source
stringlengths 727
3.71k
| target
stringlengths 9
18.4k
| lan
stringclasses 23
values |
|---|---|---|
This paper investigates the problem of finding a fixed point for a global nonexpansive operator under time-varying communication graphs in real Hilbert spaces, where the global operator is separable and composed of an aggregate sum of local nonexpansive operators. Each local operator is only privately accessible to each agent, and all agents constitute a network. To seek a fixed point of the global operator, it is indispensable for agents to exchange local information and update their solution cooperatively. To solve the problem, two algorithms are developed, called distributed Krasnosel'ski\u{\i}-Mann (D-KM) and distributed block-coordinate Krasnosel'ski\u{\i}-Mann (D-BKM) iterations, for which the D-BKM iteration is a block-coordinate version of the D-KM iteration in the sense of randomly choosing and computing only one block-coordinate of local operators at each time for each agent. It is shown that the proposed two algorithms can both converge weakly to a fixed point of the global operator. Meanwhile, the designed algorithms are applied to recover the classical distributed gradient descent (DGD) algorithm, devise a new block-coordinate DGD algorithm, handle a distributed shortest distance problem in the Hilbert space for the first time, and solve linear algebraic equations in a novel distributed approach. Finally, the theoretical results are corroborated by a few numerical examples.
|
Este artigo investiga o problema de encontrar um ponto fixo para um operador global não expansivo sob grafos de comunicação variantes no tempo em espaços de Hilbert reais, onde o operador global é separável e composto pela soma agregada de operadores não expansivos locais. Cada operador local é acessível apenas privadamente a cada agente, e todos os agentes constituem uma rede. Para buscar um ponto fixo do operador global, é indispensável que os agentes troquem informações locais e atualizem suas soluções cooperativamente. Para resolver o problema, são desenvolvidos dois algoritmos, denominados iterações distribuídas de Krasnosel'ski\u{\i}-Mann (D-KM) e iterações distribuídas de Krasnosel'ski\u{\i}-Mann por coordenadas em blocos (D-BKM), sendo que a iteração D-BKM é uma versão por coordenadas em blocos da iteração D-KM, no sentido de escolher aleatoriamente e calcular apenas uma coordenada em bloco dos operadores locais a cada instante para cada agente. Demonstra-se que os dois algoritmos propostos convergem fracamente para um ponto fixo do operador global. Paralelamente, os algoritmos projetados são aplicados para recuperar o algoritmo clássico de descida de gradiente distribuído (DGD), desenvolver um novo algoritmo DGD por coordenadas em blocos, lidar, pela primeira vez, com um problema distribuído de distância mínima no espaço de Hilbert e resolver equações algébricas lineares por uma nova abordagem distribuída. Por fim, os resultados teóricos são corroborados por alguns exemplos numéricos.
|
pt
|
Given $\beta\in(1,2)$ the fat Sierpinski gasket $\mathcal S_\beta$ is the self-similar set in $\mathbb R^2$ generated by the iterated function system (IFS) \[ f_{\beta,d}(x)=\frac{x+d}{\beta},\quad d\in\mathcal A:=\{(0, 0), (1,0), (0,1)\}. \] Then for each point $P\in\mathcal S_\beta$ there exists a sequence $(d_i)\in\mathcal A^\mathbb N$ such that $P=\sum_{i=1}^\infty d_i/\beta^i$, and the infinite sequence $(d_i)$ is called a \emph{coding} of $P$. In general, a point in $\mathcal S_\beta$ may have multiple codings since the overlap region $\mathcal O_\beta:=\bigcup_{c,d\in\mathcal A, c\ne d}f_{\beta,c}(\Delta_\beta)\cap f_{\beta,d}(\Delta_\beta)$ has non-empty interior, where $\Delta_\beta$ is the convex hull of $\mathcal S_\beta$. In this paper we are interested in the invariant set \[ \widetilde{\mathcal U}_\beta:=\left\{\sum_{i=1}^\infty \frac{d_i}{\beta^i}\in \mathcal S_\beta: \sum_{i=1}^\infty\frac{d_{n+i}}{\beta^i}\notin\mathcal O_\beta~\forall n\ge 0\right\}. \] Then each point in $ \widetilde{\mathcal U}_\beta$ has a unique coding. We show that there is a transcendental number $\beta_c\approx 1.55263$ related to the Thue-Morse sequence, such that $\widetilde{\mathcal U}_\beta$ has positive Hausdorff dimension if and only if $\beta>\beta_{c}$. Furthermore, for $\beta=\beta_c$ the set $\widetilde{\mathcal U}_\beta$ is uncountable but has zero Hausdorff dimension, and for $\beta<\beta_c$ the set $\widetilde{\mathcal U}_\beta$ is at most countable. Consequently, we also answer a conjecture of Sidorov (2007). Our strategy is using combinatorics on words based on the lexicographical characterization of $\widetilde{\mathcal U}_\beta$.
|
Для заданного $\beta\in(1,2)$ толстый ковёр Сьерпинского $\mathcal S_\beta$ — это самоподобное множество на $\mathbb R^2$, порождённое системой итерируемых функций (IFS)
\[
f_{\beta,d}(x)=\frac{x+d}{\beta},\quad d\in\mathcal A:=\{(0, 0), (1,0), (0,1)\}.
\]
Тогда для каждой точки $P\in\mathcal S_\beta$ существует последовательность $(d_i)\in\mathcal A^\mathbb N$, такая что $P=\sum_{i=1}^\infty d_i/\beta^i$, и бесконечная последовательность $(d_i)$ называется \emph{кодировкой} точки $P$. Вообще, точка из $\mathcal S_\beta$ может иметь несколько кодировок, поскольку область перекрытия $\mathcal O_\beta:=\bigcup_{c,d\in\mathcal A, c\ne d}f_{\beta,c}(\Delta_\beta)\cap f_{\beta,d}(\Delta_\beta)$ имеет непустую внутренность, где $\Delta_\beta$ — выпуклая оболочка множества $\mathcal S_\beta$. В данной работе нас интересует инвариантное множество
\[
\widetilde{\mathcal U}_\beta:=\left\{\sum_{i=1}^\infty \frac{d_i}{\beta^i}\in \mathcal S_\beta: \sum_{i=1}^\infty\frac{d_{n+i}}{\beta^i}\notin\mathcal O_\beta~\forall n\ge 0\right\}.
\]
Тогда каждая точка из $\widetilde{\mathcal U}_\beta$ имеет единственную кодировку. Мы показываем, что существует трансцендентное число $\beta_c\approx 1.55263$, связанное с последовательностью Ту-Морса, такое что $\widetilde{\mathcal U}_\beta$ имеет положительную хаусдорфову размерность тогда и только тогда, когда $\beta>\beta_{c}$. Более того, при $\beta=\beta_c$ множество $\widetilde{\mathcal U}_\beta$ несчётно, но имеет нулевую хаусдорфову размерность, а при $\beta<\beta_c$ множество $\widetilde{\mathcal U}_\beta$ не более чем счётно. Следовательно, мы также отвечаем на гипотезу Сидорова (2007). Наша стратегия основана на комбинаторике слов, использующей лексикографическую характеризацию множества $\widetilde{\mathcal U}_\beta$.
