aid stringlengths 9 15 | mid stringlengths 7 10 | abstract stringlengths 78 2.56k | related_work stringlengths 92 1.77k | ref_abstract dict |
|---|---|---|---|---|
1907.08736 | 2962812134 | Most of privacy protection studies for textual data focus on removing explicit sensitive identifiers. However, personal writing style, as a strong indicator of the authorship, is often neglected. Recent studies on writing style anonymization can only output numeric vectors which are difficult for the recipients to inte... | Writing Style obfuscation studies try to hide the identity of the author. Anonymouth @cite_40 is a tool that utilizes to generate writing attributes. It gives users suggestions on which way they are able to anonymize their text according to two reference dataset. also propose a similar architecture to anonymize text, h... | {
"cite_N": [
"@cite_40"
],
"mid": [
"160636586"
],
"abstract": [
"This paper presents Anonymouth, a novel framework for anonymizing writing style. Without accounting for style, anonymous authors risk identification. This framework is necessary to provide a tool for testing the consistency of anon... |
1907.08895 | 2963587483 | Convolutional Neural Network (CNN) based image segmentation has made great progress in recent years. However, video object segmentation remains a challenging task due to its high computational complexity. Most of the previous methods employ a two-stream CNN framework to handle spatial and motion features separately. In... | Convolutional neural networks have been demonstrated to achieve excellent results in video action understanding @cite_10 @cite_1 @cite_13 . Video should not be treated as a set of independent frames, since the connection between frames provides extra temporal information for understanding. Simonyan al @cite_43 propose ... | {
"cite_N": [
"@cite_13",
"@cite_36",
"@cite_1",
"@cite_44",
"@cite_43",
"@cite_0",
"@cite_47",
"@cite_10",
"@cite_11"
],
"mid": [
"",
"2883429621",
"",
"2963524571",
"2156303437",
"2963155035",
"1522734439",
"",
"2962934715"
],
"abstract... |
1907.08895 | 2963587483 | Convolutional Neural Network (CNN) based image segmentation has made great progress in recent years. However, video object segmentation remains a challenging task due to its high computational complexity. Most of the previous methods employ a two-stream CNN framework to handle spatial and motion features separately. In... | The success of CNN-based approaches for image classification @cite_2 @cite_22 have led to dramatic advances in image segmentation @cite_37 @cite_24 . Many of the segmentation approaches leverage recognition models trained on ImageNet and replace the fully-connected layers with 1 @math 1 kernel convolutions to generate ... | {
"cite_N": [
"@cite_37",
"@cite_22",
"@cite_7",
"@cite_48",
"@cite_21",
"@cite_32",
"@cite_24",
"@cite_2",
"@cite_16",
"@cite_25"
],
"mid": [
"1903029394",
"",
"",
"1901129140",
"2963840672",
"2560023338",
"",
"2016053056",
"2963881378",... |
1907.08895 | 2963587483 | Convolutional Neural Network (CNN) based image segmentation has made great progress in recent years. However, video object segmentation remains a challenging task due to its high computational complexity. Most of the previous methods employ a two-stream CNN framework to handle spatial and motion features separately. In... | Video object segmentation @cite_5 @cite_26 aims to delineate the foreground object(s) from the background in each frame. Semi-supervised segmentation pipelines @cite_15 @cite_42 assume that the segmentation mask of the first frame in the sequence during testing is given, and exploit temporal consistency in video sequen... | {
"cite_N": [
"@cite_5",
"@cite_15",
"@cite_26",
"@cite_42"
],
"mid": [
"2137981002",
"2964343881",
"",
"2963253279"
],
"abstract": [
"Real-world actions occur often in crowded, dynamic environments. This poses a difficult challenge for current approaches to video event det... |
1907.08895 | 2963587483 | Convolutional Neural Network (CNN) based image segmentation has made great progress in recent years. However, video object segmentation remains a challenging task due to its high computational complexity. Most of the previous methods employ a two-stream CNN framework to handle spatial and motion features separately. In... | In the more challenging unsupervised setting, as we address in this paper, no object mask is provided as initialization during the test phase. Unsupervised segmentation has been addressed by several variants of CNN-based models, such as the two-stream architecture @cite_14 @cite_29 , recurrent neural networks @cite_40 ... | {
"cite_N": [
"@cite_30",
"@cite_14",
"@cite_4",
"@cite_28",
"@cite_29",
"@cite_3",
"@cite_19",
"@cite_40",
"@cite_34"
],
"mid": [
"",
"2564998703",
"2566030665",
"2096979710",
"",
"2963983744",
"",
"2895340898",
"2610147486"
],
"abstract... |
1907.08895 | 2963587483 | Convolutional Neural Network (CNN) based image segmentation has made great progress in recent years. However, video object segmentation remains a challenging task due to its high computational complexity. Most of the previous methods employ a two-stream CNN framework to handle spatial and motion features separately. In... | Action segmentation provides pixel-level localization for actions ( action segmentation maps), which are more accurate than bounding boxes for action localization. Lu al @cite_35 propose supervoxel hierarchy to enforce the consistency of the human segmentation in video. Gavrilyuk al @cite_17 infer pixel-level segmentat... | {
"cite_N": [
"@cite_35",
"@cite_17"
],
"mid": [
"1912148408",
"2963354481"
],
"abstract": [
"Detailed analysis of human action, such as action classification, detection and localization has received increasing attention from the community; datasets like JHMDB have made it plausible to con... |
1907.08845 | 2962687069 | Predicting future frames in natural video sequences is a new challenge that is receiving increasing attention in the computer vision community. However, existing models suffer from severe loss of temporal information when the predicted sequence is long. Compared to previous methods focusing on generating more realistic... | The task of video frame prediction has received growing attention in the computer vision community. Early work investigates object motion prediction @cite_28 . Advanced neural network approaches were then applied to directly predict future frames @cite_32 @cite_1 @cite_11 @cite_10 . Mathieu al @cite_32 proposed a multi... | {
"cite_N": [
"@cite_28",
"@cite_1",
"@cite_32",
"@cite_2",
"@cite_31",
"@cite_10",
"@cite_11"
],
"mid": [
"2056120433",
"2401640538",
"2248556341",
"2765363933",
"2796303840",
"2738136547",
"2520707650"
],
"abstract": [
"In this paper we present a c... |
1907.08845 | 2962687069 | Predicting future frames in natural video sequences is a new challenge that is receiving increasing attention in the computer vision community. However, existing models suffer from severe loss of temporal information when the predicted sequence is long. Compared to previous methods focusing on generating more realistic... | Several works utilise shuffle based self-supervised learning methods on videos, which do not require external annotations @cite_40 @cite_5 @cite_16 . In @cite_40 , based on ordinal supervision provided by visual tracking, Wang and Gupta designed a Siamese-triplet network with a ranking loss function to learn the visual... | {
"cite_N": [
"@cite_5",
"@cite_40",
"@cite_16"
],
"mid": [
"2487442924",
"219040644",
"2964037671"
],
"abstract": [
"In this paper, we present an approach for learning a visual representation from the raw spatiotemporal signals in videos. Our representation is learned without supe... |
1907.08941 | 2962880793 | This paper conducts research on the short-term electric load forecast method under the background of big data. It builds a new electric load forecast model based on Deep Auto-Encoder Networks (DAENs), which takes into account multidimensional load-related data sets including historical load value, temperature, day type... | Auto-Encoder networks have been widely explored in the past years. Reference @cite_4 introduces a new method to analyze the human immunodeficiency virus using a combination of Auto-Encoder networks and genetic algorithms, which outperforms the conventional feedforward neural network models and is a much better classifi... | {
"cite_N": [
"@cite_30",
"@cite_4",
"@cite_8",
"@cite_19",
"@cite_16"
],
"mid": [
"2153750031",
"2237036",
"2025768430",
"2022163189",
"2517380078"
],
"abstract": [
"HMM-TTS synthesis is a popular approach toward flexible, low-footprint, data driven systems that pr... |
1907.08873 | 2963235411 | Cyberbullying and cyberaggression are increasingly worrisome phenomena affecting people across all demographics. More than half of young social media users worldwide have been exposed to such prolonged and or coordinated digital harassment. Victims can experience a wide range of emotions, with negative consequences suc... | Cyberbullying in Social Media. Focusing more on cyberbullying behaviors, @cite_67 initially decompose such a phenomenon into a set of sensitive topics, i.e., race, culture, sexuality, and intelligence. Then, they analyze YouTube comments from controversial videos based on a bag-of-words-driven text classification. Also... | {
"cite_N": [
"@cite_67",
"@cite_41",
"@cite_70",
"@cite_50",
"@cite_31",
"@cite_20"
],
"mid": [
"2209227144",
"",
"2216854803",
"1823790170",
"",
"2283668614"
],
"abstract": [
"The scourge of cyberbullying has assumed alarming proportions with an ever-incre... |
1907.08873 | 2963235411 | Cyberbullying and cyberaggression are increasingly worrisome phenomena affecting people across all demographics. More than half of young social media users worldwide have been exposed to such prolonged and or coordinated digital harassment. Victims can experience a wide range of emotions, with negative consequences suc... | Abusive Incidents in Game Communities. The rise of cyberbullying and abusive incidents, in general, is also evident in online game communities. Since these communities are widely used by people of all ages, such a phenomenon has attracted the interest of the research community. For instance, @cite_61 studies cyberbully... | {
"cite_N": [
"@cite_61",
"@cite_13"
],
"mid": [
"2078178105",
"2130529515"
],
"abstract": [
"In this work we explore cyberbullying and other toxic behavior in team competition online games. Using a dataset of over 10 million player reports on 1.46 million toxic players along with correspo... |
1907.08873 | 2963235411 | Cyberbullying and cyberaggression are increasingly worrisome phenomena affecting people across all demographics. More than half of young social media users worldwide have been exposed to such prolonged and or coordinated digital harassment. Victims can experience a wide range of emotions, with negative consequences suc... | Abusive Detection Methods. Various supervised approaches have been used for monitoring different instances of online abusive behaviors. For instance, authors in @cite_74 use a regression model, whereas @cite_50 @cite_67 @cite_20 rely on other methods like Naive Bayes, Support Vector Machines (SVM), and Decision Trees (... | {
"cite_N": [
"@cite_67",
"@cite_70",
"@cite_94",
"@cite_50",
"@cite_74",
"@cite_20"
],
"mid": [
"2209227144",
"2216854803",
"2949557692",
"1823790170",
"2340954483",
"2283668614"
],
"abstract": [
"The scourge of cyberbullying has assumed alarming proportion... |
1907.08873 | 2963235411 | Cyberbullying and cyberaggression are increasingly worrisome phenomena affecting people across all demographics. More than half of young social media users worldwide have been exposed to such prolonged and or coordinated digital harassment. Victims can experience a wide range of emotions, with negative consequences suc... | Remarks. This article presents in a unified way and, more importantly, extends our previous work on aggressive behavior in Twitter, published in @cite_47 @cite_60 @cite_64 . Compared to the literature, we advance the state-of-the-art on cyberbullying and cyberaggression analysis and detection along the following dimens... | {
