision process used in this paper was chosen for illustrative purposes, naturally several other Markov processes of selecting, comparing, and discarding choice alternatives are possible, all having the appropriate invariant distribution for the choice probabilities. Depending on the particular choice, some processes may be more suitable than others.
Metropolis choice process
-------------------------
For a choice with $n$ alternatives, $2^n$ different configurations for the vector $\boldsymbol{\sigma}$ exist. We use $\boldsymbol{\sigma}_i$ to denote the $i^{th}$ of the $2^n$ possible states. Now, let $\mathbf{P}$ be a square matrix of order $2^n$, where element $P_{ij}$ contains the probability of transitioning from $\boldsymbol{\sigma}_i$ to $\boldsymbol{\sigma}_j$ in one step of the Metropolis algorithm. $P_{ii}$ contains the probability of staying in the current state, either because the proposal state equals the current state, or the proposal state is rejected.
$$\begin{aligned}
P_{ij} &= s(\boldsymbol{\sigma}_{j} \mid \boldsymbol{\sigma}_i) p(\boldsymbol{\sigma}_i \rightarrow \boldsymbol{\sigma}_j) \\
P_{ij} &= s(\boldsymbol{\sigma}_i \mid \boldsymbol{\sigma}_i) + \sum_{j \neq i} s(\boldsymbol{\sigma}_{j} \mid \boldsymbol{\sigma}_i) (1-p(\boldsymbol{\sigma}_i \rightarrow \boldsymbol{\sigma}_j))
\label{eq:04}\end{aligned}$$
From $\mathbf{P}$ we can obtain the expected rational response probabilities, contained in the stationary distribution, by taking the values of the elements of the first eigenvector of $\mathbf{P}$ that correspond to states in which the choice conditions are met (i.e., $\sigma_c = 1$ and $\sigma_+ = 2$), and dividing these by their sum. We will not go into this approach at length as these probabilities can much more simply be obtained from equation \[eq:02\].
To obtain the expected response probabilities, for when the response process stops as soon as the choice conditions $\sigma_c = 1$ and $\sigma_+ = 2$ are met for the first time, we reformulate the Markov chain with transition matrix $\mathbf{P}$ as an absorbing chain with $n$ absorbing states, i.e., those states where only one alternative is active, and $2^n - n$ transient states, i.e., those states where more than one alternative or no alternatives are active. $\mathbf{P}$ has canonical form
$$\begin{aligned}
\mathbf{P} =
\begin{bmatrix}
\mathbf{Q} & \mathbf{R} \\
\mathbf{0} & \mathbf{I}
\end{bmatrix} \, ,
\label{eq:05}\end{aligned}$$
where $\mathbf{Q}$ contains the transition probabilities from one transient state to another transient state, $\mathbf{R}$ contains the transition probabilities from the transient states to absorbing states and $\mathbf{I}$ is an identity matrix of order $n$ representing that once an absorbing state is entered it cannot be left, and $\mathbf{0}$ is the zero matrix, which represents the transition probabilities from an absorbing state to a transient state (which are all zero). Rearranging $\mathbf{P}$ in its canonical form allows us to derive the expected progression of the Markov chain more easily [@diederich2003].
Let $x \in \{1,2,\dots,n\}$ denote the alternative that is chosen. Furthermore, we use $\mathbf{z} = [z_1, z_2, \dots, z_{2^n}]$ to denote the resting state probabilities, where $z_i$ denotes the probability for the system to have configuration $\boldsymbol{\sigma}_i$ for the alternatives in the resting state distribution. That is, $\mathbf{z}$, contains the probabilities of starting the choice process in some particular configuration $(\boldsymbol{\sigma})$ of the alternatives. This vector of probabilities $\mathbf{z}$ can be divided into the part containing the probabilities of starting in an absorbing state($\mathbf{z}_a$), or transient states ($\mathbf{z}_t$). Finally, we let $s$ denote the number of iterations of the Metropolis algorithm.
We can now write the (marginal) probability that alternative $x$ is chosen as
$$\begin{aligned}
p(x) = \mathbf{z}_a \, \mathbf{I} + \mathbf{z}_t \, \sum\limits_{s = 0}^{\infty} \, \mathbf{Q}^s \, \mathbf{R}_x \,,
\label{eq:06}\end{aligned}$$
and using geometric series to rewrite the infinite sum over $\mathbf{Q}^s$ we can simplify this expression as
$$\begin{aligned}
p(\text{x}) = \mathbf{z}_a \, \mathbf{I_x} + \mathbf{z}_t \, (\mathbf{I} - \mathbf{Q})^{-1} \, \mathbf{R}_x \,.
\label{eq:07}\end{aligned}$$
As an added bonus we can derive the expected number of Metropolis iterations before alternative $x$ is chosen. While this is beyond the scope of the current paper, one can imagine using the number of iterations as a proxy for the expected (qualitative ordering of) response times. Such predictions can in turn be used to validate/compare different choice structures, or different response processes for the same choice structure. The expected number of Metropolis iterations before particular alternative is chosen is given by
$$\begin{aligned}
E(\text{s} \mid \text{x}) &= \frac{1}{p(\text{x})} \, \left[ \mathbf{z}_t \, \sum\limits_{s = 1}^{\infty} \, s \, \mathbf{Q}^{s-1} \, \mathbf{R}_x \right] \, .
\label{eq:08}\end{aligned}$$
We can again use geometric series to rewrite the infinite sum over $s \, \mathbf{Q}^{s-1}$ and simplify the expression as
$$\begin{aligned}
E(\text{s} \mid \text{x}) &= \frac{1}{p(\text{x})} \, \left[ \mathbf{z}_t \, (\mathbf{I} - \mathbf{Q})^{-2} \, \mathbf{R}_x \right] \, .
\label{eq:09}\end{aligned}$$
Parameter values
----------------
For each of the examples in Figures \[fig:02\] to \[fig:04\], we set $\beta = 1$, and $\varepsilon =0$ to make them easily comparable. Furthermore, in all examples the strength of the relationship between the cue’s and the alternatives was set to one $(\omega_{ca} = 1)$. All base appeals, and relations between alternatives were set to zero $(\mu = 0, \omega_{aa} = 0)$ unless otherwise specified. The parameters for which non-default values were used are given in Table \[tab:01\].
Data availability statement
---------------------------
Data sharing not applicable to this article as no datasets were generated or analysed during the current study.
Code availability
-----------------
Software code used for calculating the expected choice probabilities and expected number of Metropolis iterations was written for the R programming environment[@R], and is available from Github at https://github.com/JoostKruis/CIRM/. The code is dependent on the plyr (https://cran.r-project.org/package=plyr), and qgraph (https://cran.r-project.org/package=qgraph) packages. This repository also contains the setup to run this code for all examples discussed in the paper.
This research was supported by NWO (The Dutch organisation for scientific research); No. 022.005.0 (J.K.), No. CI1-12-S037 (G.M.), No. 451-17-017 (M.M.), No. 451-15-008 (D.M.), No. 314-99-107 (H.M.).
J.K. and G.M. wrote the manuscript; J.K. created the analytic model with contributions from G.M., M.M., and M.B.; H.M. gave conceptual advice. All authors discussed the results and implications and commented on the manuscript.
The authors declare that they have no competing interests.
Correspondence and requests for materials should be addressed to Joost Kruis (email: j.kruis@uva.nl).