|
ru
|
Super-resolution fluorescence microscopy, with a resolution beyond the diffraction limit of light, has become an indispensable tool to directly visualize biological structures in living cells at a nanometer-scale resolution. Despite advances in high-density super-resolution fluorescent techniques, existing methods still have bottlenecks, including extremely long execution time, artificial thinning and thickening of structures, and lack of ability to capture latent structures. Here we propose a novel deep learning guided Bayesian inference approach, DLBI, for the time-series analysis of high-density fluorescent images. Our method combines the strength of deep learning and statistical inference, where deep learning captures the underlying distribution of the fluorophores that are consistent with the observed time-series fluorescent images by exploring local features and correlation along time-axis, and statistical inference further refines the ultrastructure extracted by deep learning and endues physical meaning to the final image. Comprehensive experimental results on both real and simulated datasets demonstrate that our method provides more accurate and realistic local patch and large-field reconstruction than the state-of-the-art method, the 3B analysis, while our method is more than two orders of magnitude faster. The main program is available at https://github.com/lykaust15/DLBI
|
La microscopie de fluorescence en super-résolution, dont la résolution dépasse la limite de diffraction de la lumière, est devenue un outil indispensable pour visualiser directement des structures biologiques dans des cellules vivantes avec une résolution à l'échelle du nanomètre. Malgré les progrès réalisés dans les techniques de fluorescence en super-résolution à haute densité, les méthodes existantes présentent encore des goulots d'étranglement, notamment un temps d'exécution extrêmement long, un amincissement ou un épaississement artificiel des structures, et une incapacité à capturer les structures latentes. Nous proposons ici une nouvelle approche fondée sur l'inférence bayésienne guidée par apprentissage profond, DLBI, pour l'analyse en série temporelle d'images fluorescentes à haute densité. Notre méthode combine les avantages de l'apprentissage profond et de l'inférence statistique : l'apprentissage profond capture la distribution sous-jacente des fluorophores, cohérente avec les images fluorescentes en série temporelle observées, en exploitant les caractéristiques locales et les corrélations le long de l'axe temporel, tandis que l'inférence statistique affine davantage l'ultrastructure extraite par l'apprentissage profond et confère un sens physique à l'image finale. Des résultats expérimentaux complets, obtenus à partir de jeux de données réels et simulés, démontrent que notre méthode permet une reconstruction plus précise et plus réaliste des zones locales et des grands champs que la méthode de référence actuelle, l'analyse 3B, tout en étant plus rapide d'au moins deux ordres de grandeur. Le programme principal est disponible à l'adresse suivante : https://github.com/lykaust15/DLBI
|
fr
|
This paper is concerned with efficiently coloring sparse graphs in the distributed setting with as few colors as possible. According to the celebrated Four Color Theorem, planar graphs can be colored with at most 4 colors, and the proof gives a (sequential) quadratic algorithm finding such a coloring. A natural problem is to improve this complexity in the distributed setting. Using the fact that planar graphs contain linearly many vertices of degree at most 6, Goldberg, Plotkin, and Shannon obtained a deterministic distributed algorithm coloring $n$-vertex planar graphs with 7 colors in $O(\log n)$ rounds. Here, we show how to color planar graphs with 6 colors in $\mbox{polylog}(n)$ rounds. Our algorithm indeed works more generally in the list-coloring setting and for sparse graphs (for such graphs we improve by at least one the number of colors resulting from an efficient algorithm of Barenboim and Elkin, at the expense of a slightly worst complexity). Our bounds on the number of colors turn out to be quite sharp in general. Among other results, we show that no distributed algorithm can color every $n$-vertex planar graph with 4 colors in $o(n)$ rounds.
|
Ang papel na ito ay tungkol sa mahusay na pagkakulay ng mga sparse graph sa isang distributed na kalagayan gamit ang kasing kaunti ng posibleng kulay. Ayon sa kilalang Four Color Theorem, ang mga planar graph ay maaaring kulayan gamit ang hanggang sa 4 na kulay, at ang patunay ay nagbibigay ng isang (sequential) quadratic algorithm upang mahanap ang gayong pagkakulay. Isang likas na problema ay ang pagpapabuti ng kumplikadong ito sa isang distributed na kalagayan. Gamit ang katotohanan na ang mga planar graph ay naglalaman ng linear na bilang ng mga vertex na may degree na hindi lalagpas sa 6, nakamit nina Goldberg, Plotkin, at Shannon ang isang deterministic distributed algorithm na nagkukulay ng mga planar graph na may $n$-vertex gamit ang 7 kulay sa loob ng $O(\log n)$ na mga pagkakabisa. Dito, ipinapakita namin kung paano kulayan ang mga planar graph gamit ang 6 na kulay sa loob ng $\mbox{polylog}(n)$ na mga pagkakabisa. Ang aming algorithm ay talagang gumagana nang mas pangkalahatan sa list-coloring na kalagayan at para sa mga sparse graph (para sa mga ganitong graph, binabawasan namin ng hindi bababa sa isa ang bilang ng mga kulay na nagmumula sa isang mahusay na algorithm ni Barenboim at Elkin, sa kabila ng bahagyang mas mataas na kumplikado). Ang aming mga hangganan sa bilang ng mga kulay ay lumalabas na medyo matalim sa pangkalahatan. Bukod sa iba pang mga resulta, ipinapakita namin na walang anumang distributed algorithm na kayang kulayan ang bawat $n$-vertex na planar graph gamit ang 4 na kulay sa loob ng $o(n)$ na mga pagkakabisa.
|
tl
|
Self-assembly of colloidal particles due to elastic interactions in nematic liquid crystals promises tunable composite materials and can be guided by exploiting surface functionalization, geometric shape and topology, though these means of controlling self-assembly remain limited. Here, we realize low-symmetry achiral and chiral elastic colloids in the nematic liquid crystals using colloidal polygonal concave and convex prisms. We show that the controlled pinning of disclinations at the prisms edges alters the symmetry of director distortions around the prisms and their orientation with respect to the far-field director. The controlled localization of the disclinations at the prism's edges significantly influences anisotropy of the diffusion properties of prisms dispersed in liquid crystals and allows one to modify their self-assembly. We show that elastic interactions between polygonal prisms can be switched between repulsive and attractive just by controlled re-pinning the disclinations at different edges using laser tweezers. Our findings demonstrate that elastic interactions between colloidal particles dispersed in nematic liquid crystals are sensitive to the topologically equivalent but geometrically rich controlled configurations of the particle-induced defects.
|
Penggabungan sendiri zarah koloid akibat interaksi elastik dalam hablur cecair nematik menjanjikan bahan komposit boleh laras dan boleh diarahkan melalui penggunaan pengfungsian permukaan, bentuk geometri dan topologi, walaupun kaedah mengawal penggabungan sendiri ini masih terhad. Di sini, kami mewujudkan koloid elastik simetri-rendah yang tidak kiral dan kiral dalam hablur cecair nematik menggunakan prisma poligon cekung dan cembung koloid. Kami menunjukkan bahawa pengekalan terkawal disklusi pada tepi prisma mengubah simetri gangguan pengarah di sekeliling prisma dan orientasinya berbanding pengarah medan-jauh. Penempatan terkawal disklusi pada tepi prisma memberi kesan besar terhadap anisotropi sifat resapan prisma yang terserak dalam hablur cecair dan membolehkan pengubahsuaian penggabungan sendiri mereka. Kami menunjukkan bahawa interaksi elastik antara prisma poligon boleh dialihkan antara tolakan dan tarikan hanya dengan mengawal semula pengekalan disklusi pada tepi yang berbeza menggunakan pengetip laser. Penemuan kami menunjukkan bahawa interaksi elastik antara zarah koloid yang terserak dalam hablur cecair nematik adalah sensitif terhadap konfigurasi terkawal yang setara secara topologi tetapi kaya secara geometri bagi cacat yang diaruhkan zarah.
|
ms
|
We propose an algebraic framework for generalized graph Laplacians which unifies the study of resistor networks, the critical group, and the eigenvalues of the Laplacian and adjacency matrices. Given a graph with boundary $G$ together with a generalized Laplacian $L$ with entries in a commutative ring $R$, we define a generalized critical group $\Upsilon_R(G,L)$. We relate $\Upsilon_R(G,L)$ to spaces of harmonic functions on the network using the Hom, Tor, and Ext functors of homological algebra. We study how these algebraic objects transform under combinatorial operations on the network $(G,L)$, including harmonic morphisms, layer-stripping, duality, and symmetry. In particular, we use layer-stripping operations from the theory of resistor networks to systematize discrete harmonic continuation. This leads to an algebraic characterization of the graphs with boundary that can be completely layer-stripped, an algorithm for simplifying computation of $\Upsilon_R(G,L)$, and upper bounds for the number of invariant factors in the critical group and the multiplicity of Laplacian eigenvalues in terms of geometric quantities.