"cite_N": [
"@cite_47",
"@cite_64",
"@cite_60"
],
"mid": [
"2592393350",
"2594902547",
"2612949191"
],
"abstract": [
"Over the past few years, online aggression and abusive behaviors have occurred in many different forms and on a variety of platforms. In extreme cases, these inci... |
1907.08985 | 2962953210 | Real-time Deep Neural Network (DNN) inference with low-latency requirement has become increasingly important for numerous applications in both cloud computing (e.g., Apple's Siri) and edge computing (e.g., Google Waymo's driverless car). FPGA-based DNN accelerators have demonstrated both superior flexibility and perfor... | The development of FPGA-based DNNs accelerator evolves in three stages. At the early stage @cite_4 @cite_30 @cite_19 @cite_5 @cite_21 @cite_33 , the whole FPGA is designed as one accelerator, and a controller iteratively moves data from off-chip DRAM to the accelerator to be executed. In the second stage, it is observe... | {
"cite_N": [
"@cite_30",
"@cite_4",
"@cite_33",
"@cite_21",
"@cite_24",
"@cite_19",
"@cite_5",
"@cite_25",
"@cite_12"
],
"mid": [
"2094756095",
"2474119684",
"2963948425",
"2799912737",
"2518660313",
"2513568085",
"2796625795",
"2466675884",
... |
1907.08985 | 2962953210 | Real-time Deep Neural Network (DNN) inference with low-latency requirement has become increasingly important for numerous applications in both cloud computing (e.g., Apple's Siri) and edge computing (e.g., Google Waymo's driverless car). FPGA-based DNN accelerators have demonstrated both superior flexibility and perfor... | Most recently, with the growing demand in time performance, it is a trend to employ a cluster of FPGAs to execute DNNs @cite_28 @cite_20 @cite_6 @cite_15 @cite_26 @cite_17 @cite_9 @cite_7 . In @cite_28 @cite_15 , authors construct multiple FPGAs as a pipeline to execute a set of input images in a pipeline fashion. In @... | {
"cite_N": [
"@cite_26",
"@cite_7",
"@cite_28",
"@cite_9",
"@cite_6",
"@cite_15",
"@cite_20",
"@cite_17"
],
"mid": [
"2944950984",
"2957020430",
"2475840367",
"",
"2890068895",
"2945306514",
"",
"2916975147"
],
"abstract": [
"Three-dimension... |
1907.08985 | 2962953210 | Real-time Deep Neural Network (DNN) inference with low-latency requirement has become increasingly important for numerous applications in both cloud computing (e.g., Apple's Siri) and edge computing (e.g., Google Waymo's driverless car). FPGA-based DNN accelerators have demonstrated both superior flexibility and perfor... | To satisfy the low latency requirement for real-time DNN inference, Microsoft in Brainwave @cite_10 @cite_34 devise techniques to pin weights on different FPGAs. Such an approach can work well for RNNs with small intermediate data, but awkward for CNN implementations due to the large intermediate data and complicated d... | {
"cite_N": [
"@cite_34",
"@cite_10"
],
"mid": [
"2883929540",
"2798956872"
],
"abstract": [
"Interactive AI-powered services require low-latency evaluation of deep neural network (DNN) models—aka \"\"real-time AI\"\". The growing demand for computationally expensive, state-of-the-art DNNs... |
1907.08985 | 2962953210 | Real-time Deep Neural Network (DNN) inference with low-latency requirement has become increasingly important for numerous applications in both cloud computing (e.g., Apple's Siri) and edge computing (e.g., Google Waymo's driverless car). FPGA-based DNN accelerators have demonstrated both superior flexibility and perfor... | Another branch of related work is to deploy CNNs on multi-core mobile devices or multi-processor system on-chip (MPSoC) @cite_31 @cite_27 @cite_13 @cite_29 @cite_3 @cite_14 . Unlike FPGA-based implementation that requires designers to determine the designs of communication and computation sub-systems, processing elemen... | {
"cite_N": [
"@cite_14",
"@cite_29",
"@cite_3",
"@cite_27",
"@cite_31",
"@cite_13"
],
"mid": [
"2944864513",
"",
"2889354425",
"2963547613",
"2915364523",
"2886457641"
],
"abstract": [
"",
"",
"Deep convolutional neural networks (CNNs) are widely ad... |
1907.08938 | 2964314696 | An @math maximum distance separable (MDS) code has optimal repair access if the minimum number of symbols accessed from @math surviving nodes is achieved, where @math . Existing results show that the sub-packetization @math of an @math high code rate (i.e., @math ) MDS code with optimal repair access is at least @math ... | Many constructions of MSR codes @cite_11 @cite_18 @cite_5 @cite_25 @cite_32 @cite_15 @cite_28 @cite_3 have been proposed in the literature. For example, product-matrix MSR codes @cite_11 support the parameters that satisfy @math , and are subsequently extended with lower computational complexity @cite_18 . Another cons... | {
"cite_N": [
"@cite_18",
"@cite_28",
"@cite_32",
"@cite_3",
"@cite_5",
"@cite_15",
"@cite_25",
"@cite_11"
],
"mid": [
"2338815034",
"2963754880",
"2832241482",
"2963781977",
"2126295689",
"2606114015",
"2069026918",
"2150777202"
],
"abstract": [... |
1907.08938 | 2964314696 | An @math maximum distance separable (MDS) code has optimal repair access if the minimum number of symbols accessed from @math surviving nodes is achieved, where @math . Existing results show that the sub-packetization @math of an @math high code rate (i.e., @math ) MDS code with optimal repair access is at least @math ... | MSR codes with high code rates (i.e, @math ) are important in practice. Some existing constructions of high-code-rate MSR codes are found in @cite_25 @cite_32 @cite_15 @cite_28 @cite_3 . It is shown in @cite_22 that a tight lower bound of the sub-packetization level of high-code-rate MSR codes with optimal repair acces... | {
"cite_N": [
"@cite_22",
"@cite_28",
"@cite_32",
"@cite_3",
"@cite_15",
"@cite_25"
],
"mid": [
"2765728635",
"2963754880",
"2832241482",
"2963781977",
"2606114015",
"2069026918"
],
"abstract": [
"The first focus of the present paper, is on lower bounds on t... |
1907.08938 | 2964314696 | An @math maximum distance separable (MDS) code has optimal repair access if the minimum number of symbols accessed from @math surviving nodes is achieved, where @math . Existing results show that the sub-packetization @math of an @math high code rate (i.e., @math ) MDS code with optimal repair access is at least @math ... | There are other practical concerns in distributed storage systems, such as how to mitigate the computational complexity. Binary MDS array codes are a special class of MDS codes that have low computational complexity, since the encoding and decoding procedures only involve XOR operations. Typical examples of binary MDS ... | {
"cite_N": [
"@cite_30",
"@cite_14",
"@cite_26",
"@cite_33",
"@cite_8",
"@cite_9",
"@cite_29",
"@cite_6",
"@cite_24",
"@cite_19",
"@cite_0",
"@cite_23",
"@cite_10",
"@cite_12"
],
"mid": [
"2963502306",
"2794738250",
"",
"1622719107",
"21... |
1907.08915 | 2963092220 | We propose a method for automatic segmentation of individual muscles from a clinical CT. The method uses Bayesian convolutional neural networks with the U-Net architecture, using Monte Carlo dropout that infers an uncertainty metric in addition to the segmentation label. We evaluated the performance of the proposed met... | Segmentation of muscle tissue and fat tissue has been studied extensively for the analysis of muscle fat composition. (Note that we refer to muscle tissue here as an object including all muscles, not an individual muscle.) @cite_9 and @cite_18 implemented an algorithm for automated segmentation of the muscle and fat ti... | {
"cite_N": [
"@cite_13",
"@cite_9",
"@cite_18",
"@cite_12"
],
"mid": [
"2643397491",
"2606023840",
"2109154320",
"1667869507"
],
"abstract": [
"Pretreatment risk stratification is key for personalized medicine. While many physicians rely on an “eyeball test” to assess whet... |
1907.08915 | 2963092220 | We propose a method for automatic segmentation of individual muscles from a clinical CT. The method uses Bayesian convolutional neural networks with the U-Net architecture, using Monte Carlo dropout that infers an uncertainty metric in addition to the segmentation label. We evaluated the performance of the proposed met... | Segmentation of individual muscles is a much more difficult problem due to the low tissue contrast at the border between neighboring muscles, especially in the area where many muscles are contiguously packed such as in the hip and thigh regions. @cite_15 manually performed segmentation of 35 individual muscles from MRI... | {
"cite_N": [
"@cite_15",
"@cite_31",
"@cite_39"
],
"mid": [
"2162583443",
"1986399714",
"74568156"
],
"abstract": [
"Skelet al muscle is the most abundant tissue in the body and serves various physiological functions including the generation of movement and support. Whole body mot... |
1907.08915 | 2963092220 | We propose a method for automatic segmentation of individual muscles from a clinical CT. The method uses Bayesian convolutional neural networks with the U-Net architecture, using Monte Carlo dropout that infers an uncertainty metric in addition to the segmentation label. We evaluated the performance of the proposed met... | In CT images, due to the lower soft tissue contrast compared to MRI, segmentation of individual muscles is even more difficult. @cite_21 addressed the automated segmentation of individual muscles from CTs of the hip and thigh regions. The target region was broader than @cite_31 covering the origin to insertion of 19 mu... | {
"cite_N": [
"@cite_31",
"@cite_21"
],
"mid": [
"1986399714",
"2796249795"
],
"abstract": [
"We present a novel probabilistic shape representation that implicitly includes prior anatomical volume and adjacency information, termed the generalized log-ratio (GLR) representation. We demonstr... |
1907.08915 | 2963092220 | We propose a method for automatic segmentation of individual muscles from a clinical CT. The method uses Bayesian convolutional neural networks with the U-Net architecture, using Monte Carlo dropout that infers an uncertainty metric in addition to the segmentation label. We evaluated the performance of the proposed met... | In order to enhance the accuracy and speed of the muscle segmentation in CT, we propose an application of convolutional neural networks (CNNs). We investigate the segmentation accuracy as well as a metric indicating uncertainty of the segmentation using the framework of Bayesian deep learning. Yarin @cite_40 found that... | {
"cite_N": [
"@cite_35",
"@cite_4",
"@cite_22",
"@cite_3",
"@cite_6",
"@cite_40",
"@cite_23",
"@cite_16",
"@cite_11"
],
"mid": [
"2095705004",
"2625559849",
"2600383743",
"2887926055",
"2594608309",
"2964059111",
"",
"2488552272",
"288615508... |
1907.08931 | 2969534524 | Regularization in the optimization of deep neural networks is often critical to avoid undesirable over-fitting leading to better generalization of model. One of the most popular regularization algorithms is to impose @math penalty on the model parameters resulting in the decay of parameters, called weight-decay, and th... | Learning Rate Annealing: The stochastic gradient, calculated by a subset of data, gives a noise to gradient and provides an implicit regularization effect @cite_41 . In SGD, parameters are updated by subtracting the gradient with the stochastic noise multiplied by the learning rate. The learning rate should shrink in o... | {
"cite_N": [
"@cite_41",
"@cite_36",
"@cite_53",
"@cite_9",
"@cite_44",
"@cite_31"
],
"mid": [
"2804386825",
"",
"2963702144",
"1522301498",
"2146917784",
"2146502635"
],
"abstract": [
"",
"",
"It is common practice to decay the learning rate. Here ... |