Figure: 2 3a 3b 4
----------------------- --------------------------- ---------------------------- ---------------------------- ------------------------
$\mu_{B_F} = 3$ $\mu_{\$} = 4.25$ $\mu_{\$} = 1.75$ $\mu_{C_L} = 7.5$
$\boldsymbol{\mu}$ $\mu_{B_K} = 3$ $\mu_{P_+} = 3.675$ $\mu_{P_+} = 1.175$ $\mu_{C_M} = 7.5$
$\mu_{D_C} \approx 3.405$ $\mu_{P_-} = 0.825$ $\mu_{P_-} = -1.675$ $\mu_{C_H} = 7.5$
$\boldsymbol{\omega}$ $\omega_{B_F, B_K} = -10$ $\omega_{\$, P_+} = -0.59$ $\omega_{P_+, P_-} = 2.75$ $\mu_{C_L, C_H} = -15$
$\omega_{\$, P_-} = -2.75$
: Parameter values used in examples.
\[tab:01\]
[^1]: The corresponding parameter values $\{\boldsymbol{\mu}, \boldsymbol{\Omega}, \beta, \varepsilon\}$ are provided in the methods section, together with the computer code that allows for the calculation of the expected response probabilities. However, as these examples are for illustrative purposes, we will not discuss the exact values in the current example, nor the subsequent examples.
| 2023-08-05T01:26:57.367551 | https://example.com/article/3881 |
A Pilot Investigation of Cognitive Behavioural Therapy for Clinical Perfectionism in Obsessive Compulsive Disorder.
Perfectionism is strongly associated with obsessive compulsive disorder (OCD). Cognitive behavioural therapy for perfectionism (CBT-P) has been found to result in reductions in a range of symptoms in individuals with anxiety disorders, depression and eating disorders. To pilot-test the efficacy of group CBT for perfectionism in participants with OCD and elevated perfectionism. Participants were randomized to receive immediate 8-week group CBT-P (n = 4) or an 8-week waitlist followed by CBT-P (n = 7). Reliable reductions and a large effect size indicated that CBT-P was associated with improvements in perfectionism and OCD severity at post-test. However, these changes were not clinically significant and drop-out was high, resulting in a small final sample. CBT-P may be effective in reducing perfectionism and disorder-specific OCD symptoms. However, the high drop-out rate and lack of clinically significant findings suggest that further research needs to be conducted to determine the efficacy of CBT for perfectionism in OCD. | 2024-07-17T01:26:57.367551 | https://example.com/article/6559 |
Q:
SQL - Filter rows by attributes
I have 3 tables USERS, WORKFLOWS, ATTRIBUTES. WORKFLOWS tables values will be NEW, UPDATE and DELETE and ATTRIBUTES table contains workflow attributes like comment, doneBy, dateOfDone etc there are many. Users and Workflows table are connected by USR_ID and workflows and Attributes are connected by WF_ID.
+--------+------------+------------+ +--------+---------+------------+
|USR_ID | USR_NAME | IS_GENUINE | |WF_ID | WF_NAME | USR_ID |
+--------|------------+------------+ +--------|---------+------------+
| 1 | John | Y | | 1 | NEW | 1 |
| 2 | King | Y | | 2 | Update | 1 |
| 3 | Mark | N | | 3 | Delete | 1 |
| 4 | Smith | N | | 4 | Delete | 2 |
| 5 | Zack | Y | | 5 | Update | 2 |
+---------------------+------------+ | 6 | New | 3 |
| 7 | Update | 5 |
+------------------+------------+
+--------+--------------+----------------+-------+
|ATTR_ID | ATTR_NAME | ATTR_VALUE | WF_ID |
+--------|------------- +----------------+-------+
| 1 | comment | good | 1 |
| 2 | doneBy | suresh | 1 |
| 3 | comment | good | 2 |
| 4 | doneBy | suresh | 2 |
| 5 | comment | bad | 3 |
| 6 | doneBy | Raj | 3 |
| 7 | comment | Wow | 4 |
| 8 | doneBy | Amit | 4 |
| 9 | comment | good | 7 |
| 10 | doneBy | suresh | 7 |
+-----------------------+----------------+-------+
Now comes the major question I want to fetch the user info along with workflows and attributes if the IS_GENUINE = 'y' and comment = 'good' and doneBy = 'suresh'.
Expected Output
+--------+------------+---------+-----------+------------+
|USR_ID | USR_NAME | WF_NAME | ATTR_NAME | ATTR_VALUE |
+--------|------------+---------+-----------|------------+
| 1 | John | NEW | comment | good |
+--------|------------+---------+-----------|------------+
| 1 | John | NEW | doneBy | suresh |
+--------|------------+---------+-----------|------------+
| 1 | John | UPDATE | comment | good |
+--------|------------+---------+-----------|------------+
| 1 | John | UPDATE | doneBy | suresh |
+--------|------------+---------+-----------|------------+
| 1 | John | DELETE | comment | bad |
+--------|------------+---------+-----------|------------+
| 1 | John | DELETE | doneBy | Raj |
+--------|------------+---------+-----------|------------+
| 5 | Zack | UPDATE | comment | good |
+--------|------------+---------+-----------|------------+
| 5 | Zack | UPDATE | doneBy | suresh |
+--------|------------+---------+-----------|------------+
A:
You can use windows function count as follows:
Select USR_ID, USR_NAME, WF_NAME, ATTR_NAME, ATTR_VALUE from
(Select U.USR_ID, U.USR_NAME, W.WF_NAME, A.ATTR_NAME, A.ATTR_VALUE,
count(case when a.attr_name = 'comment' and attr_value = 'good' then 1 end) over (partition by a.wf_id) as comm_cnt,
count(case when a.attr_name = 'doneBy' and attr_value = 'suresh' then 1 end) over (partition by a.wf_id) as suresh_cnt
from
Users u join workflow w
On u.user_id = w.user_id
Join attributes a on a.wf_id = w.wf_id
Where is_genuine = 'Y')
Where comm_cnt = 1 and suresh_cnt = 1;
| 2024-01-12T01:26:57.367551 | https://example.com/article/8636 |
<test-metadata>
<benchmark-version>1.2</benchmark-version>
<category>ldapi</category>
<test-number>00021</test-number>
<vulnerability>true</vulnerability>
<cwe>90</cwe>
</test-metadata>
| 2024-07-29T01:26:57.367551 | https://example.com/article/6782 |
Electrophysiological effects of sodium channel blockers on guinea pig left atrium.
The electrophysiological effects of five sodium channel blockers (mexiletine, lidocaine, disopyramide, aprindine and flecainide) on the guinea pig left atrium were investigated by recording the action potential and its maximum rate of rise (Vmax). The onset and offset kinetics of use-dependent block of Vmax were analyzed. Lidocaine, aprindine and flecainide were classified clearly as fast, intermediate and slow, respectively. Mexiletine and disopyramide had two components in onset and offset of use-dependent block. Mexiletine showed fast and intermediate kinetics, whereas disopyramide showed intermediate and slow kinetics. Action potential duration at 90% repolarization (APD) was prolonged by disopyramide and mexiletine. The other drugs did not change the action potential duration. Effective refractory period was prolonged by all drugs with relative potency in the following order: disopyramide greater than mexiletine greater than lidocaine greater than aprindine = flecainide. In conclusion, the modes of actions of sodium channel blockers on the atrium were disclosed to be different from those on the ventricle. The pharmacological therapy for atrial arrhythmias should be based on the electrophysiological effects of the drugs on the atrium, not on the ventricle. | 2023-09-26T01:26:57.367551 | https://example.com/article/7586 |
[Evaluation of sonography in diseases of the paranasal sinuses].