|
ພວກເຮົາສະເໜີເຂດແບບພີດຄະນິດສຳລັບຕົວດຳເນີນການລາພາສຽນ (Laplacian) ທົ່ວໄປ ເຊິ່ງເຮັດໃຫ້ການສຶກສາເຄືອຂ່າຍຕ້ານທານ, ກຸ່ມຄຳວ່າສຳຄັນ (critical group), ແລະ ຄ່າລັກສະນະຂອງຕົວດຳເນີນການລາພາສຽນ ແລະ ຕົວດຳເນີນການຕໍ່ກັນ (adjacency matrices) ສາມາດສຶກສາໄດ້ພາຍໃນເຂດດຽວກັນ. ໃຫ້ເຄືອຂ່າຍ G ທີ່ມີຂອບ ແລະ ຕົວດຳເນີນການລາພາສຽນທົ່ວໄປ L ທີ່ມີສ່ວນປະກອບຢູ່ໃນວົງກົມຄະນິດສາດ (commutative ring) R, ພວກເຮົາກຳນົດກຸ່ມຄຳວ່າສຳຄັນທົ່ວໄປ \Upsilon_R(G,L). ພວກເຮົາເຊື່ອມຕໍ່ \Upsilon_R(G,L) ກັບພື້ນທີ່ຂອງຟັງຊັ່ນຮາມົນິກ (harmonic functions) ໃນເຄືອຂ່າຍໂດຍໃຊ້ຕົວດຳເນີນການ Hom, Tor, ແລະ Ext ຈາກພີດຄະນິດເຊິ່ງກ່ຽວຂ້ອງກັບໂຄງສ້າງທາງພີດຄະນິດ (homological algebra). ພວກເຮົາສຶກສາວ່າ ວັດຖຸທາງພີດຄະນິດເຫຼົ່ານີ້ ປ່ຽນແປງແນວໃດ ໃຕ້ການດຳເນີນງານທາງປະສົມປະສານ (combinatorial operations) ຕໍ່ເຄືອຂ່າຍ (G,L), ລວມທັງ morphisms ຮາມົນິກ, ການຖອກຊັ້ນ (layer-stripping), ຄວາມສຳພັນກັນ (duality), ແລະ ຄວາມສົມດຸນ (symmetry). ໂດຍສະເພາະ, ພວກເຮົາໃຊ້ການດຳເນີນງານຖອກຊັ້ນ ຈາກທິດສະດີຂອງເຄືອຂ່າຍຕ້ານທານ ເພື່ອຈັດລະບົບການຕໍ່ເນື່ອງຮາມົນິກແບບດິສເຄີດ (discrete harmonic continuation). ສິ່ງນີ້ນຳໄປສູ່ການອະທິບາຍແບບພີດຄະນິດຂອງເຄືອຂ່າຍທີ່ມີຂອບ ເຊິ່ງສາມາດຖອກຊັ້ນອອກໄດ້ທັງໝົດ, ອະລະກໍລິດຝີມື (algorithm) ສຳລັບການງ່າຍຂຶ້ນໃນການຄຳນວນ \Upsilon_R(G,L), ແລະ ຂອບເຂດສູງສຸດ ສຳລັບຈຳນວນຂອງປັດໃຈຄົງທີ່ (invariant factors) ໃນກຸ່ມຄຳວ່າສຳຄັນ ແລະ ຄວາມຖີ່ຂອງຄ່າລັກສະນະຂອງຕົວດຳເນີນການລາພາສຽນ ໂດຍອີງໃສ່ປະລິມານທາງເລຂາຄະນິດ.
|
lo
|
Semantic segmentation is an extensively studied task in computer vision, with numerous methods proposed every year. Thanks to the advent of deep learning in semantic segmentation, the performance on existing benchmarks is close to saturation. A natural question then arises: Does the superior performance on the closed (and frequently re-used) test sets transfer to the open visual world with unconstrained variations? In this paper, we take steps toward answering the question by exposing failures of existing semantic segmentation methods in the open visual world under the constraint of very limited human labeling effort. Inspired by previous research on model falsification, we start from an arbitrarily large image set, and automatically sample a small image set by MAximizing the Discrepancy (MAD) between two segmentation methods. The selected images have the greatest potential in falsifying either (or both) of the two methods. We also explicitly enforce several conditions to diversify the exposed failures, corresponding to different underlying root causes. A segmentation method, whose failures are more difficult to be exposed in the MAD competition, is considered better. We conduct a thorough MAD diagnosis of ten PASCAL VOC semantic segmentation algorithms. With detailed analysis of experimental results, we point out strengths and weaknesses of the competing algorithms, as well as potential research directions for further advancement in semantic segmentation. The codes are publicly available at \url{https://github.com/QTJiebin/MAD_Segmentation}.
|
سیمینٹک سیگمنٹیشن کمپیوٹر ویژن میں وسیع پیمانے پر مطالعہ کیا جانے والا کام ہے، جس کے لیے ہر سال بے شمار طریقے تجویز کیے جاتے ہیں۔ سیمینٹک سیگمنٹیشن میں گہری سیکھنے (ڈیپ لرننگ) کی آمد کی بدولت موجودہ بینچ مارکس پر کارکردگی تقریباً اطمینان بخش حد تک پہنچ چکی ہے۔ اس کے نتیجے میں ایک قدرتی سوال یہ اٹھتا ہے: کیا بند (اور بار بار استعمال ہونے والے) ٹیسٹ سیٹس پر بہتر کارکردگی کھلے ویژل دنیا میں غیر محدود تغیرات کے ساتھ بھی برقرار رہتی ہے؟ اس مقالے میں، ہم بہت محدود انسانی لیبلنگ کی کوششوں کی پابندی کے تحت کھلی ویژل دنیا میں موجودہ سیمینٹک سیگمنٹیشن طریقوں کی ناکامیوں کو عیاں کر کے اس سوال کا جواب دینے کی کوشش کرتے ہیں۔ ماڈل کی تکذیب (model falsification) پر پچھلی تحقیقات سے متاثر ہو کر، ہم ایک اعتباطی طور پر بڑے امیج سیٹ سے شروع کرتے ہیں، اور دو سیگمنٹیشن طریقوں کے درمیان فرق (Discrepancy) کو زیادہ سے زیادہ (MAximizing the Discrepancy (MAD)) کر کے خودکار طریقے سے ایک چھوٹا امیج سیٹ منتخب کرتے ہیں۔ منتخب شدہ امیجز میں دونوں طریقوں میں سے کسی ایک (یا دونوں) کی تکذیب کی زیادہ سے زیادہ صلاحیت ہوتی ہے۔ ہم ناکامیوں کو متنوع بنانے کے لیے کئی شرائط کو بھی صریح طور پر نافذ کرتے ہیں، جو مختلف بنیادی وجوہات کے مطابق ہوتی ہیں۔ ایک سیگمنٹیشن طریقہ جس کی ناکامیوں کو MAD مقابلے میں عیاں کرنا مشکل ہو، کو بہتر سمجھا جاتا ہے۔ ہم دس PASCAL VOC سیمینٹک سیگمنٹیشن الخوارزمیوں کا مکمل MAD تشخیص کرتے ہیں۔ تجرباتی نتائج کے تفصیلی تجزیے کے ذریعے، ہم مقابلہ کرنے والی الخوارزمیوں کی طاقتوں اور کمزوریوں کی نشاندہی کرتے ہیں، ساتھ ہی سیمینٹک سیگمنٹیشن میں مزید ترقی کے ممکنہ تحقیقی رخ بھی بتاتے ہیں۔ کوڈز عوامی طور پر \url{https://github.com/QTJiebin/MAD_Segmentation} پر دستیاب ہیں۔
|
ur
|
OSD PSE is the Indonesian Government Certification Authority (CA) for National e-Procurement System and later named OSD PSE G2. It has a unique hierarchical structure under the OSD Lemsaneg. As an Issuing CA, the OSD PSE G2 publishes and guarantee the quality of the Certificate Policy and Certification Practice Statement (CP-CPS) in order to gain the PKI user trustworthy. In this article, we analyze the CP-CPS version 1.0 that published by OSD PSE G2. For this purpose, we apply the methodology of PKI Assessment Guidelines (PAG). The quality assessment of this CP-CPS, including its compliance to the related reference/standard, namely: CP OSD Lemsaneg v.1.1; RFC 3647; and CA Business Practice Disclosure Principle on Trust Service Principles and Criteria for Certification Authorities (BPDP-TSPCCA) version 2.0. We finally found that the CP-CPS version 1.0 does not comply with related standard and reference. Hence, the CP-CPS need to be updated following the current condition of OSD PSE G2.
|
OSD PSE è l'Autorità di Certificazione (CA) del Governo indonesiano per il Sistema Nazionale di e-Procurement, successivamente denominato OSD PSE G2. Ha una struttura gerarchica unica sotto l'OSD Lemsaneg. In qualità di CA emittente, OSD PSE G2 pubblica e garantisce la qualità della Policy dei Certificati e della Dichiarazione sulle Pratiche di Certificazione (CP-CPS), al fine di ottenere la fiducia degli utenti PKI. In questo articolo analizziamo la versione 1.0 della CP-CPS pubblicata da OSD PSE G2. A tal fine, applichiamo la metodologia delle PKI Assessment Guidelines (PAG). La valutazione della qualità della CP-CPS include il rispetto delle norme e riferimenti pertinenti, ossia: CP OSD Lemsaneg v.1.1; RFC 3647; e Principi sulla divulgazione delle pratiche aziendali delle CA e Principi e criteri per i servizi di fiducia delle Autorità di Certificazione (BPDP-TSPCCA) versione 2.0. Alla fine, abbiamo riscontrato che la CP-CPS versione 1.0 non è conforme ai riferimenti e alle norme pertinenti. Pertanto, la CP-CPS deve essere aggiornata in base alle attuali condizioni di OSD PSE G2.