1907.08931 | 2969534524 | Regularization in the optimization of deep neural networks is often critical to avoid undesirable over-fitting leading to better generalization of model. One of the most popular regularization algorithms is to impose @math penalty on the model parameters resulting in the decay of parameters, called weight-decay, and th... | Dropout is another regularization technique that is in particular used with classical shallow networks. The dropout zeros the activation of randomly selected nodes with a certain probability during the training process @cite_50 . The dropping rate is generally set to be constant but its variants have been considered wi... | {
"cite_N": [
"@cite_35",
"@cite_26",
"@cite_4",
"@cite_50",
"@cite_47",
"@cite_13"
],
"mid": [
"",
"2331143823",
"2136836265",
"2095705004",
"2963334011",
"1826234144"
],
"abstract": [
"",
"Very deep convolutional networks with hundreds of layers have l... |
1907.08931 | 2969534524 | Regularization in the optimization of deep neural networks is often critical to avoid undesirable over-fitting leading to better generalization of model. One of the most popular regularization algorithms is to impose @math penalty on the model parameters resulting in the decay of parameters, called weight-decay, and th... | Energy Landscape: The geometrical property of energy surface is helpful in optimization of highly complex non-convex problems associated with deep network architecture. It is preferred to drive a solution toward local minima on a flat energy surface that is considered to yield better generalization @cite_40 @cite_23 @c... | {
"cite_N": [
"@cite_40",
"@cite_25",
"@cite_23"
],
"mid": [
"",
"2605372163",
"2552194003"
],
"abstract": [
"",
"Despite their overwhelming capacity to overfit, deep learning architectures tend to generalize relatively well to unseen data, allowing them to be deployed in pract... |
1907.08931 | 2969534524 | Regularization in the optimization of deep neural networks is often critical to avoid undesirable over-fitting leading to better generalization of model. One of the most popular regularization algorithms is to impose @math penalty on the model parameters resulting in the decay of parameters, called weight-decay, and th... | Variance Reduction: The variance of stochastic gradients is detrimental to SGD, motivating variance reduction techniques @cite_10 @cite_33 @cite_29 @cite_15 @cite_34 @cite_1 @cite_48 that aim to reduce the variance incurred due to their stochastic process of estimation, and improve the convergence rate mainly for conve... | {
"cite_N": [
"@cite_30",
"@cite_38",
"@cite_37",
"@cite_33",
"@cite_48",
"@cite_29",
"@cite_42",
"@cite_1",
"@cite_32",
"@cite_3",
"@cite_15",
"@cite_34",
"@cite_10"
],
"mid": [
"2963139509",
"2962795635",
"",
"2107438106",
"2963734024",
... |
1907.08931 | 2969534524 | Regularization in the optimization of deep neural networks is often critical to avoid undesirable over-fitting leading to better generalization of model. One of the most popular regularization algorithms is to impose @math penalty on the model parameters resulting in the decay of parameters, called weight-decay, and th... | Weight-Decay: is an explicit way of regularization such that a regularization term is added into the energy function. Specifically @math -norm is used as the regularization term in order to penalize large weight values. Different with the other implicit methods, e.g., stochastic update and dropout, one can directly con... | {
"cite_N": [
"@cite_35",
"@cite_31",
"@cite_14",
"@cite_4",
"@cite_22",
"@cite_26",
"@cite_7",
"@cite_36",
"@cite_9",
"@cite_21",
"@cite_52",
"@cite_0",
"@cite_24",
"@cite_47",
"@cite_13",
"@cite_12",
"@cite_11"
],
"mid": [
"",
"21465026... |
1901.07880 | 2913243196 | The Chinese pronunciation system offers two characteristics that distinguish it from other languages: deep phonemic orthography and intonation variations. We are the first to argue that these two important properties can play a major role in Chinese sentiment analysis. Particularly, we propose two effective features to... | One-hot representation is the initial numeric word representation method in NLP. However, it usually leads to a problem of high dimensionality and sparsity. To solve this problem, distributed representation (or word embedding) @cite_25 is proposed. Word embedding is a representation which maps words into low dimensiona... | {
"cite_N": [
"@cite_25"
],
"mid": [
"2132339004"
],
"abstract": [
"A goal of statistical language modeling is to learn the joint probability function of sequences of words in a language. This is intrinsically difficult because of the curse of dimensionality: a word sequence on which the model wil... |
1901.07880 | 2913243196 | The Chinese pronunciation system offers two characteristics that distinguish it from other languages: deep phonemic orthography and intonation variations. We are the first to argue that these two important properties can play a major role in Chinese sentiment analysis. Particularly, we propose two effective features to... | Chinese text differs from English text for two key aspects: it does not have word segmentations and it has a characteristic of compositionality due to its pictogram nature. Based on the former aspect, word segmentation tools are always employed before text representation, such as ICTCLAS @cite_45 , THULAC @cite_19 , Ji... | {
"cite_N": [
"@cite_37",
"@cite_53",
"@cite_57",
"@cite_19",
"@cite_45",
"@cite_31",
"@cite_34"
],
"mid": [
"2566150155",
"2251131401",
"2394700483",
"",
"1982498087",
"2952935105",
"1594229598"
],
"abstract": [
"",
"Languages using Chinese char... |
1901.07880 | 2913243196 | The Chinese pronunciation system offers two characteristics that distinguish it from other languages: deep phonemic orthography and intonation variations. We are the first to argue that these two important properties can play a major role in Chinese sentiment analysis. Particularly, we propose two effective features to... | Sentiment analysis has raised growing interest both within the scientific community, leading to many exciting open challenges, as well as in the business world, due to the remarkable benefits to be had from financial @cite_28 and political @cite_15 forecasting, user profiling @cite_26 and community detection @cite_4 , ... | {
"cite_N": [
"@cite_30",
"@cite_31",
"@cite_26",
"@cite_4",
"@cite_11",
"@cite_7",
"@cite_8",
"@cite_28",
"@cite_46",
"@cite_55",
"@cite_32",
"@cite_6",
"@cite_43",
"@cite_23",
"@cite_15",
"@cite_17"
],
"mid": [
"2166706824",
"2952935105",
... |
1901.07880 | 2913243196 | The Chinese pronunciation system offers two characteristics that distinguish it from other languages: deep phonemic orthography and intonation variations. We are the first to argue that these two important properties can play a major role in Chinese sentiment analysis. Particularly, we propose two effective features to... | Shu and Anderson conducted a study on Chinese phonetic awareness in @cite_0 . The study involved 113 participants of Chinese 2nd, 4th, and 6th graders enrolled in a working-class Beijing, China elementary school. Their task was to represent the pronunciation of 60 semantic phonetic compound characters. Results showed t... | {
"cite_N": [
"@cite_0"
],
"mid": [
"2001970020"
],
"abstract": [
"This study investigated the development of phonetic awareness, meaning insight into the structure and function of the component of Chinese characters that gives a clue to pronunciation. Participants were 113 Chinese 2nd, 4th, and 6... |
1901.07914 | 2913315775 | Modern lightweight dual-arm robots bring the physical capabilities to quickly take over tasks at typical industrial workplaces designed for workers. In times of mass-customization, low setup times including the instructing specifying of new tasks are crucial to stay competitive. We propose a constraint programming appr... | The state-of-the-art optimal TAPF method @cite_15 cannot solve STAAMS problems in general, as it does not compute kinematically feasible motions for agents, nor can it be applied in cases requiring ordering decisions about task assignments. Online methods for multi-agent task assignment and scheduling algorithms have b... | {
"cite_N": [
"@cite_19",
"@cite_15",
"@cite_4",
"@cite_23"
],
"mid": [
"2512486815",
"2482025661",
"2537545174",
"68215201"
],
"abstract": [
"We study the application of constraint programming (CP) to the planning and scheduling of multiple social robots interacting with r... |
1901.07914 | 2913315775 | Modern lightweight dual-arm robots bring the physical capabilities to quickly take over tasks at typical industrial workplaces designed for workers. In times of mass-customization, low setup times including the instructing specifying of new tasks are crucial to stay competitive. We propose a constraint programming appr... | @cite_2 present a survey about the task sequencing problem for industrial robots, where sources for execution variants are systematically identified for a given task specification (e.g., multiple inverse kinematic solutions, partial ordering) and optimized based on various cost functions. The survey, however, lacks the... | {
"cite_N": [
"@cite_22",
"@cite_12",
"@cite_2"
],
"mid": [
"2086550339",
"2479648088",
"2013457645"
],
"abstract": [
"In this paper, we investigate the problem of scheduling a 6 DOF robotic arm to carry out a sequence of spray painting tasks. The duration of any given painting tas... |
1901.08019 | 2912987408 | We present a variation of the Autoencoder (AE) that explicitly maximizes the mutual information between the input data and the hidden representation. The proposed model, the InfoMax Autoencoder (IMAE), by construction is able to learn a robust representation and good prototypes of the data. IMAE is compared both theore... | The first neural network defined to explicitly maximize the information between the input and the hidden layer was proposed by Linsker @cite_5 , giving to the objective function the name, . This model is a linear model that actually maximizes only the entropy of the representation, performing Principal Component Analys... | {
"cite_N": [
"@cite_5",
"@cite_9"
],
"mid": [
"2122925692",
"2108384452"
],
"abstract": [
"The emergence of a feature-analyzing function from the development rules of simple, multilayered networks is explored. It is shown that even a single developing cell of a layered network exhibits a ... |
1901.08019 | 2912987408 | We present a variation of the Autoencoder (AE) that explicitly maximizes the mutual information between the input data and the hidden representation. The proposed model, the InfoMax Autoencoder (IMAE), by construction is able to learn a robust representation and good prototypes of the data. IMAE is compared both theore... | Both these models are quite restrictive, indeed they work only under the assumption that the visible data is a linear combination of hidden features. A more general way to extract hidden features is given by the Autoencoder (AE), a NN that is a composition of an encoder and a decoder map, respectively @math and @math .... | {
"cite_N": [
"@cite_6"
],
"mid": [
"2078626246"
],
"abstract": [
"Abstract We consider the problem of learning from examples in layered linear feed-forward neural networks using optimization methods, such as back propagation, with respect to the usual quadratic error function E of the connection ... |
1901.08019 | 2912987408 | We present a variation of the Autoencoder (AE) that explicitly maximizes the mutual information between the input data and the hidden representation. The proposed model, the InfoMax Autoencoder (IMAE), by construction is able to learn a robust representation and good prototypes of the data. IMAE is compared both theore... | An information theoretic description of an AE was given by @cite_4 where, with restrictive assumptions, they observed that reducing the reconstruction loss of an AE is related to maximizing the mutual information between the visible and hidden variables. | {