The diagnosis of disorders of the paranasal sinuses by ultrasound has grown in importance during the last few years and has become easier due to developments in ultrasonic equipment. Nevertheless it seemed to be necessary to evaluate the advertising claims that the method is as reliable as radiography in the diagnosis of pathological conditions of the paranasal sinuses. We investigated 188 maxillary sinuses of adults by sinuscopy and compared the findings to those of radiography and ultrasound. The results are given and discussed. | 2023-11-21T01:26:57.367551 | https://example.com/article/6591 |
Seth Meyers’s Home
New York real estate is no joke. But, Seth Meyers is! The long-time head writer and cast member of “Saturday Night Live” just purchased a nice little 994 sq ft. condo in the West Village area along with the wifey, who he married about a few months ago. Oh alright, we’ll give you a name. Alexi Ashe, and she’s quite the hottie. But enough about her. Seth’s set to take over for Jimmy Fallon this February, while Fallon’s set to take over for the literal old-timer, Jay Leno. That’s quite a chain reaction there. Then again, Seth needs to keep working, especially now that he spent $3.525 million on that 2 bed and 2 bath condo in NYC. Good luck, Mr. Meyers. In the meantime, we’d like to see your new home.
The home on 8th avenue is nothing but a cozy and warm place. Might be due to the fireplace that can be found in the living room. Three white generous windows can also be found here, where Seth and the wife can peer through and observe the wonderful view of the neighborhood and distant cityscapes. Seth and Alexi’s home are a few blocks away from Union Square and shops and restaurants are within a foot’s reach. Seth…..Alexi…….not sure how you managed to pummel your way through all of that snow, but touche. | 2024-07-10T01:26:57.367551 | https://example.com/article/7376 |
Q:
Trying to install nacl (native client) from google but getting syntax error
C:\nacl_sdk>naclsdk list
File "C:\nacl_sdk\sdk_tools\sdk_update_main.py", line 720
print "Native Client SDK Updater, version r%s" % (REVISION,)
^
SyntaxError: invalid syntax
File "C:\nacl_sdk\sdk_tools\sdk_update_main.py", line 720
print "Native Client SDK Updater, version r%s" % (REVISION,)
^
SyntaxError: invalid syntax
C:\nacl_sdk>
This is what I am getting on the command line. I have done nothing to the unzipped folder I downloaded except moving it to the C:\ directory. Any advice, solutions or help would be appreciated.
A:
You need to use Python 2 instead of Python 3.
| 2024-05-30T01:26:57.367551 | https://example.com/article/4148 |
The glass is made in Japan. Although the final product will look toward the future, the first steps call on glassmaking techniques of centuries past.
Craftsmen fill large handmade clay pots with one and a half tons of silica and boron oxide powders. The sandy mix goes into a great furnace, where it is heated for more than two days at temperatures as high as 1,500 degrees Celsius (2,732 Fahrenheit), melting the raw material until the 800-liter pots are brimming with molten glass. Over the course of 13 days, the material cools as the crystal-clear glass hardens uniformly, forming a structure that is resistant to heat and pressure.
The clay pots are then shattered with a sledgehammer, revealing the transparent monoliths within. Each slab weighs about a ton. The Ohara E6 glass is perhaps the purest optical glass in the world, known for its resistance to thermal expansion.
A slab of E6 glass. Ohara Group
The Japanese glass is cut into smaller chunks that weigh four to five kilograms (9 to 11 lbs.) each, which are then shipped to Tucson, Arizona. Here, technicians will use the glass to cast gargantuan mirrors, each one spanning 8.4 meters in diameter—that's 27 feet from edge to edge—and weighing 17 tons. In time, seven of these segments will combine to form an enormous primary telescope mirror spanning 24.5 meters (80 feet) across.
The justification for all this globetrotting, painstaking glass work is the Giant Magellan Telescope, which will become the largest optical telescope in the world. When complete in the mid-2020s, GMT will perch on the summit of Chile's Cerro Las Campanas at 8,200 feet, in a 22-story-tall building that is already under construction. Here, in the high altitudes of the Atacama Desert, the ultra-dry air offers some of the clearest views of the night sky anywhere on the planet.
The giant telescope will have 10 times the resolving power of the Hubble Space Telescope, revealing distant galaxies, the birth of stars, and the compositions of exoplanet atmospheres—a key field of research in the search for extraterrestrial life.
Artist rendering of the Giant Magellan Telescope. GMTO
The beginning of telescope operations, known as first light, is planned for 2023 and will use just four mirrors. The first observations with all seven mirrors will follow in 2025. As for today: The pots are filled in Japan, the furnace is lit in Arizona, the ground is broken in Chile, and the Giant Magellan Telescope is under way.
Polishing Potato Chips
"We had to make the first one before they let us make the second," says Dae Wook Kim, "because they didn't think it was possible to make such a potato chip."
Kim, an assistant professor at the College of Optical Sciences at the University of Arizona, has taken me deep into the bowels of telescope technology. We're standing in the Richard F. Caris Mirror Laboratory, a big bunker beneath the university's 55,000-seat football stadium. Here, below the gridiron of the Arizona Wildcats, is where the largest telescope mirrors in the world come together.
The Mirror Lab has "possibly the only people who can make large single mirrors anymore," says Patrick McCarthy, vice president for operations of the Giant Magellan Telescope Organization (GMTO). Arizona's lab has made mirrors for some of the most powerful telescopes on the planet, including the twin 6.5-meter mirrors for the Magellan Telescopes in Chile, the 8.4-meter ones for Arizona's Large Binocular Telescope (LBT), and the 8.4-meter mirror for the Large Synoptic Survey Telescope (LSST), an ambitious project to take 1,000 pairs of exposures of the night sky with a 3.8-gigapixel digital camera, the largest in the world.
The glass for the 5th GMT mirror segment in the University of Arizona’s revolving furnace for spin-casting giant telescope mirrors, housed underneath the university football stadium. GMTO
At the moment, though, Kim is wrestling with his potato chip problem. The central mirror of the GMT will be a perfectly parabolic concave structure. The outer six segments, however, need to be shaped differently to work as a single primary mirror once aligned together. The six will circle the central mirror like flower petals, and each one needs to reflect the light it collects to the very center of the aperture, which means they must be ground to a potato chip shape.
The Mirror Lab houses a giant rotating furnace to cast the slabs of glass, equally enormous grinding and polishing machines, and a measuring tower to scan the mirrors with lasers. The whole process starts with the chunks of glass from Japan. Technicians fill the industrial merry-go-round furnace by hand. For the fifth mirror, they laid out precisely 17,481 kg (38,539 lbs.) of Ohara E6 glass.
This content is imported from YouTube. You may be able to find the same content in another format, or you may be able to find more information, at their web site.
The furnace begins to spin as the heat is cranked up to 700 C (1,292 F). The machine spins faster and grows hotter over the course of weeks, until the large rotating furnace hits 4.9 rpm and 1,165 C (2,129 F), hot enough to melt the glass blocks entirely.
The lab's secret is spin-casting mirrors in a rotating furnace over a mold to form a honeycomb structure, which makes their creations much lighter and sturdier than a mirror would be with no structural cells. The molten material pools around a mold of individual cells to create the honeycomb on the back, while a smooth surface of uninterrupted glass forms on the front.
The furnace then slows to 1/2 rpm for a three-month cooling process. The slow cooling of the glass allows it to anneal, relieving internal stresses and improving durability. Once the mirror cools and hardens, a year or perhaps longer of grinding and polishing can begin.
The rotating furnace at the Richard F. Caris Mirror Lab, loaded with Ohara E6 glass and about to start the casting process. GMTO
Despite all this telescope wizardry, the technicians at Arizona's Mirror Lab didn't know for sure if they'd be able to make something so precise as the potato chip. Grinding the shallow bowls into the saddle shape required for the outer GMT mirrors had never been attempted. "We don't get there in one shot," says Hubert M. "Buddy" Martin, an associate research professor of optical sciences at the University of Arizona.