|
it
|
CPU branch prediction has hit a wall--existing techniques achieve near-perfect accuracy on 99% of static branches, and yet the mispredictions that remain hide major performance gains. In a companion report, we show that a primary source of mispredictions is a handful of systematically hard-to-predict branches (H2Ps), e.g. just 10 static instructions per SimPoint phase in SPECint 2017. The lost opportunity posed by these mispredictions is significant to the CPU: 14.0% in instructions-per-cycle (IPC) on Intel SkyLake and 37.4% IPC when the pipeline is scaled four-fold, on par with gains from process technology. However, up to 80% of this upside is unreachable by the best known branch predictors, even when afforded exponentially more resources. New approaches are needed, and machine learning (ML) provides a palette of powerful predictors. A growing body of work has shown that ML models are deployable within the microarchitecture to optimize hardware at runtime, and are one way to customize CPUs post-silicon by training to customer applications. We develop this scenario for branch prediction using convolutional neural networks (CNNs) to boost accuracy for H2Ps. Step-by-step, we (1) map CNNs to the global history data used by existing branch predictors; (2) show how CNNs improve H2P prediction in SPEC 2017; (3) adapt 2-bit CNN inference to the constraints of current branch prediction units; and (4) establish that CNN helper predictors are reusable across application executions on different inputs, enabling us to amortize offline training and deploy ML pattern matching to improve IPC.
|
CPU 분기 예측은 한계에 부딪혔다. 기존 기법들은 정적 분기의 99%에서 거의 완벽한 정확도를 달성하지만, 남아 있는 오측정들에는 상당한 성능 향상 잠재력이 숨어 있다. 동반 보고서에서 우리는 이러한 오측정의 주요 원인이 체계적으로 예측하기 어려운 소수의 분기(H2P)들, 예를 들어 SPECint 2017의 각 시뮬레이션 지점(SimPoint) 단계에서 단지 10개의 정적 명령어에 불과하다는 것을 보여준다. 이러한 오측정으로 인해 CPU가 놓치는 기회는 상당하다. 인텔 스카이레이크(SkyLake)에서는 사이클당 명령어 수(IPC)가 14.0% 감소하며, 파이프라인이 4배 확장된 환경에서는 IPC가 37.4% 감소하는데, 이는 공정 기술 향상으로 얻는 성능 향상과 맞먹는 수준이다. 그러나 이 모든 개선 잠재력의 최대 80%는 기존에 알려진 최고의 분기 예측기로도, 지수적으로 더 많은 자원을 투입하더라도 달성할 수 없다. 새로운 접근이 필요하며, 머신러닝(ML)은 강력한 예측 기법들을 제공한다. 점점 늘어나는 연구들은 머신러닝 모델이 마이크로아키텍처 내에 구현되어 하드웨어를 런타임에 최적화할 수 있음을 보여주었으며, 고객 애플리케이션에 맞춰 훈련함으로써 실리콘 제작 이후에도 CPU를 맞춤화할 수 있는 방법 중 하나임을 입증하고 있다. 우리는 분기 예측 분야에 대해 이 시나리오를 발전시켜, 합성곱 신경망(CNN)을 이용해 H2P의 예측 정확도를 향상시킨다. 단계별로 (1) 기존 분기 예측기가 사용하는 전역 히스토리 데이터에 CNN을 매핑하고, (2) CNN이 SPEC 2017에서 H2P 예측을 어떻게 향상시키는지 보여주며, (3) 현재의 분기 예측 유닛 제약 조건에 맞게 2비트 CNN 추론을 조정하고, (4) CNN 보조 예측기가 서로 다른 입력을 가진 애플리케이션 실행 간에 재사용 가능하다는 것을 입증하여 오프라인 훈련 비용을 분산시킬 수 있으며, ML 기반 패턴 매칭을 배치해 IPC 향상을 실현할 수 있음을 확인한다.
|
ko
|
This paper is an exploratory study of two epidemiological questions on a worldwide basis. How fast is the disease spreading? Are the restrictions (especially mobility restrictions) for people bring the expected effect? To answer the first question, we propose a tool for estimating the reproduction number of epidemic (the number of secondary infections $R_t$) based on the SEIR model and compare it with an non-model $R_t$ estimation. To measure the $R_t$ of COVID-19 for different countries, real-time data on coronavirus daily cases of infections, recoveries, deaths are retrieved from the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University. To assess the effectiveness of mobility restrictions for the COVID-19 pandemic in 2020, the correlations between the $R_t$ and people's mobility (based on the Apple mobility index) are presented. The correlations were considered for 12 countries and for most of them, the correlations are negative. This shows a delay in the implementation of mobility restrictions - the countries imposed them in response to growth of new COVID-19 cases, rather than preventively.
|
এই গবেষণাপত্রটি বিশ্বব্যাপী দুটি মহামারী বিজ্ঞানের প্রশ্ন নিয়ে একটি অনুসন্ধানমূলক গবেষণা। রোগটি কত দ্রুত ছড়াচ্ছে? মানুষের জন্য আরোপিত বিধিনিষেধ (বিশেষ করে চলাচল সংক্রান্ত বিধিনিষেধ) আশানুরূপ ফল দিচ্ছে কি? প্রথম প্রশ্নের উত্তর দিতে, আমরা SEIR মডেলের উপর ভিত্তি করে মহামারীর পুনরুৎপাদন সংখ্যা (দ্বিতীয় সংক্রমণের সংখ্যা $R_t$) অনুমান করার জন্য একটি সরঞ্জাম প্রস্তাব করি এবং এটিকে একটি মডেলবিহীন $R_t$ অনুমানের সাথে তুলনা করি। বিভিন্ন দেশের জন্য COVID-19 এর $R_t$ পরিমাপ করার জন্য, জনস হপকিন্স বিশ্ববিদ্যালয়ের সিস্টেম সায়েন্স অ্যান্ড ইঞ্জিনিয়ারিং সেন্টার (CSSE) থেকে করোনাভাইরাসের দৈনিক আক্রান্ত, সুস্থ এবং মৃত্যুর বাস্তব-সময়ের তথ্য সংগ্রহ করা হয়। 2020 সালে কোভিড-19 মহামারীর ক্ষেত্রে চলাচল সংক্রান্ত বিধিনিষেধের কার্যকারিতা মূল্যায়ন করার জন্য, $R_t$ এবং মানুষের চলাচল (অ্যাপল মোবিলিটি সূচকের ভিত্তিতে) এর মধ্যে সম্পর্কগুলি উপস্থাপন করা হয়। 12টি দেশের জন্য এই সম্পর্কগুলি বিবেচনা করা হয় এবং তাদের অধিকাংশের ক্ষেত্রে সম্পর্কগুলি ঋণাত্মক। এটি চলাচল সংক্রান্ত বিধিনিষেধ বাস্তবায়নে বিলম্ব দেখায় - দেশগুলি নতুন কোভিড-19 আক্রান্তের সংখ্যা বৃদ্ধির প্রতিক্রিয়া হিসাবে বিধিনিষেধ আরোপ করেছে, প্রতিরোধমূলকভাবে নয়।
|
bn
|
We study the solar energetic particle (SEP) event associated with the 2012 July 23 extreme solar storm, for which STEREO and the spacecraft at L1 provide multi-point remote sensing and in situ observations. The extreme solar storm, with a superfast shock and extremely enhanced ejecta magnetic fields observed near 1 AU at STEREO A, was caused by the combination of successive coronal mass ejections (CMEs). Meanwhile, energetic particles were observed by STEREO and near-Earth spacecraft such as ACE and SOHO, suggestive of a wide longitudinal spread of the particles at 1 AU. Combining the SEP observations with in situ plasma and magnetic field measurements we investigate the longitudinal distribution of the SEP event in connection with the associated shock and CMEs. Our results underscore the complex magnetic configuration of the inner heliosphere formed by solar eruptions. The examinations of particle intensities, proton anisotropy distributions, element abundance ratios, magnetic connectivity and spectra also give important clues for the particle acceleration, transport and distribution.
|
Біз STEREO және L1 нүктесіндегі ғарыштық аппараттар арқылы көп нүктелі алыстан бақылау және тікелей бақылаулар жасалған 2012 жылғы шілденің 23-індегі аса күшті күн даңғылымен байланысты күннің энергиялық бөлшектерінің (SEP) оқиғасын зерттейміз. STEREO A жақын аймағында 1 АЕ маңында бақыланған супер жылдам соққы мен едәуір күшейген ежілу магнит өрістерімен сипатталатын аса күшті күн даңғылы бірнеше қатарлас короналық массалардың ежілуінің (CME) қосылуы нәтижесінде туындады. Сол уақытта STEREO, сонымен қатар Жерге жақын ғарыштық аппараттар – ACE және SOHO энергиялық бөлшектерді бақылады, бұл бөлшектердің 1 АЕ-де бойлық бағытта кең таралуын көрсетеді. SEP бақылауларын тікелей бақыланатын плазма мен магнит өрісі өлшемдерімен біріктіре отырып, біз байланысты соққы мен CME-лерге қатысты SEP оқиғасының бойлық таралуын зерттейміз. Нәтижелеріміз күн жарқылдары нәтижесінде қалыптасқан гелиосфераның ішкі аймағының күрделі магниттік конфигурациясын айқындайды. Бөлшектердің интенсивтілігін, протондардың анизотропия таралуын, элементтердің мөлшерлерінің қатынасын, магниттік байланысты және спектрлерді тексеру бөлшектердің үдеуі, тасымалдануы мен таралуы туралы маңызды кілттік деректерді береді.