"cite_N": [
"@cite_4"
],
"mid": [
"2025768430"
],
"abstract": [
"Previous work has shown that the difficulties in learning deep generative or discriminative models can be overcome by an initial unsupervised learning step that maps inputs to useful intermediate representations. We introduce and m... |
1901.08019 | 2912987408 | We present a variation of the Autoencoder (AE) that explicitly maximizes the mutual information between the input data and the hidden representation. The proposed model, the InfoMax Autoencoder (IMAE), by construction is able to learn a robust representation and good prototypes of the data. IMAE is compared both theore... | The IMAE, as we will see in the next section, is an AE that is able to learn a robust representation. In the literature there are many models in this family that are defined to learn a good representation, see e.g. @cite_13 and references therein. For practical reasons, in the following section we compare the IMAE with... | {
"cite_N": [
"@cite_13"
],
"mid": [
"2616536158"
],
"abstract": [
"Abstract Deep neural networks with several layers have recently become a highly successful and popular research topic in machine learning due to their excellent performance in many benchmark problems and applications. A key idea i... |
1901.08101 | 2950236485 | Can faces acquired by low-cost depth sensors be useful to catch some characteristic details of the face? Typically the answer is no. However, new deep architectures can generate RGB images from data acquired in a different modality, such as depth data. In this paper, we propose a new , trained on annotated RGB-D face d... | GANs have been defined very recently and tested in several contexts. Our work is inspired by the first idea of Goodfellow , of Generative Adversarial Networks @cite_8 with some variant in terms of conditional and discriminative GANs. GANs have been successfully used for Image-to-Image translation; they have been initia... | {
"cite_N": [
"@cite_8",
"@cite_17",
"@cite_6",
"@cite_19",
"@cite_11"
],
"mid": [
"2099471712",
"2605287558",
"2771102363",
"2552465644",
"2963794138"
],
"abstract": [
"We propose a new framework for estimating generative models via an adversarial process, in which... |
1901.08101 | 2950236485 | Can faces acquired by low-cost depth sensors be useful to catch some characteristic details of the face? Typically the answer is no. However, new deep architectures can generate RGB images from data acquired in a different modality, such as depth data. In this paper, we propose a new , trained on annotated RGB-D face d... | The latest spread of high-quality, cheap and accurate commercial depth sensors has encouraged the researchers of the computer vision community. Depth data are a useful source of information especially for systems that have to work in presence of darkness or dramatic light changes. Besides, recent depth sensors usually ... | {
"cite_N": [
"@cite_18",
"@cite_4",
"@cite_9",
"@cite_1",
"@cite_5",
"@cite_15",
"@cite_12",
"@cite_11"
],
"mid": [
"2158211626",
"",
"2763649054",
"2032354375",
"2593748903",
"2132947399",
"2963670480",
"2963794138"
],
"abstract": [
"We con... |
1901.07860 | 2914377850 | Policy evaluation is a key process in reinforcement learning. It assesses a given policy using estimation of the corresponding value function. When using a parameterized function to approximate the value, it is common to optimize the set of parameters by minimizing the sum of squared Bellman Temporal Differences errors... | Bayesian Neural Networks (BNNs): There are several works on Bayesian methods for placing uncertainty on the approximator parameters @cite_39 @cite_40 . have used BNNs for learning MDP dynamics in RL tasks. In these works a fully factorized Gaussian distribution on parameters is assumed while we consider possible correl... | {
"cite_N": [
"@cite_40",
"@cite_39"
],
"mid": [
"2964059111",
"2951266961"
],
"abstract": [
"Deep learning tools have gained tremendous attention in applied machine learning. However such tools for regression and classification do not capture model uncertainty. In comparison, Bayesian mod... |
1901.07860 | 2914377850 | Policy evaluation is a key process in reinforcement learning. It assesses a given policy using estimation of the corresponding value function. When using a parameterized function to approximate the value, it is common to optimize the set of parameters by minimizing the sum of squared Bellman Temporal Differences errors... | Kalman filters: Outside of the RL framework, the use of Kalman filter as an optimization method is discussed in @cite_11 @cite_1 @cite_30 . solve the dynamics of each parameter with Kalman filtering. use Kalman filter for normalizing batches. In our work we use Kalman filtering for VF optimization in the context of RL.... | {
"cite_N": [
"@cite_30",
"@cite_1",
"@cite_3",
"@cite_23",
"@cite_11"
],
"mid": [
"2886639315",
"2898416980",
"2605635402",
"2221583060",
"1581808154"
],
"abstract": [
"We specialize the decoupled extended Kalman filter (DEKF) for online parameter learning in facto... |
1901.07860 | 2914377850 | Policy evaluation is a key process in reinforcement learning. It assesses a given policy using estimation of the corresponding value function. When using a parameterized function to approximate the value, it is common to optimize the set of parameters by minimizing the sum of squared Bellman Temporal Differences errors... | Trust region for policies: The natural gradient method, when applied to RL tasks, is mostly used in policy gradient algorithms to estimate the parameters of the policy @cite_20 @cite_21 @cite_38 . Trust region methods in RL have been developed for parameterized policies @cite_38 @cite_31 . Despite that, trust region me... | {
"cite_N": [
"@cite_38",
"@cite_31",
"@cite_21",
"@cite_20"
],
"mid": [
"2949608212",
"2736601468",
"",
"2404067440"
],
"abstract": [
"We describe an iterative procedure for optimizing policies, with guaranteed monotonic improvement. By making several approximations to the... |
1901.07860 | 2914377850 | Policy evaluation is a key process in reinforcement learning. It assesses a given policy using estimation of the corresponding value function. When using a parameterized function to approximate the value, it is common to optimize the set of parameters by minimizing the sum of squared Bellman Temporal Differences errors... | Distributional perspective on values and observations: Distributional RL @cite_6 treats the full (general) distribution of total return, and considers VF parameters as deterministic. In our work we assume Gaussian distribution over the total return and in addition Gaussian distribution over the VF parameters. | {
"cite_N": [
"@cite_6"
],
"mid": [
"2739473244"
],
"abstract": [
"In this paper we argue for the fundamental importance of the value distribution: the distribution of the random return received by a reinforcement learning agent. This is in contrast to the common approach to reinforcement learning... |
1901.07860 | 2914377850 | Policy evaluation is a key process in reinforcement learning. It assesses a given policy using estimation of the corresponding value function. When using a parameterized function to approximate the value, it is common to optimize the set of parameters by minimizing the sum of squared Bellman Temporal Differences errors... | Our work may be seen as a modern extension of GPTD @cite_37 @cite_5 for DRL domains with continuous state and action spaces. GPTD uses Gaussian Processes (GPs) for both VF and total return, for solving the RL problem of value estimation. We introduce here several improvements and generalizations over their work: (1) Ou... | {
"cite_N": [
"@cite_5",
"@cite_37"
],
"mid": [
"2156974606",
"2132849848"
],
"abstract": [
"Gaussian Process Temporal Difference (GPTD) learning offers a Bayesian solution to the policy evaluation problem of reinforcement learning. In this paper we extend the GPTD framework by addressing ... |
1901.08113 | 2913856657 | Network modeling is a critical component for building self-driving Software-Defined Networks, particularly to find optimal routing schemes that meet the goals set by administrators. However, existing modeling techniques do not meet the requirements to provide accurate estimations of relevant performance metrics such as... | Finally, an early attempt to use Graph Neural Networks for computer networks can be found in @cite_22 . In this case the authors use a GNN to learn shortest-path routing and max-min routing using supervised learning. While this approach is able to generalize to different topologies it cannot generalize to different rou... | {
"cite_N": [
"@cite_22"
],
"mid": [
"2885459608"
],
"abstract": [
"Automated network control and management has been a long standing target of network protocols. We address in this paper the question of automated protocol design, where distributed networked nodes have to cooperate to achieve a co... |
1901.07807 | 2905445635 | Despite technological advances, most smart objects in the Internet of Things (IoT) cannot be accessed using technologies designed and developed for interacting with powerful Internet servers. IoT use cases involve devices that not only have limited resources, but also they are not always connected to the Internet and a... | Prior work on blockchain-assisted access control has proposed schemes that store access control policies in the blockchain. For example, @cite_4 use the Bitcoin blockchain to store Right Transfer Transactions'', i.e., a transaction that indicates that a user is allowed to access a particular resource. These transaction... | {
"cite_N": [
"@cite_4",
"@cite_6",
"@cite_3"
],
"mid": [
"2620085947",
"2620284884",
"1559136758"
],
"abstract": [
"Access Control systems are used in computer security to regulate the access to critical or valuable resources. The rights of subjects to access such resources are ty... |
1901.07807 | 2905445635 | Despite technological advances, most smart objects in the Internet of Things (IoT) cannot be accessed using technologies designed and developed for interacting with powerful Internet servers. IoT use cases involve devices that not only have limited resources, but also they are not always connected to the Internet and a... | A growing body of work propose the use of custom blockhains in order to overcome similar challenges. For example, @cite_0 implement a custom made blockchain for a smart home application and consider per-home miners, which also act as trusted proxies for the home devices. Similarly, @cite_5 propose a blockchain solution... | {
"cite_N": [
"@cite_0",
"@cite_5"
],
"mid": [
"2611626082",
"2588585573"
],
"abstract": [
"Internet of Things (IoT) security and privacy remain a major challenge, mainly due to the massive scale and distributed nature of IoT networks. Blockchain-based approaches provide decentralized secu... |
1901.07827 | 2912648642 | The success of convolutional neural networks (CNNs) in computer vision applications has been accompanied by a significant increase of computation and memory costs, which prohibits its usage on resource-limited environments such as mobile or embedded devices. To this end, the research of CNN compression has recently bec... | Early works in network compression mainly focus on compressing the fully-connected layers @cite_44 @cite_10 @cite_17 @cite_68 @cite_12 . For instance, LeCun @cite_44 and Hassibi @cite_10 proposed a saliency measurement by computing the Hessian matrix of the loss function with respect to the parameters, based on which n... | {
"cite_N": [
"@cite_18",
"@cite_44",
"@cite_68",
"@cite_10",
"@cite_12",
"@cite_17"
],
"mid": [
"2155893237",
"",
"2964299589",
"2125389748",
"2963674932",
"992687842"
],
"abstract": [
"Caffe provides multimedia scientists and practitioners with a clean and... |