The mirrors are measured with lasers in a 28-meter-tall (92 feet) test tower. The tower is structurally isolated from the rest of the stadium. Three meters of solid concrete form its foundation, and pneumatic isolators prevent any vibrations from affecting the precision laser interferometer.
The test tower for measuring large telescope mirrors at the Mirror Lab. GMTO
"We're pretty much immune to vibrations in the building, football stadium, traffic outside, helicopters coming into the hospital," says Martin. "If we weren't isolated from these vibrations, they would wreak havoc on the measurements."
The test tower beams lasers down onto the mirror and collects the rebounding light with a series of sensors. The system creates a contour map of the glass surface. "You can think of it as a sheet of light designed to be the perfect mirror surface," Martin says. "When the light comes back you can see irregularities."
The test tower data feeds into computers that control an industrial grinder and a polisher. The grinder does quick work using 80-grit diamond sandpaper, a noisy process that requires hearing protection. The 35-inch-thick mirror is shaved down from 20 tons to more like 17.
Polishing, however, is a more delicate process. The Mirror Lab worked for a year and a half to polish the GMT's first mirror, moving the gargantuan slab of glass back and forth between the test tower and the polishing machine some 50 times. While the grinder works with a precision of about 75 micrometers, the polishing machine refines the glassy surface to perfection within 25 nanometers, less than a millionth of an inch.
Dr. Dae Wook Kim giving a tour of the Mirror Lab with mirror segment 2 under the test tower. GMTO
"If you made this mirror the entire size of the United States, the Rocky Mountains would be a coin width up," Kim says, standing before the enormous polishing machine. A rotating arm mounted to a control rig waxes a small part of the mirror, like a robotic gearhead meticulously polishing its '63 Sting Ray.
The polisher uses off-the-shelf Silly Putty that the Mirror Lab buys in 5-pound blocks. The MacGyver trick was Kim's idea, and it's an example of the inventive solutions you need in a one-of-a-kind mirror-casting lab. Despite all their advanced gadgets, the lab technicians use fishing wire to remove pads glued to the mirrors, and the furnace is loaded by hand.
Moving GMT mirror segment 4 out of the furnace with glue-on pads. GMTO
The first GMT mirror is now complete. It sits in a climate-controlled storage container on a trailer in Tucson, waiting to begin its journey to Chile. With one mirror in the basket, the Mirror Lab has been given the green light to cast more. There are four GMT mirrors in various stages of completion at the Arizona lab. The team hopes to make the potato chips faster now that the process has been proven.
At the Mirror Lab, you'll hear the technicians joke that they don't actually make mirrors. "We just make really huge pieces of glass," an employee said. The final step of the process—which will take place at the telescope's construction site in Chile—is coating the mirror blanks with a thin layer of aluminum to actually make them reflective.
The Giant Magellan Telescope inches toward first light, when these mirrors will turn to the heavens for the first time. "We want to take a picture of the universe when it was just being born," Kim says.
The Planet Detective
GMTO
When it all comes together in the 2020s, the GMT will be able to resolve an object the size of a dime at 60 miles away, which is an order of magnitude more resolving power than Hubble. Such viewing power opens a universe of possibilities.
"Our favorite science use for GMT is looking at planets in habitable zones," says Jared Males, assistant astronomer at the University of Arizona's Steward Observatory. "What we want to do is look for biosignatures. We want to look for life on these planets."
Sophisticated spectrometers will measure the light collected by GMT to identify the compositions of celestial objects. When an exoplanet passes in front of the star it orbits from our perspective, the giant telescope's instruments will be sensitive enough to separate the light passing through the planet's atmosphere from the rest of the starlight. The light that passes through the atmosphere will have gaps in the electromagnetic spectrum—absorption lines that reveal the presence of specific molecules in that world's skies.
This content is imported from YouTube. You may be able to find the same content in another format, or you may be able to find more information, at their web site.
"A really good biosignature is diatomic oxygen," says Patrick McCarthy. "All the oxygen in Earth's atmosphere is due to life."
Photosynthetic life is the only reason molecular oxygen exists in such abundance in Earth's atmosphere, but there are also natural geologic processes that can produce significant amounts of atmospheric oxygen (though if this were the case on another planet, other clues would tell us the oxygen wasn't from a biological source). More problematic, however, is the fact that life existed on Earth for more than a billion years before photosynthesis filled our skies with oxygen, and similar anaerobic microbes on other planets could happily thrive with no O 2 to speak of.
If that life is anything like the early life on Earth, however, it will produce methane, another intriguing potential biosignature for alien worlds. In fact, a recent study in Science Advances found that a combination of methane and carbon dioxide in an exoplanet's atmosphere would be a convincing biosignature of pre-photosynthetic life.
This artist’s concept shows what the TRAPPIST-1 planetary system may look like, based on available data about the planets’ diameters, masses and distances from the host star, as of February 2018. NASA/JPL-Caltech
The GMT will turn its gaze toward nearby exoplanets, such as Proxima b orbiting the closest star to us and the seven Earth-sized planets orbiting the star TRAPPIST-1 less than 40 light-years away. The scope will also look to thousands of other planets that astronomers are discovering all across the Milky Way to search for signs of life in alien atmospheres.
"You can ask whether the sky is blue on an exoplanet," says McCarthy. An atmosphere similar to Earth's would cause blue light to scatter more than other wavelengths.
The enormous telescope, perched in the high Atacama, will also be able to detect incredibly dim objects with its vast light-collecting surface. Early star formation will be studied with unprecedented precision, and galaxies will be discovered that are farther away, older, and less bright than any previously detected. In addition to searching the galaxy for life, the GMT will allow astronomers to study the most ancient eons of the universe, probing the firmament to test our fundamental laws of physics.
A Telescope for the Future
GMT diagram. Damien Jemison/GMTO
Since the 1990s, two revolutionary technologies have enabled telescopes to measure the cosmos with more precision than ever before. The first is interferometry, which uses multiple telescope observations to stitch together a picture with the same resolution as one enormous telescope, which is useful for doing things like imaging the surfaces of other stars. The test tower at the Mirror Lab uses interferometry to create contour maps of the glass surfaces.
The second major breakthrough in telescope technology, which the GMT will use, is adaptive optics.
"The telescope is basically always getting knocked out of focus from the atmosphere," says Laird Close, professor of astronomy at the University of Arizona. "So if you could use actuators to adjust the focus 1,000 times per second, you can have a telescope that is always in focus."
Keeping the telescope in focus is exactly what the adaptive secondary mirror is designed to do. This 3.25 meter diameter mirror will also have seven circular segments to match the primary mirror segments, except the secondary mirror segments will only be 1.6 mm thick.
"It's so thin that if you picked it up with your hands, it would break," says Close. "We spend a lot of time trying not to break the mirrors."
The adaptive secondary mirror will be constantly warped by hundreds of actuators to correct for the atmospheric refraction that bends light before it reaches the surface of our planet. Sodium lasers beamed from the telescope up into the sky will serve as guide stars to calibrate the adaptive optics system, which uses algorithms to predict atmospheric changes and warp the secondary mirror in real time.
GMTO
Even in the high Atacama Desert, atmospheric interference can significantly reduce the precision of astronomy observations. Adaptive optics systems allow telescopes like GMT to achieve unprecedented accuracy, similar to a telescope orbiting in space.