|
kk
|
The light rays from a source are subject to a local inhomogeneous geometry generated by inhomogeneous matter distribution as well as the existence of collapsed objects. In this paper we investigate the effect of inhomogeneities and the existence of collapsed objects on the propagation of light rays and evaluate changes in the magnitude-redshift relation from the standard relationship found in a homogeneous FRW universe. We give the expression of the correlation function and the variance for the perturbation of apparent magnitude, and calculate it numerically by using the non-linear matter power spectrum. We use the lognormal probability distribution function for the density contrast and spherical collapse model to truncate the power spectrum in order to estimate the blocking effect by collapsed objects. We find that the uncertainties in $\Omega_m$ is $\sim 0.02$, and that of $w$ is $\sim 0.04$. We also discuss a possible method to extract these effects from real data which contains intrinsic ambiguities associated with the absolute magnitude.
|
একটি উৎস থেকে আগত আলোক রশ্মিগুলি অসম পদার্থ বন্টন এবং সঙ্কুচিত বস্তুর অস্তিত্বের কারণে উৎপন্ন স্থানীয় অসম জ্যামিতির অধীন। এই প্রবন্ধে, আমরা আলোক রশ্মির প্রসারণের উপর অসমতার প্রভাব এবং সঙ্কুচিত বস্তুর অস্তিত্বের প্রভাব অনুসন্ধান করি এবং একটি সম এফআরডব্লিউ (FRW) মহাবিশ্বে পাওয়া আদর্শ সম্পর্কের তুলনায় দৃশ্যমান মাত্রা-রেডশিফট সম্পর্কের পরিবর্তনগুলি মূল্যায়ন করি। আমরা দৃশ্যমান মাত্রার ব্যতিক্রমের জন্য সম্পর্ক অপেক্ষক এবং ভেদাঙ্কের রাশিমালা দিয়েছি এবং অরৈখিক পদার্থ পাওয়ার স্পেকট্রাম ব্যবহার করে এর সংখ্যাগত গণনা করেছি। ঘনত্ব পার্থক্যের জন্য আমরা লগনরমাল সম্ভাব্যতা বন্টন অপেক্ষক এবং সঙ্কুচিত বস্তুগুলির দ্বারা উৎপন্ন অবরোধের প্রভাব আনুমানিক করার জন্য গোলাকার সঙ্কোচন মডেল ব্যবহার করে পাওয়ার স্পেকট্রাম কাটছাঁট করেছি। আমরা পেয়েছি যে $\Omega_m$-এর অনিশ্চয়তা $\sim 0.02$ এবং $w$-এর অনিশ্চয়তা $\sim 0.04$। আমরা আলোচনা করেছি বাস্তব তথ্য থেকে এই প্রভাবগুলি নিষ্কাশনের একটি সম্ভাব্য পদ্ধতি যা পরম মাত্রার সঙ্গে সম্পর্কিত অন্তর্নিহিত অস্পষ্টতা ধারণ করে।
|
bn
|
The discovery of new families of exfoliatable 2D crystals that have diverse sets of electronic, optical, and spin-orbit coupling properties, enables the realization of unique physical phenomena in these few-atom thick building blocks and in proximity to other materials. Herein, using NaSn2As2 as a model system, we demonstrate that layered Zintl phases having the stoichiometry ATt2Pn2 (A = Group 1 or 2 element, Tt = Group 14 tetrel element and Pn = Group 15 pnictogen element) and feature networks separated by van der Waals gaps can be readily exfoliated with both mechanical and liquid-phase methods. We identified the symmetries of the Raman active modes of the bulk crystals via polarized Raman spectroscopy. The bulk and mechanically exfoliated NaSn2As2 samples are resistant towards oxidation, with only the top surface oxidizing in ambient conditions over a couple of days, while the liquid-exfoliated samples oxidize much more quickly in ambient conditions. Employing angle-resolved photoemission spectroscopy (ARPES), density functional theory (DFT), and transport on bulk and exfoliated samples, we show that NaSn2As2 is a highly conducting 2D semimetal, with resistivities on the order of 10-6 {\Omega} m. Due to peculiarities in the band structure, the dominating p-type carriers at low temperature are nearly compensated by the opening of n-type conduction channels as temperature increases. This work further expands the family of exfoliatable 2D materials to layered van der Waals Zintl phases, opening up opportunities in electronics and spintronics.
|
বিভিন্ন ধরনের ইলেকট্রনিক, আলোকীয় এবং স্পিন-অরবিটাল যুক্ত বৈশিষ্ট্যের অধিকারী নতুন পরিবারের বিচ্ছায়নযোগ্য 2D স্ফটিকের আবিষ্কার এই কয়েকটি পরমাণুর পুরুত্বের গাঠনিক ব্লকগুলি এবং অন্যান্য উপকরণগুলির কাছাকাছি অনন্য পদার্থবিজ্ঞানের ঘটনাগুলি বাস্তবায়নের সুযোগ করে দেয়। এখানে, NaSn2As2-কে মডেল সিস্টেম হিসাবে ব্যবহার করে, আমরা দেখাই যে স্তরযুক্ত জিন্টল ফেজগুলি যাদের স্টোয়াকিওমেট্রি হল ATt2Pn2 (A = গ্রুপ 1 বা 2 এর মৌল, Tt = গ্রুপ 14 টেট্রেল মৌল এবং Pn = গ্রুপ 15 প্নিক্টোজেন মৌল) এবং যারা ভ্যান ডার ওয়ালস ফাঁক দ্বারা পৃথক নেটওয়ার্ক নিয়ে গঠিত, তাদের সহজেই যান্ত্রিক এবং তরল-পর্যায় পদ্ধতি দুটি দ্বারা বিচ্ছায়িত করা যায়। ধ্রুবীকৃত রামান স্পেকট্রোস্কোপির মাধ্যমে আমরা বাল্ক ক্রিস্টালগুলির রামান সক্রিয় মোডগুলির প্রতিসাম্যতা চিহ্নিত করেছি। বাল্ক এবং যান্ত্রিকভাবে বিচ্ছায়িত NaSn2As2 নমুনাগুলি জারণের প্রতি প্রতিরোধী, কয়েকদিন ধরে পরিবেশের শর্তে শুধুমাত্র উপরের পৃষ্ঠটি জারিত হয়, অন্যদিকে তরল-বিচ্ছায়িত নমুনাগুলি পরিবেশের শর্তে অনেক দ্রুত জারিত হয়। কোণ-নির্ভর ফটো-নিঃসরণ স্পেকট্রোস্কোপি (ARPES), ঘনত্ব কার্য তত্ত্ব (DFT) এবং বাল্ক ও বিচ্ছায়িত নমুনাগুলির উপর পরিবহন পরীক্ষা ব্যবহার করে, আমরা দেখাই যে NaSn2As2 একটি অত্যন্ত পরিবাহী 2D অর্ধধাতু, যার রোধত্ব প্রায় 10-6 {\Omega} m ক্রমের। ব্যান্ড গঠনের বিশেষ বৈশিষ্ট্যের কারণে, নিম্ন তাপমাত্রায় প্রাধান্য পাওয়া p-প্রকার বাহকগুলি তাপমাত্রা বৃদ্ধির সাথে সাথে n-প্রকার পরিবহন চ্যানেল খোলার মাধ্যমে প্রায় ক্ষতিপূরণ পায়। এই কাজটি বিচ্ছায়নযোগ্য 2D উপকরণের পরিবারকে স্তরযুক্ত ভ্যান ডার ওয়ালস জিন্টল ফেজগুলিতে আরও প্রসারিত করে, ইলেকট্রনিক্স এবং স্পিনট্রনিক্সে সুযোগগুলি খুলে দেয়।
|
bn
|
Successful applications of reinforcement learning in real-world problems often require dealing with partially observable states. It is in general very challenging to construct and infer hidden states as they often depend on the agent's entire interaction history and may require substantial domain knowledge. In this work, we investigate a deep-learning approach to learning the representation of states in partially observable tasks, with minimal prior knowledge of the domain. In particular, we propose a new family of hybrid models that combines the strength of both supervised learning (SL) and reinforcement learning (RL), trained in a joint fashion: The SL component can be a recurrent neural networks (RNN) or its long short-term memory (LSTM) version, which is equipped with the desired property of being able to capture long-term dependency on history, thus providing an effective way of learning the representation of hidden states. The RL component is a deep Q-network (DQN) that learns to optimize the control for maximizing long-term rewards. Extensive experiments in a direct mailing campaign problem demonstrate the effectiveness and advantages of the proposed approach, which performs the best among a set of previous state-of-the-art methods.