1901.07925 | 2953303055 | With the rapid development of spaceborne imaging techniques, object detection in optical remote sensing imagery has drawn much attention in recent decades. While many advanced works have been developed with powerful learning algorithms, the incomplete feature representation still cannot meet the demand for effectively ... | Channel Features refer to a collection of spatially discriminative features by linear or non-linear transformations of the input image. Over the past decades, channel features extraction techniques have been received an increasing interest with successful applications in pedestrian detection @cite_0 @cite_52 and face d... | {
"cite_N": [
"@cite_4",
"@cite_15",
"@cite_21",
"@cite_52",
"@cite_39",
"@cite_0",
"@cite_24",
"@cite_57",
"@cite_5",
"@cite_31"
],
"mid": [
"2323755471",
"2081865073",
"2013081398",
"2159386181",
"2041497292",
"2125556102",
"2161969291",
"2... |
1901.07925 | 2953303055 | With the rapid development of spaceborne imaging techniques, object detection in optical remote sensing imagery has drawn much attention in recent decades. While many advanced works have been developed with powerful learning algorithms, the incomplete feature representation still cannot meet the demand for effectively ... | With a theoretical guarantee, Liu @cite_50 proposed a fourier histogram of oriented gradients (FourierHOG) with a rigorous mathematical proof. It models the rotation-invariant descriptor in a continuous frequency domain rather than in the discrete spatial domain using a Fourier-based convolutionally-manipulated tensor-... | {
"cite_N": [
"@cite_38",
"@cite_50"
],
"mid": [
"2035875937",
"2008213480"
],
"abstract": [
"Abstract Palmprint is one important biometric feature with uniqueness, stability and high distinguishability, and its study has attracted much attention in the past decades. Although many palmprin... |
1901.07925 | 2953303055 | With the rapid development of spaceborne imaging techniques, object detection in optical remote sensing imagery has drawn much attention in recent decades. While many advanced works have been developed with powerful learning algorithms, the incomplete feature representation still cannot meet the demand for effectively ... | In the field of machine learning, the boosting methods have been widely used with great success for decades in various applications, e.g. object detection @cite_53 @cite_28 @cite_59 , face detection @cite_30 , and pose detection @cite_19 @cite_58 . Unlike other powerful classifiers (e.g., Rotation-based SVM @cite_11 , ... | {
"cite_N": [
"@cite_30",
"@cite_35",
"@cite_28",
"@cite_53",
"@cite_29",
"@cite_19",
"@cite_59",
"@cite_58",
"@cite_11"
],
"mid": [
"2137401668",
"2041403134",
"2565366929",
"2086355237",
"2130627644",
"2292006437",
"2547609471",
"2894950214",
... |
1901.07542 | 2911511992 | In almost every election cycle, the validity of the United States Electoral College is brought into question. The 2016 Presidential Election again brought up the issue of a candidate winning the popular vote but not winning the Electoral College, with Hillary Clinton receiving close to three million more votes than Don... | Significant research has been done by Stanford University scholars like Andy Hall, Shanto Iyengar, Bruce Cain, and David Brady, on polarization and election systems. Hall, Political Science Professor at Stanford, claims that "U.S. legislatures have become increasingly polarized and dysfunctional in part because of how ... | {
"cite_N": [
"@cite_14",
"@cite_1",
"@cite_8"
],
"mid": [
"2138174666",
"2793302648",
"2792314235"
],
"abstract": [
"This article studies the interplay of U.S. primary and general elections. I examine how the nomination of an extremist changes general-election outcomes and legisla... |
1901.07621 | 2912660065 | Counterfactual Regret Minimization (CFR) is the most successful algorithm for finding approximate Nash equilibria in imperfect information games. However, CFR's reliance on full game-tree traversals limits its scalability. For this reason, the game's state- and action-space is often abstracted (i.e. simplified) for CFR... | was not the first algorithm that used deep learning with the goal of solving large games efficiently. @cite_23 applies function approximation to estimate regret values in CFR and CFR @math . Unfortunately, despite promising expectations, recent work failed to apply R-CFR in combination with sampling @cite_2 . @cite_9 i... | {
"cite_N": [
"@cite_9",
"@cite_23",
"@cite_2"
],
"mid": [
"2766768413",
"1726776292",
"2951301202"
],
"abstract": [
"Deep reinforcement learning algorithms that estimate state and state-action value functions have been shown to be effective in a variety of challenging domains, inc... |
1901.07621 | 2912660065 | Counterfactual Regret Minimization (CFR) is the most successful algorithm for finding approximate Nash equilibria in imperfect information games. However, CFR's reliance on full game-tree traversals limits its scalability. For this reason, the game's state- and action-space is often abstracted (i.e. simplified) for CFR... | A very successful application of deep learning to CFR in the domain of poker is @cite_4 , an algorithm that was able to defeat professional poker players in head-to-head gameplay in the game of Heads-Up No-Limit Hold'em Poker (HUNL) with statistical significance. However, while DeepStack works very well in poker, it st... | {
"cite_N": [
"@cite_4"
],
"mid": [
"2574978968"
],
"abstract": [
"Artificial intelligence has seen several breakthroughs in recent years, with games often serving as milestones. A common feature of these games is that players have perfect information. Poker, the quintessential game of imperfect i... |
1901.07621 | 2912660065 | Counterfactual Regret Minimization (CFR) is the most successful algorithm for finding approximate Nash equilibria in imperfect information games. However, CFR's reliance on full game-tree traversals limits its scalability. For this reason, the game's state- and action-space is often abstracted (i.e. simplified) for CFR... | To the best of our knowledge, @cite_30 was the first algorithm to soundly apply deep reinforcement learning (deep RL) from single trajectory samples to large extensive-form games. While not showing record-breaking results in terms of exploitability, NFSP was able to learn a competitive strategy in Limit Texas Hold'em P... | {
"cite_N": [
"@cite_30"
],
"mid": [
"2291986326"
],
"abstract": [
"Many real-world applications can be described as large-scale games of imperfect information. To deal with these challenging domains, prior work has focused on computing Nash equilibria in a handcrafted abstraction of the domain. I... |
1901.07621 | 2912660065 | Counterfactual Regret Minimization (CFR) is the most successful algorithm for finding approximate Nash equilibria in imperfect information games. However, CFR's reliance on full game-tree traversals limits its scalability. For this reason, the game's state- and action-space is often abstracted (i.e. simplified) for CFR... | Recent literature elaborates on the convergence properties of multi-agent deep reinforcement learning (deep RL) @cite_14 . A theoretical discussion around the relationship between game theoretic approaches to finding Nash equilibria and the convergence of actor-critic reinforcement learning methods in multi-agent imper... | {
"cite_N": [
"@cite_14",
"@cite_2"
],
"mid": [
"2963937357",
"2951301202"
],
"abstract": [
"There has been a resurgence of interest in multiagent reinforcement learning (MARL), due partly to the recent success of deep neural networks. The simplest form of MARL is independent reinforcement... |
1901.07683 | 2913745806 | Existing method generates class activation map (CAM) by a set of fixed classes (i.e., using all the classes), while the discriminative cues between class pairs are not considered. Note that activation maps by considering different class pair are complementary, and therefore can provide more discriminative cues to overc... | Several class activation map generation methods have been proposed recently @cite_12 @cite_21 @cite_0 @cite_16 @cite_10 . Some methods first introduce the local mechanisms such as global max pooling @cite_14 and log-sum-exp pooling @cite_1 to highlight the activation regions. Then, @cite_9 propose CAM method that uses ... | {
"cite_N": [
"@cite_13",
"@cite_14",
"@cite_9",
"@cite_21",
"@cite_1",
"@cite_0",
"@cite_16",
"@cite_10",
"@cite_12"
],
"mid": [
"2962858109",
"1994488211",
"2295107390",
"2951308125",
"",
"2963609017",
"2793149861",
"2796992393",
"296298156... |
1901.07698 | 2913287814 | In manufacturing and automation settings, robots often have to perform highly-repetitive manipulation tasks in structured environments. In this work we are interested in settings where tasks are similar, yet not identical (e.g., due to uncertain orientation of objects) and motion planning needs to be extremely fast. Pr... | A straightforward approach to efficiently preprocess a known environment is using the PRM algorithm @cite_8 which generates a A roadmap is a graph embedded in the configuration space where vertices correspond to configurations and edges correspond to paths connecting close-by configurations. . Once a a dense roadmap ha... | {
"cite_N": [
"@cite_6",
"@cite_25",
"@cite_23",
"@cite_8"
],
"mid": [
"",
"2096636360",
"2606507580",
"2128990851"
],
"abstract": [
"",
"Provides an analysis of a path planning method which uses probabilistic roadmaps. This method has proven very successful in practice... |
1901.07698 | 2913287814 | In manufacturing and automation settings, robots often have to perform highly-repetitive manipulation tasks in structured environments. In this work we are interested in settings where tasks are similar, yet not identical (e.g., due to uncertain orientation of objects) and motion planning needs to be extremely fast. Pr... | Recently,the repetition roadmap @cite_19 was suggested as a way to extend the PRM for the case of multiple highly-similar scenarios. While this approach exhibits significant speedup in computation time, it still suffers from the previously-mentioned shortcomings. | {
"cite_N": [
"@cite_19"
],
"mid": [
"2883982988"
],
"abstract": [
"We present the Repetition Roadmap, a motion planner that effectively exploits the repetitiveness of a set of tasks with small variations to efficiently compute new motions. The method learns an abstract roadmap of probability dist... |
1901.07698 | 2913287814 | In manufacturing and automation settings, robots often have to perform highly-repetitive manipulation tasks in structured environments. In this work we are interested in settings where tasks are similar, yet not identical (e.g., due to uncertain orientation of objects) and motion planning needs to be extremely fast. Pr... | An alternative approach to address our problem is to precompute a set of complete paths into a library and given a query, attempt to match complete paths from the library to the new query @cite_20 @cite_14 . Using paths from previous search episodes (also known as using experience) has also been an active line of work ... | {
"cite_N": [
"@cite_14",
"@cite_11",
"@cite_9",
"@cite_5",
"@cite_16",
"@cite_13",
"@cite_20",
"@cite_17"
],
"mid": [
"2079148749",
"1485661534",
"2293052688",
"2013329639",
"2108232656",
"",
"1971458750",
"2104761931"
],
"abstract": [
"Traj... |
1901.07698 | 2913287814 | In manufacturing and automation settings, robots often have to perform highly-repetitive manipulation tasks in structured environments. In this work we are interested in settings where tasks are similar, yet not identical (e.g., due to uncertain orientation of objects) and motion planning needs to be extremely fast. Pr... | Finally, our notion of attractor states is similar to control-based methods that ensure safe operation over local regions of the free configuration space @cite_22 @cite_2 . These regions are then used within a high-level motion planner to compute collision-free paths. | {
"cite_N": [
"@cite_22",
"@cite_2"