Other large ground-based telescopes—such as the Thirty Meter Telescope (TMT) planned for construction on the volcano Mauna Kea in Hawaii and the Extremely Large Telescope (ELT) which will live to the north of GMT in Chile—will have even larger primary mirrors than the Giant Magellan Telescope. These two telescopes, however, will use hundreds of hexagonal mirror segments that are each about a meter and a half across (492 segments for the TMT and 798 for the EMT), rather than seven gargantuan monolithic mirrors like GMT.
As a result, more gaps will exist in the light collecting areas of the TMT and ELT. The light will need to bounce around more before it reaches the instruments, and adaptive optics algorithms will be required to stitch everything together to a greater extent than with GMT, which can use a backup secondary mirror that is rigid to conduct observations while its flexible adaptive optics system is being cleaned and maintenced.
There is also the next flagship NASA space telescope, the James Webb Space Telescope (JWST), with 18 hexagonal mirror segments measuring 2.4 meters across for a total primary mirror aperture of 6.5 meters. James Webb will be dwarfed by the GMT's 24.5 meter aperture, but out beyond the atmosphere it will have an unparalleled view of the stars. Additionally, James Webb will take observations primarily in infrared, while GMT is optimized for light in the visible spectrum.
GMT construction site as of February 2017. GMTO
"I think it will be very complimentary to Webb," says McCarthy. "The visible part of the spectrum is very rich in information."
James Webb, for example, will be particularly suited to studying the very early universe, as light that has been traveling for billions of light-years is Doppler shifted significantly into the infrared part of the spectrum. Infrared light can also be detected piercing through clouds and nebulae in space, revealing the objects lurking behind. Optical observations, like those of Hubble and GMT, are well-suited to spectroscopy and detecting the compositions of bodies such as exoplanets. The sheer size of GMT will also make it capable of detecting incredibly dim and distant objects.
Before the Giant Magellan Telescope can power on, however, there is still much work to be done. The fifth mirror is currently cooling in the furnace, and three more need to be cast—one extra outer mirror to keep the telescope at full capability as mirrors are swapped out to be cleaned and recoated with aluminum. The 17-ton mirrors in their articulated, temperature-controlled, shock-absorbing cases also need to be shipped from Arizona to Chile—by way of port in either Houston or Los Angeles.
Everything is at stake. McCarthy stresses the need to avoid a situation like the first light of Hubble, when it was discovered that the mirrors were not perfectly shaped, requiring new instrumentation to be installed to account for the incorrectly focused beams of light. And, of course, there is the possibility of losing a mirror on the voyage across the ocean and up into the mountains.
"You have all your eggs," says McCarthy. "Do you put them all in one basket, or do you put them in multiple baskets? ... Maybe I'll just go with them because if the boat goes down, it won't be my problem."
GMTO
The Giant Magellan Telescope is truly a telescope of a new era. The first large-scale observatories were established by the scientific societies of the Enlightenment with the financial backing of monarchs, such as the Royal Observatory in Greenwich and the Paris Observatory. In subsequent centuries, wealthy benefactors sponsored the construction of giant telescopes, such as James Lick who is buried underneath the concrete of his eponymous observatory on the summit of Mt. Hamilton near San Jose, California. In the modern era, taxpayers foot the bill for major astronomy projects from agencies like NSF, NASA, and the European Southern Observatory (ESO).
The GMT project represents a paradigm shift. Private interests, international partners, universities, and science institutions across the world have come together to build this great monument to science in the high Atacama. Boeing is conducting fluid dynamics research for the telescope's housing structure, and NASA's Jet Propulsion Laboratory is providing mirror insight gleaned from the mistakes made on Hubble. Other than the University of Arizona's Steward Observatory, the GMT is sponsored by the Carnegie Institution for Science in Washington D.C., the São Paulo Research Foundation of Brazil, the Korea Astronomy and Space Science Institute, Harvard, the Smithsonian Institution, Arizona State University, the McDonald Observatory of the University of Texas at Austin, Texas A&M, the University of Chicago, Astronomy Australia Limited, and the Australian National University, among other private benefactors.
"We're going to be fighting tooth and nail for time," said Males, referring to the University of Arizona's need to share the telescope's observation time with so many other institutions. But the very fact that science agencies around the world are chomping at the bit to get involved with the Giant Magellan Telescope is a testament to its capability, versatility, and the profundity that astronomers expect to encounter through its 119-ton mirror, glinting on a mountaintop in the Atacama Desert.
"We do not make this to sell for a profit," says Dae Wook Kim. "We make it for all mankind."
| 2024-02-28T01:26:57.367551 | https://example.com/article/7787 |
Was 2015 the beginning of the end for SSDs?
This site may earn affiliate commissions from the links on this page. Terms of use.
The advent of SSDs has arguably done more to transform the experience of using a computer than any other event in the past eight years. Faster GPUs and CPUs benefit the high-end users that need such horsepower, but solid state disks can breathe new performance into virtually any hardware. The earliest drives may have had performance issues, but once those were ironed out, it was clear that NAND flash’s ascension to the top of the storage market was a question of when, not if, and the “when” depended on questions of reliability and cost-per-bit — not fundamental speed. This fundamental argument has been accepted at every level, from individual PCs to high-end enterprise deployments.
That’s why it’s surprising to see ZDNet instead arguing that 2015 was the “beginning of the end” for NAND flash in the enterprise. This argument hinges on a number of different papers that were published in 2015 concerning NAND reliability, performance, and suitability for datacenter applications. We covered some of these findings when the papers were new, but will summarize the collective findings here:
Facebook and Carnegie-Mellon found that higher temperatures can negatively impact SSD reliability and that this correlates with higher bus power consumption as well. Interestingly, this study found that failure rates did not monotonically increase (only increase) with the amount of data written to NAND flash, that sparse data layouts and dense data layouts can both increase failure rates under certain conditions, and that SSDs that don’t throttle and experience high temperatures have higher failure rates.
A major Korean study on VM performance found that SSD garbage collection didn’t mesh well with existing algorithms for that purpose, leading to significant performance degradations in some cases. The paper concluded that it’s currently impossible to guarantee a set number of IOPS when multiple VMs are hosted on a single drive. While this paper used consumer hardware, the flaws it found in how garbage collection is handled would have applied to enterprise equipment as well.
A new Sandisk study found that the use of multiple layers of log-structured applications “affects sequentiality and increases write pressure to flash devices through randomization of workloads, unaligned segment sizes, and uncoordinated multi-log garbage collection. All of these effects can combine to negate the intended positive affects of using a log.”
When you put these reports together, they point to issues with SSD reliability, performance, and suitability for certain workloads. But I’m much less certain than ZDnet that this stacks up to NAND’s rapid retreat from the data center.
Teething problems vs. cataclysmic deficiencies
I strongly suspect that if we could rewind the clock to the beginning of the HDD era, we’d see similar comments made about the suitability of hard drives to replace tape. In the 1970s and early 1980s, tape was the proven technology and HDDs, particularly HDDs in consumer systems, was the upstart newcomer. It’s difficult to find comparative costs (and it’s highly segment dependent), but the March 4, 1985 issue of Computerworld suggests that tape drives were far cheaper than their HDD equivalents.
The advent of 3D NAND flash has the potential to improve NAND reliability
I don’t want to stretch this analogy too far, but I think there’s a lesson here. The pace of hardware innovation is always faster than the software that follows it; you can’t write software to take advantage of hardware that doesn’t exist yet (at least, not very well). It’s not surprising to see that it’s taken years to suss out some of the nuance of SSD use in the enterprise, and it’s also not surprising to discover that there are distinct best practices that need to be implemented in order for SSDs to perform optimally.