|
غالبًا ما تتطلب التطبيقات الناجحة للتعلم المعزز في المشكلات الواقعية التعامل مع حالات جزئية الملاحظة. وعمومًا، فإن بناء الحالات المخفية واستنتاجها يُعد أمرًا بالغ الصعوبة، لأنها غالبًا ما تعتمد على تاريخ تفاعل العامل بأكمله وقد تتطلب معرفة واسعة بالنطاق. في هذا العمل، ندرس نهجًا يعتمد على التعلم العميق لتعلم تمثيل الحالات في المهام جزئية الملاحظة، مع الحد الأدنى من المعرفة المسبقة بالنطاق. وتحديدًا، نقترح عائلة جديدة من النماذج الهجينة التي تجمع بين قوة التعلم الإشرافي (SL) والتعلم المعزز (RL)، ويتم تدريبها بشكل مشترك: يمكن أن يكون المكوّن الإشرافي عبارة عن شبكة عصبية متكررة (RNN) أو نسختها للذاكرة قصيرة وطويلة المدى (LSTM)، والتي تمتلك الخاصية المطلوبة المتمثلة في القدرة على التقاط التبعيات طويلة المدى للتاريخ، وبالتالي توفر وسيلة فعالة لتعلم تمثيل الحالات المخفية. أما المكوّن المعزز فهو شبكة Q العميقة (DQN) التي تتعلم تحسين التحكم بهدف تعظيم المكافآت طويلة المدى. وتُظهر تجارب واسعة النطاق في مشكلة حملة البريد المباشر فعالية ومزايا النهج المقترح، الذي حقق الأداء الأفضل مقارنة بمجموعة من الطرق الرائدة السابقة.
|
ar
|
We initiate the study of multi-layered cake cutting with the goal of fairly allocating multiple divisible resources (layers of a cake) among a set of agents. The key requirement is that each agent can only utilize a single resource at each time interval. Several real-life applications exhibit such restrictions on overlapping pieces; for example, assigning time intervals over multiple facilities and resources or assigning shifts to medical professionals. We investigate the existence and computation of envy-free and proportional allocations. We show that envy-free allocations that are both feasible and contiguous are guaranteed to exist for up to three agents with two types of preferences, when the number of layers is two. We also show that envy-free feasible allocations where each agent receives a polynomially bounded number of intervals exist for any number of agents and layers under mild conditions on agents' preferences. We further devise an algorithm for computing proportional allocations for any number of agents and layers.
|
우리는 다수의 가변 자원(케이크의 여러 층)을 일련의 행위자들 사이에 공평하게 배분하는 것을 목표로 하는 다층 케이크 절단 문제의 연구를 시작한다. 핵심 요구사항은 각 행위자가 각 시간 구간에서 단 하나의 자원만 사용할 수 있다는 점이다. 일상 생활의 여러 실제 응용 분야에서 중복되는 조각에 대한 제한이 나타나는데, 예를 들어 여러 시설 및 자원에 대한 시간 구간 할당이나 의료 전문가에게 근무 교대를 할당하는 경우가 있다. 우리는 질투가 없는(jealousy-free) 및 비례적(proportional) 배분의 존재성과 계산 가능성을 조사한다. 우리는 층의 수가 두 개인 경우, 두 가지 유형의 선호를 갖는 최대 세 명의 행위자에 대해, 실현 가능하고 동시에 연속적인 질투 없는 배분이 반드시 존재함을 보인다. 또한, 행위자들의 선호에 대한 완화된 조건 하에서, 행위자 수와 층 수가 임의일 때 각 행위자가 다항식 수준으로 제한된 수의 구간을 받는 실현 가능한 질투 없는 배분이 존재함을 보인다. 더 나아가, 임의의 수의 행위자와 층에 대해 비례적 배분을 계산하는 알고리즘을 고안하였다.
|
ko
|
We prove the computational weakness of a model of tile assembly that has so far resisted many attempts of formal analysis or positive constructions. Specifically, we prove that, in Winfree's abstract Tile Assembly Model, when restricted to use only noncooperative bindings, any long enough path that can grow in all terminal assemblies is pumpable, meaning that this path can be extended into an infinite, ultimately periodic path. This result can be seen as a geometric generalization of the pumping lemma of finite state automata, and closes the question of what can be computed deterministically in this model. Moreover, this question has motivated the development of a new method called visible glues. We believe that this method can also be used to tackle other long-standing problems in computational geometry, in relation for instance with self-avoiding paths. Tile assembly (including non-cooperative tile assembly) was originally introduced by Winfree and Rothemund in STOC 2000 to understand how to program shapes. The non-cooperative variant, also known as temperature 1 tile assembly, is the model where tiles are allowed to bind as soon as they match on one side, whereas in cooperative tile assembly, some tiles need to match on several sides in order to bind. In this work, we prove that only very simple shapes can indeed be programmed, whereas exactly one known result (SODA 2014) showed a restriction on the assemblies general non-cooperative self-assembly could achieve, without any implication on its computational expressiveness. With non-square tiles (like polyominos, SODA 2015), other recent works have shown that the model quickly becomes computationally powerful.
|
Kami membuktikan kelemahan komputasi dari suatu model asamblil ubin yang hingga kini telah menahan banyak upaya analisis formal maupun konstruksi positif. Secara khusus, kami membuktikan bahwa dalam Model Asamblil Ubin Abstrak Winfree, apabila dibatasi hanya menggunakan ikatan nonkooperatif, setiap lintasan yang cukup panjang yang dapat tumbuh dalam semua asamblil terminal bersifat pompa-boleh, artinya lintasan ini dapat diperluas menjadi lintasan tak hingga yang pada akhirnya periodik. Hasil ini dapat dipandang sebagai generalisasi geometris dari lemma pompa pada automata berhingga, serta menutup pertanyaan mengenai apa yang dapat dihitung secara deterministik dalam model ini. Selain itu, pertanyaan ini telah mendorong pengembangan suatu metode baru yang disebut lem yang terlihat (visible glues). Kami meyakini bahwa metode ini juga dapat digunakan untuk menangani masalah-masalah lama lainnya dalam geometri komputasi, terkait misalnya dengan lintasan yang menghindari diri sendiri. Asamblil ubin (termasuk asamblil ubin nonkooperatif) pada awalnya diperkenalkan oleh Winfree dan Rothemund dalam STOC 2000 untuk memahami bagaimana memprogram bentuk. Variannya yang nonkooperatif, juga dikenal sebagai asamblil ubin suhu 1, merupakan model di mana ubin diperbolehkan melekat begitu saja selama cocok pada satu sisi, sedangkan dalam asamblil ubin kooperatif, beberapa ubin harus cocok pada beberapa sisi agar dapat melekat. Dalam karya ini, kami membuktikan bahwa hanya bentuk-bentuk sangat sederhana yang sebenarnya dapat diprogram, sementara sebelumnya hanya ada satu hasil yang diketahui (SODA 2014) yang menunjukkan pembatasan pada asamblil yang dapat dicapai oleh asamblil diri nonkooperatif secara umum, tanpa implikasi terhadap ekspresivitas komputasionalnya. Dengan ubin nonpersegi (seperti polyomino, SODA 2015), karya-karya terkini lainnya telah menunjukkan bahwa model ini dengan cepat menjadi sangat kuat secara komputasi.
|
id
|
A novel algorithm for wide-baseline matching called MODS - Matching On Demand with view Synthesis - is presented. The MODS algorithm is experimentally shown to solve a broader range of wide-baseline problems than the state of the art while being nearly as fast as standard matchers on simple problems. The apparent robustness vs. speed trade-off is finessed by the use of progressively more time-consuming feature detectors and by on-demand generation of synthesized images that is performed until a reliable estimate of geometry is obtained. We introduce an improved method for tentative correspondence selection, applicable both with and without view synthesis. A modification of the standard first to second nearest distance rule increases the number of correct matches by 5-20% at no additional computational cost. Performance of the MODS algorithm is evaluated on several standard publicly available datasets, and on a new set of geometrically challenging wide baseline problems that is made public together with the ground truth. Experiments show that the MODS outperforms the state-of-the-art in robustness and speed. Moreover, MODS performs well on other classes of difficult two-view problems like matching of images from different modalities, with wide temporal baseline or with significant lighting changes.
|
Представлен новый алгоритм сопоставления при большом базисе, названный MODS — сопоставление по требованию с синтезом видов (Matching On Demand with view Synthesis). Экспериментально показано, что алгоритм MODS решает более широкий спектр задач с большим базисом по сравнению с современными методами, оставаясь при этом почти таким же быстрым, как стандартные сопоставители, на простых задачах. Очевидный компромисс между устойчивостью и скоростью обходится за счёт использования постепенно более трудоёмких детекторов признаков и генерации синтезированных изображений по требованию, которая продолжается до получения надёжной оценки геометрии. Вводится усовершенствованный метод отбора предварительных соответствий, применимый как с синтезом видов, так и без него. Модификация стандартного правила отношения расстояний от первого до второго ближайшего соседа увеличивает количество правильных сопоставлений на 5–20% без дополнительных вычислительных затрат. Производительность алгоритма MODS оценивается на нескольких стандартных общедоступных наборах данных, а также на новом наборе геометрически сложных задач с большим базисом, который публикуется вместе с эталонными данными. Эксперименты показывают, что MODS превосходит современные методы по устойчивости и скорости. Кроме того, MODS хорошо справляется с другими классами трудных задач сопоставления двух изображений, например, при сопоставлении изображений из разных модальностей, с большим временным интервалом или при значительных изменениях освещения.