],
"mid": [
"2165744968",
"2164402964"
],
"abstract": [
"This paper develops a method of composing simple control policies, applicable over a limited region in a dynamical system's free space, such that the resulting composition completely solves the nav... |
1901.07439 | 2911690298 | Recently, Graph Convolutional Networks (GCNs) have been widely studied for graph-structured data representation and learning. However, in many real applications, data are coming with multiple graphs, and it is non-trivial to adapt GCNs to deal with data representation with multiple graph structures. One main challenge ... | Our multiple graph adversarial learning model is inspired by Generative Adversarial Network (GAN) @cite_7 , which consists of a generator @math and a discriminator @math . The generator is trained to generate the samples to convince the discriminator while the discriminator aims to discriminate the samples returned by ... | {
"cite_N": [
"@cite_14",
"@cite_23",
"@cite_12",
"@cite_7"
],
"mid": [
"2963169753",
"2785662987",
"2768104274",
"2099471712"
],
"abstract": [
"",
"Graph embedding is an effective method to represent graph data in a low dimensional space for graph analytics. Most exist... |
1901.07076 | 2963131478 | In recent years, the learned local descriptors have outperformed handcrafted ones by a large margin, due to the powerful deep convolutional neural network architectures such as L2-Net [1] and triplet based metric learning [2]. However, there are two problems in the current methods, which hinders the overall performance... | Recent work on local descriptor designing has gone through a huge change from conventional hand-crafted descriptors to learning-based approaches, which ranges from SIFT @cite_27 and DAISY @cite_16 to latest methods such as DeepCompare, MatchNet, and HardNet @cite_19 @cite_7 @cite_20 @cite_5 . As for deep learning-based... | {
"cite_N": [
"@cite_7",
"@cite_19",
"@cite_27",
"@cite_5",
"@cite_16",
"@cite_20"
],
"mid": [
"1869500417",
"2128237624",
"2151103935",
"2963157250",
"2151638304",
"1929856797"
],
"abstract": [
"Deep learning has revolutionalized image-level tasks such as c... |
1901.07076 | 2963131478 | In recent years, the learned local descriptors have outperformed handcrafted ones by a large margin, due to the powerful deep convolutional neural network architectures such as L2-Net [1] and triplet based metric learning [2]. However, there are two problems in the current methods, which hinders the overall performance... | Before CNN models being broadly applied, descriptors learning methods were limited to specific machine learning descriptors. Therefore, there were various kinds of methods inspired by different aspects. Principal Components Analysis (PCA) based SIFT (PCA-SIFT) @cite_25 leads to normalized gradient patch compared to SIF... | {
"cite_N": [
"@cite_24",
"@cite_26",
"@cite_25",
"@cite_3"
],
"mid": [
"1906968832",
"2089888558",
"2145072179",
"2007178811"
],
"abstract": [
"In this paper we propose a novel approach to generate a binary descriptor optimized for each image patch independently. The appro... |
1901.07076 | 2963131478 | In recent years, the learned local descriptors have outperformed handcrafted ones by a large margin, due to the powerful deep convolutional neural network architectures such as L2-Net [1] and triplet based metric learning [2]. However, there are two problems in the current methods, which hinders the overall performance... | In the past few years, models based on CNN try to get better performance by designing various convolutional neural network architectures, e.g. @cite_20 @cite_23 . @cite_20 choose a two-branched network, a typical Siamese structure for feature extraction and three full connected layers for deep metric learning. @cite_23... | {
"cite_N": [
"@cite_23",
"@cite_20"
],
"mid": [
"1955055330",
"1929856797"
],
"abstract": [
"In this paper we show how to learn directly from image data (i.e., without resorting to manually-designed features) a general similarity function for comparing image patches, which is a task of fu... |
1901.07273 | 2914658942 | We propose a new video representation in terms of an over-segmentation of dense trajectories covering the whole video. Trajectories are often used to encode long-temporal information in several computer vision applications. Similar to temporal superpixels, a temporal slice of super-trajectories are superpixels, but the... | Image segmentation into superpixels is a widely studied problem in Computer Vision. Here we only discuss some of prominent works in this area. Normalized cuts algorithm, by Shi and Malik, uses contour and texture cues to recursively partition the image using a pixel graph @cite_14 . Meanshift, proposed by Comaniciu and... | {
"cite_N": [
"@cite_18",
"@cite_14",
"@cite_22",
"@cite_3",
"@cite_16",
"@cite_12"
],
"mid": [
"1508404128",
"2121947440",
"2067191022",
"",
"2585293115",
""
],
"abstract": [
"We show that the complexity of the recently introduced medoid-shift algorithm in ... |
1901.07046 | 2945411327 | A large number of the most-subscribed YouTube channels target children of very young age. Hundreds of toddler-oriented channels on YouTube feature inoffensive, well produced, and educational videos. Unfortunately, inappropriate content that targets this demographic is also common. YouTube's algorithmic recommendation s... | Several studies focused on understanding videos that target young children, and how they interact with such videos and the platform. Buzzi @cite_12 suggests the addition of extra parental controls on YouTube in an attempt to prevent children from accessing inappropriate content. Ara 'u @cite_26 study the audience profi... | {
"cite_N": [
"@cite_24",
"@cite_26",
"@cite_12",
"@cite_2"
],
"mid": [
"2963592843",
"2963491436",
"2066569906",
"1582800212"
],
"abstract": [
"YouTube draws large number of users who contribute actively by uploading videos or commenting on existing videos. However, being ... |
1901.07046 | 2945411327 | A large number of the most-subscribed YouTube channels target children of very young age. Hundreds of toddler-oriented channels on YouTube feature inoffensive, well produced, and educational videos. Unfortunately, inappropriate content that targets this demographic is also common. YouTube's algorithmic recommendation s... | A large body of previous work focused on the detection of malicious activity on YouTube. @cite_13 use social network analysis techniques to discover hate and extremist YouTube videos, as well as hidden communities in the ecosystem. @cite_22 develop a binary classifier trained with user and video features for detecting ... | {
"cite_N": [
"@cite_18",
"@cite_22",
"@cite_23",
"@cite_13",
"@cite_25"
],
"mid": [
"2797799909",
"2007072922",
"2057916874",
"45052841",
"2003931815"
],
"abstract": [
"As of 2018, YouTube, the major online video sharing website, hosts multiple channels promoting r... |
1901.07046 | 2945411327 | A large number of the most-subscribed YouTube channels target children of very young age. Hundreds of toddler-oriented channels on YouTube feature inoffensive, well produced, and educational videos. Unfortunately, inappropriate content that targets this demographic is also common. YouTube's algorithmic recommendation s... | With regard to spam detection, @cite_10 explore video attributes that may enable the detection of spam videos on YouTube. A similar study by Sureka @cite_20 focuses on both user features and comment activity logs to propose formulas rules that can accurately detect spamming YouTube users. Using similar features, @cite_... | {
"cite_N": [
"@cite_29",
"@cite_10",
"@cite_20",
"@cite_15"
],
"mid": [
"29377604",
"2128388280",
"1839559394",
"2052549182"
],
"abstract": [
"Fraudulent product promotion online, including online videos, is on the rise. In order to understand and defend against this ill, ... |
1901.07046 | 2945411327 | A large number of the most-subscribed YouTube channels target children of very young age. Hundreds of toddler-oriented channels on YouTube feature inoffensive, well produced, and educational videos. Unfortunately, inappropriate content that targets this demographic is also common. YouTube's algorithmic recommendation s... | O' @cite_4 use dynamic network analysis methods to identify the nature of different spam campaign strategies. @cite_9 propose two supervised classification algorithms to detect spammers, promoters, and legitimate YouTube users. Also, in an effort to improve the performance of spam filtering on the platform, @cite_14 te... | {
"cite_N": [
"@cite_19",
"@cite_9",
"@cite_14",
"@cite_4"
],
"mid": [
"2885950427",
"2028772504",
"2295416969",
"2963997978"
],
"abstract": [
"The use of deceptive techniques in user-generated video portals is ubiquitous. Unscrupulous uploaders deliberately mislabel video ... |
1901.07334 | 2913901546 | Recurrent neural networks can be difficult to train on long sequence data due to the well-known vanishing gradient problem. Some architectures incorporate methods to reduce RNN state updates, therefore allowing the network to preserve memory over long temporal intervals. To address these problems of convergence, this p... | There have been a multitude of proposed methods to improve the training of RNNs, especially for long sequences. Apart from incorporating additional gating structures, for example the LSTM and the GRU , more recently various techniques were proposed to further increase the capabilities of recurrent networks to learn on ... | {
"cite_N": [
"@cite_3",
"@cite_2"
],
"mid": [
"2793273050",
"581956982"
],
"abstract": [
"Despite recent advances in training recurrent neural networks (RNNs), capturing long-term dependencies in sequences remains a fundamental challenge. Most approaches use backpropagation through time (... |
1901.07223 | 2912693033 | As the foundation of driverless vehicle and intelligent robots, Simultaneous Localization and Mapping(SLAM) has attracted much attention these days. However, non-geometric modules of traditional SLAM algorithms are limited by data association tasks and have become a bottleneck preventing the development of SLAM. To dea... | Deep learning is considered an excellent solution to SLAM problems due to its superb performance in data association tasks. Part of recent studies makes a straight substitution of an end-to-end network for the traditional SLAM system, estimating ego-motion from monocular video @cite_41 @cite_8 @cite_42 or completing vi... | {
"cite_N": [
"@cite_7",
"@cite_8",
"@cite_41",
"@cite_42",
"@cite_34"
],
"mid": [
"2962887844",
"2963906250",
"2609883120",
"2964314455",
"2951660448"
],
"abstract": [
"Two less addressed issues of deep reinforcement learning are (1) lack of generalization capabili... |
1901.07223 | 2912693033 | As the foundation of driverless vehicle and intelligent robots, Simultaneous Localization and Mapping(SLAM) has attracted much attention these days. However, non-geometric modules of traditional SLAM algorithms are limited by data association tasks and have become a bottleneck preventing the development of SLAM. To dea... | To tackle such problems, some researchers focus on the replacement of only parts of traditional SLAM systems while keeping traditional pipelines unchanged @cite_1 @cite_28 @cite_3 @cite_39 @cite_45 . Such attempts are still in an embryonic stage and do not achieve better results than traditional ones. One of the possib... | {
"cite_N": [
"@cite_28",
"@cite_1",
"@cite_3",
"@cite_39",
"@cite_45"
],
"mid": [
"2830339951",
"2949634581",
"2200124539",
"2737630486",
"2964213180"
],
"abstract": [
"Monocular visual odometry approaches that purely rely on geometric cues are prone to scale drift... |