To cite one equivalent example — it was a 2005 paper (backed up by an amusing 2009 video) that demonstrated how shouting at hard drives could literally make them stop working. While drive OEMs were obviously aware of the need to dampen vibrations in enterprise deployments long before then, the issue bubbled up to consumer awareness in that timeframe.
Hard drives, nevertheless, continue to be sold in large numbers — even in enterprise deployments.
None of this is to suggest that NAND flash is foolproof or does not need a medium-to-long-term replacement. I’ve covered several such potential replacements just this year. It does, however, suggest that a bit more perspective is in order. It’s easy to promise huge gains on paper and extremely difficult to deliver those gains in a scaleable, cost-effective manner.
Right now, it looks as though 3D NAND adoption will drive the further evolution of SSD technology for the next 3-5 years. That, in turn, will make it more difficult for alternative technologies to find footing — a replacement storage solution will need to match the improving density of 3D NAND, or offer multiple orders of magnitude better performance in order to disrupt the NAND industry. Intel’s joint Micron venture and its 3D XPoint could disrupt the status quo when it arrives next year, but I’ll wait for benchmarks and hard data before concluding that it will.
Far from being the beginning of the end, I suspect 2015 was the end of the beginning of NAND flash, and will mark a shift towards software-level optimization and a better understanding of best practices as the technology moves deeper into data centers.
Tagged In
“Hard drives, nevertheless, continue to be sold in large numbers — even in enterprise deployments.”
Well ya. There is no large and cheaper storage medium currently. If we could get 8TB SSDs the same price, perhaps that wouldn’t be a concern as much, but also the speed isn’t ‘as’ important on these enterprise drives either. No surprise magnetic media is still widely in use for cost and storage reasons.
lord cheeseburger
“but also the speed isn’t ‘as’ important on these enterprise drives either”
Depends on the environment. Nimble, Pur, SolidFire, etc have no problem selling AFA. It’s a huge and growing market. And will continue to grow as the price of flash drops. Also as we see some big performance improvements, ie SATA Express, NVMe and the removal of the FTL.
BillBasham
I’m glad you brought a good sense of perspective to the article. I was there at the beginning of corporate hard drive use, and a hard drive failure was always a possibility at any time. Things are so much better now.
Kyle
Nice article.
I’m ‘still’ using the first SSD I ever purchased as the main OS drive of my desktop. The abuse I put that through speaks volumes (no pun intended) on the reliability of that particular drive, or my good fortune. Then again, I’ve never had an HDD fail on me either, but I always keep those under 50% filled for performance purposes.
Kary
Good points. I’d only add that it’s probably been over 15 years since I had a hard drive fail, but that doesn’t mean I haven’t replaced a drive that started making noise before it possibly failed. I’m not sure how full a drive is would affect a HDD failing due to mechanical reasons.
BTW, still using my first SSD too–I think it’s about 3 years old now.
Mayoo
I still rock an Intel 510 and an Intel 520 from when they were both new. Never had a problem. The only reason I bought a 520 is because I needed more space and didn’t wanted to go back to HDDs.
Kyle
I know keeping the used space as low as possible increases speeds, but I’ve also heard (I haven’t confirmed) that keeping a drive full tends to cause them to fail more frequently then keeping a drive partially empty.
As we learned from Blackblaze , the HDD capacity and specific model also has a very large impact on drive failure rates, so it might simply be that people use certain drives for bulk storage more than others, which could lead to this sort of rumor emerging.
lord cheeseburger
> that keeping a drive full tends to cause them to fail more frequently then keeping a drive partially empty
This would be true in theory due to the way wear leveling algorithms work.
Jason Slover
Mechanical reasons would be your drives spinning faster to access the full drive as well as arm being extended further. Thats not to say it would cause damage but the potential is there.
Jason Slover
With that being said SSD for a regular user is a huge step in performance. Can’t speak on reliablitly I have only had my SSD in use for maybe 6 months but benchmarked machine is hell of a lot faster with SSD then HDD however HDD is more affordable for much more realistate but if you don’t need the space running an OS of a SSD is incredibly faster
Maverick Guy
What’s the model of that SSD?
Mike kizaberg
Why do you lot write titles like this just to get clicks =/, you said this about hdds and they’re still around and are not going anywhere, now you’re saying the same thing about ssds, you lot should be ashamed of yourself.
Ekard
Please read the articles before posting on them. If you had read this article you would of known that Joel does not agree with Robin Harris (the author of the article that Joel is posting this article in response to) and stated he believes that SSDs are primed for software improvements (vs hardware improvements).
Though I believe the use of the term “SSD” is a bit overreaching in this case. The title should of used NAND instead.
Title: “Was 2015 the beginning of the end for SSDs?”
Ending a title with a question mark suggests that the answer = yes.
Example:
“Was Obama caught naked on the White House lawn?”
If the article is about the answer being no, then the title was clickbait designed to lure you in with the suggestion that the answer was yes. It’s shameless and people are right to call it out when they see it.
Ekard
I didn’t dispute it was clickbait and titles that are questions are normally answered as a no. This is not a new practice (started before the internet, in printed media). It is a well known and documented way to grab the readers attention (again pre-internet). Again did not dispute his labeling of the article title as clickbait.
Joel Hruska
I do. Thoroughly. I took the central assertion of the ZD headline and *questioned* it. And explained my reasoning for doing so.
There are two ways to tackle this kind of situation, where Author B is writing in response to Author A.
You can use a question mark, as I did, or you can use a strong declaration. So I could have written “No, this isn’t the beginning of the end for enterprise SSDs” or “No, enterprise SSDs aren’t being phased out.”
The reason I dislike this construction has nothing to do with clickbait. I dislike it because it drops readers in the middle of a discussion they didn’t know we were having. It’s confusing. A logical thing for a reader to do would be to ask: “Did he (or someone else at ET) recently write a story on why enterprise SSDs are being phased out?”
By rephrasing a direct statement made by the ZD author as a question, I make it clear that I am *literally* questioning his conclusion. And while this tendency is absolutely abused in other contexts “Is Obama a Muslim?” “Is the GOP racist?” “Will Mars be as close to Earth as the Moon next week?” that’s not the purpose here.
Joel Hruska
The construction methodology you refer to is used to plant doubt in the minds of the reader.
In this case, I use a question mark to signify that I am going to *engage with a topic.* I clearly identify the author of the original piece. I discuss his conclusions and my own.
In this case, because I’m engaging with a piece written by someone else, the “?” Is neither an attempt to oversell a weak conclusion nor an attempt to mislead a reader. I am questioning a statement made in headline by someone else.
Mayoo
“Facebook and Carnegie-Mellon found that higher temperatures can negatively impactSSD reliability”
SSD don’t heat … At least noting close to “high temperatures”.
Joel Hruska
Read the report. SSDs are not the only things in data centers that generate heat, and operating temperatures can exceed 40C.
Ah! If that’s the case, though, I might consider that a failure of sorts. Had single-threaded scaling continued its old course, we probably wouldn’t have needed dual cores.
Joel Hruska
I don’t think that’s true. Plenty of people loved having dual-cores — hell, I loved having dual-cores — precisely because many of us did basic tasks that could benefit.
The great advantage of Hyper-Threading on the P4, even when it *didn’t* give the chip a performance advantage over AMD’s K8, was that it made the entire system feel smoother. The ability to keep executing one thread and have another available for GUI updates or simple application launches made the entire system “creamy smooth” to quote Scott Wasson of TR.
I still remember testing a quad-core / dual-socket AMD Opteron rig with dual graphics cards. They slapped down two Nvidia chipsets on a Tyan motherboard and did the whole thing that way. Fun times.
Sheldon Cooper
Shit, I like having dual cores.