|
ru
|
Late blight disease is one of the most destructive diseases in potato crop, leading to serious yield losses globally. Accurate diagnosis of the disease at early stage is critical for precision disease control and management. Current farm practices in crop disease diagnosis are based on manual visual inspection, which is costly, time consuming, subject to individual bias. Recent advances in imaging sensors (e.g. RGB, multiple spectral and hyperspectral cameras), remote sensing and machine learning offer the opportunity to address this challenge. Particularly, hyperspectral imagery (HSI) combining with machine learning/deep learning approaches is preferable for accurately identifying specific plant diseases because the HSI consists of a wide range of high-quality reflectance information beyond human vision, capable of capturing both spectral-spatial information. The proposed method considers the potential disease specific reflectance radiation variance caused by the canopy structural diversity, introduces the multiple capsule layers to model the hierarchical structure of the spectral-spatial disease attributes with the encapsulated features to represent the various classes and the rotation invariance of the disease attributes in the feature space. We have evaluated the proposed method with the real UAV-based HSI data under the controlled field conditions. The effectiveness of the hierarchical features has been quantitatively assessed and compared with the existing representative machine learning/deep learning methods. The experiment results show that the proposed model significantly improves the accuracy performance when considering hierarchical-structure of spectral-spatial features, comparing to the existing methods only using spectral, or spatial or spectral-spatial features without consider hierarchical-structure of spectral-spatial features.
|
A doença da requeima tardia é uma das doenças mais destrutivas na cultura da batata, provocando perdas significativas na produtividade em nível global. O diagnóstico preciso da doença em estágios iniciais é essencial para o controle e manejo preciso da doença. Atualmente, as práticas agrícolas no diagnóstico de doenças de culturas baseiam-se em inspeção visual manual, que é onerosa, demorada e sujeita a vieses individuais. Avanços recentes em sensores de imagem (por exemplo, câmeras RGB, multiespectrais e hiperspectrais), sensoriamento remoto e aprendizado de máquina oferecem a oportunidade de enfrentar esse desafio. Em particular, a imagem hiperspectral (HSI) combinada com abordagens de aprendizado de máquina/aprendizado profundo é preferível para identificar com precisão doenças específicas das plantas, pois a HSI contém uma ampla gama de informações de refletância de alta qualidade para além da visão humana, sendo capaz de capturar informações espectrais e espaciais simultaneamente. O método proposto considera a possível variância na radiação refletida específica da doença causada pela diversidade estrutural da copa, introduzindo múltiplas camadas de cápsulas para modelar a estrutura hierárquica dos atributos espectrais-espaciais da doença, com características encapsuladas para representar as diversas classes e a invariância rotacional dos atributos da doença no espaço de características. Avaliamos o método proposto com dados reais de HSI baseados em UAV, sob condições de campo controladas. A eficácia das características hierárquicas foi avaliada quantitativamente e comparada com métodos representativos existentes de aprendizado de máquina/aprendizado profundo. Os resultados dos experimentos mostram que o modelo proposto melhora significativamente o desempenho de precisão ao considerar a estrutura hierárquica das características espectrais-espaciais, em comparação com os métodos existentes que utilizam apenas características espectrais, ou espaciais, ou espectrais-espaciais sem considerar a estrutura hierárquica dessas características.
|
pt
|
The effect of the intersite and interplane Coulomb interactions between the Dirac fermions on the formation of the Kohn-Luttinger superconductivity in bilayer doped graphene is studied disregarding the effects of the van der Waals potential of the substrate and both magnetic and non-magnetic impurities. The phase diagram determining the boundaries of superconductive domains with different types of symmetry of the order parameter is built using the extended Hubbard model in the Born weak-coupling approximation with allowance for the intratomic, interatomic, and interlayer Coulomb interactions between electrons. It is shown that the Kohn-Luttinger polarization contributions up to the second order of perturbation theory in the Coulomb interaction inclusively and an account for the long-range intraplane Coulomb interactions significantly affect the competition between the superconducting $f-$, $p+ip-$, and $d+id-$wave pairings. It is demonstrated that the account for the interplane Coulomb interaction enhances the critical temperature of the transition to the superconducting phase.
|
Дирак фермиондары арасындағы сайтар арасындағы және жазықтықтар арасындағы Кулон әсерлесуі екі қабатты қоспалы графенде Коэн-Латтингер суперөткізгіштігінің пайда болуына әсері зерттелген, алайда субстраттың ван-дер-Ваальс потенциалы мен магнитті және магнитті емес қоспалардың әсері ескерілмеген. Тәртіп параметрінің әртүрлі симметрия түрлеріне сәйкес келетін суперөткізгіштік аймақтарының шекараларын анықтайтын фазалық диаграмма электрондар арасындағы атомішілік, атомдараралық және қабаттараралық Кулон әсерлесулерін ескере отырып, кеңейтілген Хаббард модельінің Борнның әлсіз байланысу жуықтауын қолдану арқылы құрылған. Кулон әсерлесуі бойынша екінші дәрежеге дейінгі Кулон-Латтингер поляризациялық үлестерін және жазықтық ішіндегі ұзақ мәжбүрлі Кулон әсерлесулерін ескеру $f-$, $p+ip-$ және $d+id-$толқынды суперөткізгіш жұптар арасындағы бәсекеге маңызды әсер ететіні көрсетілген. Қабаттар арасындағы Кулон әсерлесуін ескеру суперөткізгіш фазаға өту критикалық температурасын арттыратыны көрсетілген.
|
kk
|
We present a general form of Renormalization operator $\mathcal{R}$ acting on potentials $V:\{0,1\}^\mathbb{N} \to \mathbb{R}$. We exhibit the analytical expression of the fixed point potential $V$ for such operator $\mathcal{R}$. This potential can be expressed in a naturally way in terms of a certain integral over the Hausdorff probability on a Cantor type set on the interval $[0,1]$. This result generalizes a previous one by A. Baraviera, R. Leplaideur and A. Lopes where the fixed point potential $V$ was of Hofbauer type. For the potentials of Hofbauer type (a well known case of phase transition) the decay is like $n^{-\gamma}$, $\gamma>0$. Among other things we present the estimation of the decay of correlation of the equilibrium probability associated to the fixed potential $V$ of our general renormalization procedure. In some cases we get polynomial decay like $n^{-\gamma}$, $\gamma>0$, and in others a decay faster than $n \,e^{ -\, \sqrt{n}}$, when $n \to \infty$. The potentials $g$ we consider here are elements of the so called family of Walters potentials on $\{0,1\}^\mathbb{N} $ which generalizes the potentials considered initially by F. Hofbauer. For these potentials some explicit expressions for the eigenfunctions are known. In a final section we also show that given any choice $d_n \to 0$ of real numbers varying with $n \in \mathbb{N}$ there exist a potential $g$ on the class defined by Walters which has a invariant probability with such numbers as the coefficients of correlation (for a certain explicit observable function).
|
Nous présentons une forme générale de l'opérateur de renormalisation $\mathcal{R}$ agissant sur des potentiels $V:\{0,1\}^\mathbb{N} \to \mathbb{R}$. Nous exhibons l'expression analytique du potentiel fixe $V$ pour un tel opérateur $\mathcal{R}$. Ce potentiel peut s'exprimer naturellement en termes d'une certaine intégrale par rapport à la probabilité de Hausdorff sur un ensemble de type Cantor contenu dans l'intervalle $[0,1]$. Ce résultat généralise un résultat antérieur d'A. Baraviera, R. Leplaideur et A. Lopes, dans lequel le potentiel fixe $V$ était de type Hofbauer. Pour les potentiels de type Hofbauer (un cas bien connu de transition de phase), la décroissance est de l'ordre de $n^{-\gamma}$, avec $\gamma>0$. Entre autres choses, nous présentons l'estimation de la décroissance des corrélations de la probabilité d'équilibre associée au potentiel fixe $V$ issu de notre procédure générale de renormalisation. Dans certains cas, nous obtenons une décroissance polynomiale du type $n^{-\gamma}$, $\gamma>0$, et dans d'autres une décroissance plus rapide que $n \,e^{ -\, \sqrt{n}}$, lorsque $n \to \infty$. Les potentiels $g$ que nous considérons ici sont des éléments de la famille dite des potentiels de Walters sur $\{0,1\}^\mathbb{N}$, qui généralise les potentiels initialement étudiés par F. Hofbauer. Pour ces potentiels, certaines expressions explicites des fonctions propres sont connues. Dans une section finale, nous montrons également que, pour toute suite donnée $d_n \to 0$ de nombres réels dépendant de $n \in \mathbb{N}$, il existe un potentiel $g$ dans la classe définie par Walters possédant une probabilité invariante dont les coefficients de corrélation (pour une certaine fonction observable explicite) sont précisément ces nombres $d_n$.