1901.07223 | 2912693033 | As the foundation of driverless vehicle and intelligent robots, Simultaneous Localization and Mapping(SLAM) has attracted much attention these days. However, non-geometric modules of traditional SLAM algorithms are limited by data association tasks and have become a bottleneck preventing the development of SLAM. To dea... | A simple but effective method is to directly improve the module that limits the performance of traditional SLAM, i.e., stereo matching between frames. Some of them calculate similarity confidence of local features @cite_44 @cite_48 @cite_10 , resulting in the inability to use traditional matching strategy, such as Eucl... | {
"cite_N": [
"@cite_44",
"@cite_48",
"@cite_16",
"@cite_10"
],
"mid": [
"2963502507",
"2440384215",
"2775929773",
"2794086484"
],
"abstract": [
"We present a method for extracting depth information from a rectified image pair. Our approach focuses on the first stage of man... |
1901.07223 | 2912693033 | As the foundation of driverless vehicle and intelligent robots, Simultaneous Localization and Mapping(SLAM) has attracted much attention these days. However, non-geometric modules of traditional SLAM algorithms are limited by data association tasks and have become a bottleneck preventing the development of SLAM. To dea... | Parallel with the long history of SLAM, considerable attempts have been made on local features. Based on classical hand-craft local features like SIFT @cite_26 , SURF @cite_14 , ORB @cite_37 , early combination of low-level machine learning and local feature descriptors produce PCA-SIFT @cite_24 , ASD @cite_17 , BOLD @... | {
"cite_N": [
"@cite_37",
"@cite_14",
"@cite_26",
"@cite_18",
"@cite_36",
"@cite_32",
"@cite_24",
"@cite_0",
"@cite_20",
"@cite_17"
],
"mid": [
"2117228865",
"1677409904",
"2143238378",
"1893620550",
"2126338221",
"2097934584",
"2145072179",
... |
1901.07223 | 2912693033 | As the foundation of driverless vehicle and intelligent robots, Simultaneous Localization and Mapping(SLAM) has attracted much attention these days. However, non-geometric modules of traditional SLAM algorithms are limited by data association tasks and have become a bottleneck preventing the development of SLAM. To dea... | Thanks to the booming of Deep Learning, researchers have gone further. End-to-end networks consisting of multiple independent components @cite_31 @cite_16 @cite_38 @cite_30 can not only give out local feature descriptors through one forward computation but also extract local feature detectors. | {
"cite_N": [
"@cite_30",
"@cite_38",
"@cite_31",
"@cite_16"
],
"mid": [
"2614218061",
"2804394440",
"2320444803",
"2775929773"
],
"abstract": [
"We propose an attentive local feature descriptor suitable for large-scale image retrieval, referred to as DELF (DEep Local Featu... |
1901.07223 | 2912693033 | As the foundation of driverless vehicle and intelligent robots, Simultaneous Localization and Mapping(SLAM) has attracted much attention these days. However, non-geometric modules of traditional SLAM algorithms are limited by data association tasks and have become a bottleneck preventing the development of SLAM. To dea... | Focusing only on descriptors, most researchers adopt multi-branch CNN-based architectures like Siamese and triplet networks. Multi-branch networks were first proposed to verify whether the handwritten signatures were consistent in 1994 @cite_27 . Experiments related to similarity measurements further confirm the superi... | {
"cite_N": [
"@cite_46",
"@cite_40",
"@cite_27",
"@cite_6"
],
"mid": [
"2219193941",
"1929856797",
"2171590421",
"1955055330"
],
"abstract": [
"Recent innovations in training deep convolutional neural network (ConvNet) models have motivated the design of new methods to aut... |
1901.06778 | 2912495464 | Head pose estimation, which computes the intrinsic Euler angles (yaw, pitch, roll) from a target human head, is crucial for gaze estimation, face alignment and 3D reconstruction. Traditional approaches to head pose estimation heavily relies on the accuracy of facial landmarks, and solve the correspondence problem betwe... | Another recent work done by @cite_16 achieves great performance on public datasets. They propose to use a higher level representation to regress the head pose while using deep learning architectures. They use the uncertainty maps in the form of 2D soft localization heatmap images over selected 5 facial landmarks, and p... | {
"cite_N": [
"@cite_16"
],
"mid": [
"2902228397"
],
"abstract": [
"Monocular head pose estimation requires learning a model that computes the intrinsic Euler angles for pose (yaw, pitch, roll) from an input image of human face. Annotating ground truth head pose angles for images in the wild is di... |
1901.06778 | 2912495464 | Head pose estimation, which computes the intrinsic Euler angles (yaw, pitch, roll) from a target human head, is crucial for gaze estimation, face alignment and 3D reconstruction. Traditional approaches to head pose estimation heavily relies on the accuracy of facial landmarks, and solve the correspondence problem betwe... | Although recent state-of-the-art landmark-based method has better prediction given the ground truth of landmark, they suffer from landmark invisibility and the accuracy of landmark under real scene. Robust landmark-free method introduces extra error which limits its performance. In our work, we follow the landmark-free... | {
"cite_N": [
"@cite_5"
],
"mid": [
"2761173690"
],
"abstract": [
"Estimating the head pose of a person is a crucial problem that has a large amount of applications such as aiding in gaze estimation, modeling attention, fitting 3D models to video and performing face alignment. Traditionally head p... |
1901.06904 | 2915067382 | Abstract Sound analysis research has mainly been focused on speech and music processing. The deployed methodologies are not suitable for analysis of sounds with varying background noise, in many cases with very low signal-to-noise ratio (SNR). In this paper, we present a method for the detection of patterns of interest... | Evaluation of algorithms for audio event detection on public benchmark data sets is a valuable tool for objective comparison of performance. The great attention that was dedicated to music and speech analysis determined the publication of several data sets used in scientific challenges for benchmarking of algorithms. T... | {
"cite_N": [
"@cite_30",
"@cite_79",
"@cite_16",
"@cite_25"
],
"mid": [
"",
"2416156489",
"2070707809",
"2086384421"
],
"abstract": [
"",
"This paper is a summary of the first CLEAR evaluation on CLassification of Events, Activities and Relationships - which took place... |
1901.06827 | 2912004605 | Loss functions with a large number of saddle points are one of the main obstacles to training many modern machine learning models. Gradient descent (GD) is a fundamental algorithm for machine learning and converges to a saddle point for certain initial data. We call the region formed by these initial values the "attrac... | @cite_24 showed that GD converges to a local minimizer almost surely if the initial point is randomly chosen. The proof is based on the stable manifold theorem and hence no upper bound of the number of steps is provided. How to escape from saddle points efficiently has been a core problem in non-convex optimization. Th... | {
"cite_N": [
"@cite_24"
],
"mid": [
"2283214199"
],
"abstract": [
"We show that gradient descent converges to a local minimizer, almost surely with random initialization. This is proved by applying the Stable Manifold Theorem from dynamical systems theory."
]
} |
1901.06827 | 2912004605 | Loss functions with a large number of saddle points are one of the main obstacles to training many modern machine learning models. Gradient descent (GD) is a fundamental algorithm for machine learning and converges to a saddle point for certain initial data. We call the region formed by these initial values the "attrac... | The Hessian is utilized to distinguish saddle point from local minima. By using the Hessian, Nesterov & Polyak @cite_20 designed a cubic regularization algorithm which converges to an @math -second-order stationary point in @math iterations; @cite_25 developed a class of trust region algorithms and achieved the same co... | {
"cite_N": [
"@cite_3",
"@cite_25",
"@cite_20",
"@cite_2"
],
"mid": [
"2609037894",
"2398500126",
"2009941369",
""
],
"abstract": [
"We design a non-convex second-order optimization algorithm that is guaranteed to return an approximate local minimum in time which scales li... |
1901.06827 | 2912004605 | Loss functions with a large number of saddle points are one of the main obstacles to training many modern machine learning models. Gradient descent (GD) is a fundamental algorithm for machine learning and converges to a saddle point for certain initial data. We call the region formed by these initial values the "attrac... | Since the computation of the Hessian is often too expensive in practice, algorithms without second-order information are very desirable. @cite_23 proved that stochastic gradient descent (SGD) can find local minima of strict saddle functions in polynomial time. Levy @cite_15 showed that noisy normalized gradient descent... | {
"cite_N": [
"@cite_4",
"@cite_9",
"@cite_6",
"@cite_23",
"@cite_15"
],
"mid": [
"2963416883",
"2963092340",
"2963487351",
"2964106499",
"2561193401"
],
"abstract": [
"Although gradient descent (GD) almost always escapes saddle points asymptotically [, 2016], this ... |
1901.06827 | 2912004605 | Loss functions with a large number of saddle points are one of the main obstacles to training many modern machine learning models. Gradient descent (GD) is a fundamental algorithm for machine learning and converges to a saddle point for certain initial data. We call the region formed by these initial values the "attrac... | @cite_17 recently proposed the LSGD based on the theory of Hamilton-Jacobi partial differential equations. LSGD replaces the gradient by the Laplacian smoothed surrogate which can be computed efficiently with the Thomas algorithm or the Fast Fourier Transform (FFT). LSGD can reduce the variance of stochastic gradient o... | {
"cite_N": [
"@cite_17"
],
"mid": [
"2808102858"
],
"abstract": [
"We propose a very simple modification of gradient descent and stochastic gradient descent. We show that when applied to a variety of machine learning models including softmax regression, convolutional neural nets, generative adver... |
1901.06773 | 2912976659 | Typically, Ultra-deep neural network(UDNN) tends to yield high-quality model, but its training process is usually resource intensive and time-consuming. Modern GPU's scarce DRAM capacity is the primary bottleneck that hinders the trainability and the training efficiency of UDNN. In this paper, we present "AccUDNN", an ... | The former avenue has already explored by IT companies with abundant hardware resources. @cite_6 Facebook's experiment employs 256 GPUs to train ResNet-50 with a minibatch size of 8192 and finishes in 1 hour. @cite_22 extends Facebook's experiment to 1024 GPUs and finishes in 15 minutes. However, the later avenue hasn'... | {
"cite_N": [
"@cite_22",
"@cite_6"
],
"mid": [
"2769856846",
"2622263826"
],
"abstract": [
"We demonstrate that training ResNet-50 on ImageNet for 90 epochs can be achieved in 15 minutes with 1024 Tesla P100 GPUs. This was made possible by using a large minibatch size of 32k. To maintain ... |
1901.06654 | 2911942780 | Many biological data analysis processes like Cytometry or Next Generation Sequencing (NGS) produce massive amounts of data which needs to be processed in batches for down-stream analysis. Such datasets are prone to technical variations due to difference in handling the batches possibly at different times, by different ... | Some of the recent deep learning based methods to solve this problem utilize residual networks to learn a near identity mapping from source to target by optimizing the Maximum Mean Discrepancy (MMD) between the transformed source and original target @cite_10 @cite_7 @cite_1 . MMD is one the several methods used to quan... | {
"cite_N": [
"@cite_1",