Ekard
How can it be the beginning of the end if we do not have a replacement technology available for consumers? I understand Joel is not agreeing with ZDNet but I am suspect of any story that doesn’t show the replacement for whatever is supposedly going away. Yes the original article list some technologies that “could” replace NAND but none of them are on the market. Also, why SSD? Shouldn’t the articles be about NAND and not SSD’s? There is nothing stopping companies like Intel and Micron from selling SSDs based on Xpoint rather than NAND (the whole reason NVME was created to not be based upon a technology).
Joel Hruska
Ekard,
In theory, you can replace enterprise with a tech capable of doing the job in that area, before waterfalling it to consumers. In practice, things tend to roll out simultaneously.
Ekard
My statement still stands, nothing is poised to take the place of NAND in datacenters. If performance is not the primary metric for you wouldn’t be using a NAND based storage solution in the enterprise space anyways. NAND and HDD are you main storage systems (outside archival), the former being performance and the latter being priceGB. As far as I know (which is not much when it comes to enterprise storage) nothing else is available on the market yet, even for enterprise, to replace performance solutions that are leveraging NAND.
Basically, if there is not a technology available on the market now that can fill the performance role that NAND fills, NAND based SSDs are not going anywhere anytime soon. NAND solutions fill a niche very well, more performance than HDDs and cheaper than RAM per GB.
jdwii
Couldn’t they just cool the SSD’s?
PtolemyWasWrong
Basically, SSDs suffer from TMH – Too Much Housekeeping to be useful AND reliable long term, or for typical workloads. Get them close to byte-level addressing, and things should improve markedly.
lord cheeseburger
This is demonstrably false.
High End – All flash arrays from several vendors (Nimble, Pur, etc) and every major SAN manufacturer includes flash (Hitachi, 3Par, HP, EMC, NetApp).
Low End – Every manufacturer sells SSDs in every computer variant, from desktops to high end workstations and every size and shape of laptop.
Consumer level drives have warranties lasting 5-10 years. There is absolutely no basis for the claim that they’re “unreliable long term” for “typical workloads” – whatever that means, typical where, for whom? In the datacenter? On the desktop? What are you talking about? Doesn’t matter, you’re wrong.
PtolemyWasWrong
1) Until recently, many SSDs experienced a huge decline in speed once all addressable clusters were written to once. This is crap.
2) Try pulling the power from a machine with an SSD, and see if the SSD dies. If you’re afraid to do that, they are crap.
3) Folks using SSDs to record months worth of high-frame video capture can tell you about their frustrations.
Basically, SSDs are like a new relationship – after 10 weeks of intense usage, you’re ready to buy a new one to get the speed you had at the beginning.
Jeremy Morales
I don’t own a SSD on my personal computer. The GB/$ isn’t quite there for me yet. Work computer has one though and I do enjoy it.
Basically, flash memory technology is unreliable and can fail in a multitude of different ways.
Creating reliability from unreliable components is not easy.
Crashbanksbuysilver .
I have zero problems with mine..
Reginald Peebottom
Not exactly on point but I’m curious if other readers have used whole disc encryption on SSDs like Veracrypt or the late Truecrypt? I have but it seemed to kill performance far more noticeably than it ever did on mechanical HDs – or that was my seat of the pants impression.
Has anyone else noticed that and if so is there an explanation as to why?
Joel Hruska
I have not tested it myself. Linux or Windows? Was any other form of encryption also enabled? (Bootlocker, for example).
Mangap
I hope 3D Xpoint memory will be released soon to market
benmyers
Back when SSDs were awfully expensive, I outfitted a pair of new Thinkpad T410s with the largest-at-the-time 512GB SSDs. Those Thinkpads, at the hands of the owners of a software company, are still going strong after 4 years, and they have come back to me for more SSD-equipped laptops for each of their employees. I have similarly equipped old Thinkpad X200s with smaller SSDs, creating better-than-Chromebook ultra-fast long-on-battery-life small laptops. SSDs rock!
WTF
From personal experience I wont be using them again as I’ve had 2 out of 3 SSD’s trashed for inexplicable reasons although one was turned into a brick by a Windows 10 upgrade. I don’t find them significantly faster than a normal HDD and their reliability sucks and cost per gigabyte is still too much..
DrWattsOn
@WTF: a suggestion.
I’ve done this and it NEVER has failed, tho had to use old-BIOS machines and wait like for-e v e r! USE SPINRITE. I have used it on SSDs at LEVEL2 (only once higher), though if it (the ssd) is done it’s done, so then do level4, which I had to do once, and they somehow mystically (I say, because I don’t get how) … they repair themselves.
I have maintained a Kingston 64GB system drive for since I bought it. It took a powerline spike, could not boot, ran Spinrite at level 4, it worked, and run at level 2 every 6 months or so. No further problems. So, Scoffers, do without.
I have suggested to everyone I know that had SSD failure (including at Enterprise level) and ALWAYS= SUCCESS.
So, if someone does this and it doesn’t work? Well, if
your car drives off the bridge because it got a flat, when you pull it out and fix the flat of course the car is NFG. SO WHAT? If you had maintained the tires, then … And, EVERY ANOLOGY FAILS; a perfect approximation is the thing itself.
Now upgraded to a Sam 850Pro 512GB. And backing up to HDD. “Spinning Rust”. Well, EVERY HDD I have used has had a fan bolted to its bottom, and none has failed. That’s about 10 of ’em still in use for over 10 years. Tho I knew to SpinRite ’em regularly, so cooling or SpinRite or both, they all work, though kinda small by today’s standards (80-320GB).
Jhollman
Im both a home user (Gamer) and a IT user, i can say i wont be buying an SSD in the near future because if the SSD’s performace degrades exponentially with the amount of data loaded then it is useless, HDs do not do that, so when i have to choose between buying an overpriced SSD which wont last me much or a safe and bullet-proof HD, yeah! i choose the HD. I still think SSDs have a future, if and only if they can fix all their problems.
This site may earn affiliate commissions from the links on this page. Terms of use.
ExtremeTech Newsletter
Subscribe Today to get the latest ExtremeTech news delivered right to your inbox. | 2023-09-05T01:26:57.367551 | https://example.com/article/5965 |
Who died in the Holocaust? Recovering names from official records.
The names of most Holocaust victims from Central and East Europe have remained unknown, but can be retrieved from official records, as illustrated here for the town of Liepāja, Latvia. Drawing on thirteen different sources, the authors have recovered the names and fates of about 7,000 of the 7,140 Jews once living there. The main source (5,700 names) is a census conducted during August 1941, the second month of the German occupation. Other sources are victims' lists compiled by Yad Vashem and by the Soviet Extraordinary Commission, house books, police and camp records, telephone and business directories, lists of deportees to the USSR, survivor reports, etc. All sources are incomplete and many are error-ridden, but they usually can be reconciled. Coverage of victims from Liepāja thus has increased from 20% to over 95%. For most victims, accurate data are now available on birth dates and places, prewar and wartime addresses, occupations, etc., permitting analysis of demography, survival patterns, and other trends. | 2024-02-03T01:26:57.367551 | https://example.com/article/6800 |
Q:
How do I set up polymorphic relationships using ember-data
So lets say a recipe has several ingredients of differing amounts.
Recipe Model
var Recipe = DS.Model.extend({
name: DS.attr('string'),
ingredients: DS.hasMany('ingredient')
});
Ingredient Model
var Ingredient = DS.Model.extend({
name: DS.attr('string'),
recipes: DS.hasMany('recipe'),
// amount?
});
So the amount of each ingredient would depend on the recipe. However on its own the ingredient will not have an amount.
How would you go about modeling that data? Right now I am using the FixtureAdapter until I finish building the interface.