"@cite_10",
"@cite_7"
],
"mid": [
"1487641199",
"2555502666",
"930928758"
],
"abstract": [
"We consider the problem of learning deep generative models from data. We formulate a method that generates an independent sample via a single feedforward pass throug... |
1901.06610 | 2914913241 | Document classification is a challenging task with important applications. The deep learning approaches to the problem have gained much attention recently. Despite the progress, the proposed models do not incorporate the knowledge of the document structure in the architecture efficiently and not take into account the c... | In the literature exists a variety of methods for document and text classification. More recent works employed deep learning methods. A hierarchical neural architecture was proposed by @cite_15 , whose structure mirrors the hierarchical structure of documents. The intuition underlying the model is that not all parts of... | {
"cite_N": [
"@cite_15"
],
"mid": [
"2470673105"
],
"abstract": [
"We propose a hierarchical attention network for document classification. Our model has two distinctive characteristics: (i) it has a hierarchical structure that mirrors the hierarchical structure of documents; (ii) it has two leve... |
1901.06610 | 2914913241 | Document classification is a challenging task with important applications. The deep learning approaches to the problem have gained much attention recently. Despite the progress, the proposed models do not incorporate the knowledge of the document structure in the architecture efficiently and not take into account the c... | A problem in approaches such as word2vec @cite_2 is that the model does not consider the morphology of the words, applying a distinct vector to each word. This is a limitation, especially for languages with large vocabularies and many rare words. These languages contain many word forms that rarely occur in the training... | {
"cite_N": [
"@cite_9",
"@cite_20",
"@cite_2"
],
"mid": [
"2950133940",
"2952566282",
""
],
"abstract": [
"The recently introduced continuous Skip-gram model is an efficient method for learning high-quality distributed vector representations that capture a large number of precise ... |
1901.06610 | 2914913241 | Document classification is a challenging task with important applications. The deep learning approaches to the problem have gained much attention recently. Despite the progress, the proposed models do not incorporate the knowledge of the document structure in the architecture efficiently and not take into account the c... | In @cite_19 was introduced the Dynamic Convolutional Neural Network (DCNN) for the semantic modeling of sentences. It uses a global pooling operation over linear sequences named -Max Pooling and applied in the network after the last convolutional layer to guarantees that the input to the fully connected layers is indep... | {
"cite_N": [
"@cite_19"
],
"mid": [
"2949300694"
],
"abstract": [
"The ability to accurately represent sentences is central to language understanding. We describe a convolutional architecture dubbed the Dynamic Convolutional Neural Network (DCNN) that we adopt for the semantic modelling of senten... |
1901.06610 | 2914913241 | Document classification is a challenging task with important applications. The deep learning approaches to the problem have gained much attention recently. Despite the progress, the proposed models do not incorporate the knowledge of the document structure in the architecture efficiently and not take into account the c... | A Gated Recurrent Unit - GRU more straightforward version of computing and implementing then LSTM was proposed by @cite_14 . It has two gates (reset and update) that effectively allows the hidden state to drop any information considered irrelevant, allowing a more compact representation. The GRU works similarly to the ... | {
"cite_N": [
"@cite_14"
],
"mid": [
"2950635152"
],
"abstract": [
"In this paper, we propose a novel neural network model called RNN Encoder-Decoder that consists of two recurrent neural networks (RNN). One RNN encodes a sequence of symbols into a fixed-length vector representation, and the other... |
1901.06610 | 2914913241 | Document classification is a challenging task with important applications. The deep learning approaches to the problem have gained much attention recently. Despite the progress, the proposed models do not incorporate the knowledge of the document structure in the architecture efficiently and not take into account the c... | A systematic evaluation of generic convolutional and recurrent architectures for sequence modeling, motivated by some recent results with convolutional neural networks is presented in @cite_5 . These results suggest that convolutional architectures outperform the recurrent neural networks on tasks such as audio synthes... | {
"cite_N": [
"@cite_5"
],
"mid": [
"2792764867"
],
"abstract": [
"For most deep learning practitioners, sequence modeling is synonymous with recurrent networks. Yet recent results indicate that convolutional architectures can outperform recurrent networks on tasks such as audio synthesis and mach... |
1901.06631 | 2950162495 | Community detection refers to the task of discovering groups of vertices sharing similar properties or functions so as to understand the network data. With the recent development of deep learning, graph representation learning techniques are also utilized for community detection. However, the communities can only be in... | Many community detection algorithms have been proposed from different perspectives. One direction is to design some measure of the quality of a community like modularity, and community structure can be uncovered by optimizing such measures @cite_15 @cite_18 . Another direction is to adopt the generative models to descr... | {
"cite_N": [
"@cite_35",
"@cite_18",
"@cite_21",
"@cite_19",
"@cite_15",
"@cite_11"
],
"mid": [
"362613609",
"2209225614",
"2139694940",
"1971432354",
"2151936673",
"2788730644"
],
"abstract": [
"Community detection is an important technique to understand s... |
1901.06958 | 2911953725 | Surface Electromyography (sEMG) is to record muscles' electrical activity from a restricted area of the skin by using electrodes. The sEMG-based gesture recognition is extremely sensitive of inter-session and inter-subject variances. We propose a model and a deep-learning-based domain adaptation method to approximate t... | From the perspective of contraction pattern, hand gestures can determine muscles to be contracted in an isotonic, isometric or mixed pattern. Isotonic contractions involve muscular contractions against resistance in which the length of the muscle changes. Contrary to isotonic contractions, isometric contractions create... | {
"cite_N": [
"@cite_16"
],
"mid": [
"2508819249"
],
"abstract": [
"In recent years, there has been major interest in the exposure to physical therapy during rehabilitation. Several publications have demonstrated its usefulness in clinical medical and human machine interface (HMI) applications. An... |
1901.06958 | 2911953725 | Surface Electromyography (sEMG) is to record muscles' electrical activity from a restricted area of the skin by using electrodes. The sEMG-based gesture recognition is extremely sensitive of inter-session and inter-subject variances. We propose a model and a deep-learning-based domain adaptation method to approximate t... | Muscles generate electric voltage during contraction detraction. EMG detectors measure this signal through electrodes that are attached to the skin. A digital-analogue conversion is performed with a sampling rate of 100 up to 2000 Hz and the outcome is usually normalized into a range of [-1.0, 1.0]. The typical bandwid... | {
"cite_N": [
"@cite_2"
],
"mid": [
"2762706434"
],
"abstract": [
"Hand prostheses controlled by surface electromyography are promising due to the non-invasive approach and the control capabilities offered by machine learning. Nevertheless, dexterous prostheses are still scarcely spread due to con... |
1901.06958 | 2911953725 | Surface Electromyography (sEMG) is to record muscles' electrical activity from a restricted area of the skin by using electrodes. The sEMG-based gesture recognition is extremely sensitive of inter-session and inter-subject variances. We propose a model and a deep-learning-based domain adaptation method to approximate t... | The intra-session gesture recognition have been extensively researched. Existing sEMG-based solutions utilizes time domain, frequency domain, and time-frequency domain features. Many researchers focused on presenting new sEMG features based on their domain knowledge or analyzing existing features to propose new feature... | {
"cite_N": [
"@cite_14",
"@cite_4",
"@cite_7",
"@cite_1",
"@cite_5",
"@cite_10"
],
"mid": [
"2169931829",
"",
"2898716605",
"2591690220",
"",
"2102134492"
],
"abstract": [
"Recent advances in rehabilitation robotics suggest that it may be possible for hand-... |
1901.06958 | 2911953725 | Surface Electromyography (sEMG) is to record muscles' electrical activity from a restricted area of the skin by using electrodes. The sEMG-based gesture recognition is extremely sensitive of inter-session and inter-subject variances. We propose a model and a deep-learning-based domain adaptation method to approximate t... | LSTM units contain a set of gates that are used to control the stages when information enters the memory (input gate: @math ), when it's output (output gate: @math ) and when it's forgotten (forget gate: @math ) as seen in Eq. ). This architecture allows the neural network to learn longer-term dependencies and they are... | {
"cite_N": [
"@cite_3"
],
"mid": [
"1924770834"
],
"abstract": [
"In this paper we compare different types of recurrent units in recurrent neural networks (RNNs). Especially, we focus on more sophisticated units that implement a gating mechanism, such as a long short-term memory (LSTM) unit and a... |
1901.06355 | 2909083954 | Reliably detecting anomalies in a given set of images is a task of high practical relevance for visual quality inspection, surveillance, or medical image analysis. Autoencoder neural networks learn to reconstruct normal images, and hence can classify those images as anomalies, where the reconstruction error exceeds som... | Almost all approaches for anomaly detection with autoencoders require the training data to consist of normal examples only, but this alone is no guarantee for anomalies to have large reconstruction errors. Robust deep autoencoders @cite_12 address this issue by combining denoising autoencoders with robust PCA, thereby ... | {
"cite_N": [
"@cite_24",
"@cite_18",
"@cite_12"
],
"mid": [
"2594867206",
"2340896621",
"2743138268"
],
"abstract": [
"Deep neural networks (DNNs) are powerful nonlinear architectures that are known to be robust to random perturbations of the input. However, these models are vulne... |
1901.06560 | 2914935482 | There is a disconnect between explanatory artificial intelligence (XAI) methods and the types of explanations that are useful for and demanded by society (policy makers, government officials, etc.) Questions that experts in artificial intelligence (AI) ask opaque systems provide inside explanations, focused on debuggin... | In this work, we focus on the types of questions and explanations that explanatory DNN methods can answer. Recent work has looked at ways to correct neural network judgments @cite_12 and different ways to audit such networks by detecting biases @cite_7 . But these judgments are not enough to completely understand the m... | {
"cite_N": [
"@cite_24",
"@cite_18",
"@cite_7",
"@cite_12"
],
"mid": [
"2811104224",
"",
"2765308317",
"2789209926"
],
"abstract": [
"When we are faced with challenging image classification tasks, we often explain our reasoning by dissecting the image, and pointing out pro... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.