Using Ember 1.5.1 and Ember-Data 1.0.0-beta.7+canary.b45e23ba.
A:
To answer your first question:
Define the model like so
App.Comment = DS.Model.extend({
message: DS.belongsTo('message', {
polymorphic: true
})
});
And the property needs an additional property propertyType, defining the relationship type
{
"message": 12,
"messageType": "post"
}
https://github.com/emberjs/data/blob/master/TRANSITION.md#polymorphic-relationships
Now to your second question, not sure if polymorphism would be necessary. I might just include a joining record
var RecipeIngredient = DS.Model.extend({
amount: DS.attr(),
ingredient: DS.belongsTo('ingredient')
});
var Recipe = DS.Model.extend({
name: DS.attr('string'),
ingredients: DS.hasMany('recipeIngredient')
});
| 2023-12-02T01:26:57.367551 | https://example.com/article/7578 |
Ras Khamis
Ras Khamis is a neighborhood in East Jerusalem. In 2007, construction of the Israeli West Bank Barrier cut off Ras Khamis from the rest of East Jerusalem, with the barrier completed in 2010. Presently, access to the rest of the city is via a single checkpoint. Subsequently, in 2012, municipal services such as urban planning, utilities, police and fire services, and garbage were discontinued in Ras Khamis and other neighborhoods that were separated by the barrier from the rest of the city. However, as the land was still claimed by Israel, the Palestinian Authority did not provide government services either.
Due to the housing shortage in the region, many high rise apartment buildings have been constructed in Ras Khamis, which as of 2014 is estimated to have a population of 20,000. The buildings are not approved by city planning and many have been threatened with demolition.
See also
East Jerusalem
Kowloon Walled City, a similar neighborhood in Hong Kong that existed from 1947 to 1992.
References
Category:Neighbourhoods of Jerusalem | 2023-11-29T01:26:57.367551 | https://example.com/article/6153 |
The present invention generally pertains to code sequence tracking in Code Division Multiple Access (CDMA) communication systems, also known as spread-spectrum communication systems. More particularly, the present invention pertains to a system and method for efficient tracking of multipath signal components allowing for combining of multipath signal components to improve data signal detection and despreading by reducing effects of multipath interference, and increase CDMA communication system efficiency by reducing the required transmit power.
Providing quality telecommunication services to user groups which are classified as remote, such as rural telephone systems and telephone systems in underdeveloped countries, has proved to be a challenge over recent years. The past needs created by these services have been partially satisfied by wireless radio services, such as fixed or mobile frequency division multiplex (FDM), frequency division multiple access (FDMA), time division multiplex (TDM), time division multiple access (TDMA) systems, combination frequency and time division systems (FD/TDMA), and other land mobile radio systems. Usually, these remote services are faced with more potential users than can be supported simultaneously by their frequency or spectral bandwidth capacity.
Recognizing these limitations, recent advances in wireless communications have used spread spectrum modulation techniques to provide simultaneous communication by multiple users. Spread spectrum modulation refers to modulating a information signal with a spreading code signal; the spreading code signal being generated by a code generator where the period, Tc, of the spreading code is substantially less than the period of the information data bit or symbol signal. The code may modulate the carrier frequency upon which the information has been sent, called frequency-hopped spreading, or may directly modulate the signal by multiplying the spreading code with the information data signal, called direct-sequence spreading (DS). Spread-spectrum modulation produces a signal with bandwidth substantially greater than that required to transmit the information signal. The original information is recovered by synchronously demodulating and despreading of the signal at the receiver. The synchronous demodulator uses a reference signal to synchronize the despreading circuits to the input spread-spectrum modulated signal to recover the carrier and information signals. The reference signal may be a spreading code which is not modulated by an information signal. Such use of a synchronous spread-spectrum modulation and demodulation for wireless communication is described in U.S. Pat. No. 5,228,056 entitled SYNCHRONOUS SPREAD-SPECTRUM COMMUNICATIONS SYSTEM AND METHOD by Donald L. Schilling, which is incorporated herein by reference.
One area in which spread-spectrum techniques are used is in the field of mobile cellular communications to provide personal communication services (PCS). Such systems desirably support large numbers of users, control Doppler shift and fade, and provide high speed digital data signals with low bit error rates. These systems employ a family of orthogonal or quasi-orthogonal spreading codes, with a pilot spreading code sequence that is synchronized to the family of codes. Each user is assigned one of the spreading codes from the family as a spreading function. Related problems of such a system include handling multipath fading effects. Solutions to such problems include diversity combining of multipath signals. The problems associated with spread spectrum communications, and methods to increase capacity of a multiple access, spread-spectrum system are described in U.S. Pat. No. 4,901,307 entitled SPREAD SPECTRUM MULTIPLE ACCESS COMMUNICATION SYSTEM USING SATELLITE OR TERRESTRIAL REPEATERS by Gilhousen et al. which is incorporated herein by reference.
The problems associated with the prior art systems focus around reliable reception and synchronization of the receiver despreading circuits to the received signal. The presence of multipath fading introduces a particular problem with spread spectrum receivers in that a receiver must somehow track the multipath components to maintain code-phase lock of the receiver""s despreading means with the input signal. Prior art receivers generally track only one or two of the multipath signals, but this method may not be satisfactory because the combined group of low-power multipath signal components may actually contain far more power than the one or two strongest multipath components. The prior art receivers track and combine only the strongest components to maintain a predetermined Bit Error Rate (BER) of the receiver. Such a receiver is described, for example, in U.S. Pat. No. 5,109,390 entitled DIVERSITY RECEIVER IN A CDMA CELLULAR TELEPHONE SYSTEM by Gilhousen et al. which is incorporated herein by reference. A receiver that combines all multipath components, however, is able to maintain the desired BER with a signal power that is lower than that of prior art systems because more signal power is available to the receiver. Consequently, there is a need for a spread spectrum communication system employing a receiver that tracks substantially all of the multipath signal components, so that substantially all multipath signals may be combined in the receiver. This would reduce the required transmit power of the signal for a given BER.
The present invention is embodied in a multiple access, spread-spectrum communication tracking system which processes a plurality of multipath signal components of a code-division-multiplexed (CDM) signal received over a radio frequency (RF) channel. The system and method tracks a centroid of a group of multipath spread-spectrum signal components constituting a spread-spectrum channel signal which includes a transmitted code sequence. The exemplary system and method operate by digitally sampling the spread-spectrum channel signal to produce a sequence of sample values. The sample values are divided into a set of even-numbered sample values which define a sequence of early spread-spectrum channel signal samples corresponding to the early multipath signal components and a set of odd sample number values which define a sequence of late spread-spectrum channel signal samples corresponding to the late multipath signal components
The centroid tracking receiver generates a plurality of local code sequences, each of which has a code phase and symbol period, and each of which is a code phase-shifted version of the transmitted code sequence. The centroid tracking receiver correlates each of the locally generated code sequences with the sequence of early received spread-spectrum channel signal samples to produce a group of early despread multipath signals. The tracking receiver also correlates each of the locally generated code sequences with the sequence of late received spread-spectrum channel signal samples to produce a group of late despread multipath signals. The group of early despread multipath signals are weighted according; to a predetermined algorithm and processed to produce an early tracking value, and the group of late despread multipath signals are similarly weighted and processed to produce a late tracking value.
The difference between the early tracking value and the late tracking value is calculated to produce an error signal value. Finally, the centroid tracking system adjusts the code phase of each of the locally generated code sequences responsive to the error signal value to maintain the maximum received signal energy. | 2024-01-16T01:26:57.367551 | https://example.com/article/3743 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.