|
fr
|
Gaussian process (GP) regression is a powerful interpolation technique due to its flexibility in capturing non-linearity. In this paper, we provide a general framework for understanding the frequentist coverage of point-wise and simultaneous Bayesian credible sets in GP regression. As an intermediate result, we develop a Bernstein von-Mises type result under supremum norm in random design GP regression. Identifying both the mean and covariance function of the posterior distribution of the Gaussian process as regularized $M$-estimators, we show that the sampling distribution of the posterior mean function and the centered posterior distribution can be respectively approximated by two population level GPs. By developing a comparison inequality between two GPs, we provide exact characterization of frequentist coverage probabilities of Bayesian point-wise credible intervals and simultaneous credible bands of the regression function. Our results show that inference based on GP regression tends to be conservative; when the prior is under-smoothed, the resulting credible intervals and bands have minimax-optimal sizes, with their frequentist coverage converging to a non-degenerate value between their nominal level and one. As a byproduct of our theory, we show that the GP regression also yields minimax-optimal posterior contraction rate relative to the supremum norm, which provides a positive evidence to the long standing problem on optimal supremum norm contraction rate in GP regression.
|
ຂະບວນການເງື່ອນໄຂກຸສ (Gaussian process: GP) ແມ່ນວິທີການຈັດລຽງທີ່ມີປະສິດທິພາບເນື່ອງຈາກຄວາມຍືດຍຸ່ນໃນການຈັບຮູບແບບທີ່ບໍ່ເປັນເສັ້ນຊື່. ໃນບົດຄວາມນີ້, ພວກເຮົາສະເໜີແບບແຜນທົ່ວໄປໃນການເຂົ້າໃຈການຄຸ້ມຄອງຂອງຊຸດຄວາມໝາຍເຊື່ອຖືໄດ້ຂອງ Bayesian ໃນແບບຈຸດຕໍ່ຈຸດ ແລະ ແບບພ້ອມກັນໃນການກຳນົດຂະບວນການເງື່ອນໄຂກຸສ. ເປັນຜົນລັບກາງ, ພວກເຮົາພັດທະນາຜົນໄດ້ຮັບແບບ Bernstein von-Mises ພາຍໃຕ້ນອກໝາກ (supremum norm) ໃນການກຳນົດຂະບວນການເງື່ອນໄຂກຸດທີ່ມີການອອກແບບແບບສຸ່ມ. ໂດຍການຈຳແນກທັງໝົດທັງສອງຄື ສະເລ່ຍ ແລະ ຟັງຊັ້ນຄວາມແປປວນຂອງການແຈກຢາຍຫຼັງຂອງຂະບວນການເງື່ອນໄຂກຸດເປັນ $M$-estimators ທີ່ຖືກປັບປຸງ, ພວກເຮົາສະແດງໃຫ້ເຫັນວ່າການແຈກຢາຍຕົວຢ່າງຂອງຟັງຊັ້ນສະເລ່ຍຫຼັງ ແລະ ການແຈກຢາຍຫຼັງທີ່ຖືກກຳນົດໃຫ້ຢູ່ສູນກາງ ສາມາດຖືກປະມານໂດຍ GP ສອງຊຸດໃນລະດັບປະຊາກອນ. ໂດຍການພັດທະນາອະສົມຜົນປຽບທຽບລະຫວ່າງ GP ສອງຊຸດ, ພວກເຮົາສະເໜີການອະທິບາຍຢ່າງແນ່ນອນກ່ຽວກັບຄວາມໜາແໜ້ນຂອງການຄຸ້ມຄອງຂອງຊຸດຄວາມໝາຍເຊື່ອຖືໄດ້ຂອງ Bayesian ໃນແບບຈຸດຕໍ່ຈຸດ ແລະ ຊຸດຄວາມໝາຍເຊື່ອຖືໄດ້ພ້ອມກັນຂອງຟັງຊັ້ນກຳນົດ. ຜົນໄດ້ຮັບຂອງພວກເຮົາສະແດງໃຫ້ເຫັນວ່າການສືບຄົ້ນທີ່ອີງໃສ່ຂະບວນການເງື່ອນໄຂກຸດມັກຈະມີແນວໂນ້ມທີ່ຮັກສາຄວາມລະມັດລະວັງ; ເມື່ອກ່ອນໜ້າຖືກປັບລຽບໜ້ອຍເກີນໄປ, ຊຸດຄວາມໝາຍເຊື່ອຖືໄດ້ທີ່ໄດ້ຈະມີຂະໜາດທີ່ເໝາະສົມຕາມມາດຕະຖານ minimax-optimal, ໂດຍທີ່ການຄຸ້ມຄອງຂອງມັນຈະສະທ້ອນຄືນໄປຫາຄ່າທີ່ບໍ່ແຍກຕົວອອກຈາກລະດັບທີ່ຕັ້ງໄວ້ແຕ່ໜ້ອຍກວ່າໜຶ່ງ. ເປັນຜົນຂ້າງຄຽງຂອງທິດສະດີຂອງພວກເຮົາ, ພວກເຮົາສະແດງໃຫ້ເຫັນວ່າການກຳນົດຂະບວນການເງື່ອນໄຂກຸດຍັງໃຫ້ອັດຕາການຫົດຕົວຫຼັງທີ່ເໝາະສົມຕາມມາດຕະຖານ minimax ເມື່ອປຽບທຽບກັບນອກໝາກ (supremum norm), ເຊິ່ງໃຫ້ຫຼັກຖານທີ່ເປັນບວກຕໍ່ບັນຫາທີ່ຍືນຍົງມາດົນກ່ຽວກັບອັດຕາການຫົດຕົວທີ່ເໝາະສົມຕາມນອກໝາກໃນການກຳນົດຂະບວນການເງື່ອນໄຂກຸດ.
|
lo
|
Given $\beta\in(1,2)$ the fat Sierpinski gasket $\mathcal S_\beta$ is the self-similar set in $\mathbb R^2$ generated by the iterated function system (IFS) \[ f_{\beta,d}(x)=\frac{x+d}{\beta},\quad d\in\mathcal A:=\{(0, 0), (1,0), (0,1)\}. \] Then for each point $P\in\mathcal S_\beta$ there exists a sequence $(d_i)\in\mathcal A^\mathbb N$ such that $P=\sum_{i=1}^\infty d_i/\beta^i$, and the infinite sequence $(d_i)$ is called a \emph{coding} of $P$. In general, a point in $\mathcal S_\beta$ may have multiple codings since the overlap region $\mathcal O_\beta:=\bigcup_{c,d\in\mathcal A, c\ne d}f_{\beta,c}(\Delta_\beta)\cap f_{\beta,d}(\Delta_\beta)$ has non-empty interior, where $\Delta_\beta$ is the convex hull of $\mathcal S_\beta$. In this paper we are interested in the invariant set \[ \widetilde{\mathcal U}_\beta:=\left\{\sum_{i=1}^\infty \frac{d_i}{\beta^i}\in \mathcal S_\beta: \sum_{i=1}^\infty\frac{d_{n+i}}{\beta^i}\notin\mathcal O_\beta~\forall n\ge 0\right\}. \] Then each point in $ \widetilde{\mathcal U}_\beta$ has a unique coding. We show that there is a transcendental number $\beta_c\approx 1.55263$ related to the Thue-Morse sequence, such that $\widetilde{\mathcal U}_\beta$ has positive Hausdorff dimension if and only if $\beta>\beta_{c}$. Furthermore, for $\beta=\beta_c$ the set $\widetilde{\mathcal U}_\beta$ is uncountable but has zero Hausdorff dimension, and for $\beta<\beta_c$ the set $\widetilde{\mathcal U}_\beta$ is at most countable. Consequently, we also answer a conjecture of Sidorov (2007). Our strategy is using combinatorics on words based on the lexicographical characterization of $\widetilde{\mathcal U}_\beta$.
|
بالنظر إلى $\beta\in(1,2)$، فإن شكل المشبك السيربينسكي السميك $\mathcal S_\beta$ هو مجموعة متشابهة ذاتيًا في $\mathbb R^2$ تُنشأ بواسطة نظام الدوال المتكررة (IFS)
\[
f_{\beta,d}(x)=\frac{x+d}{\beta},\quad d\in\mathcal A:=\{(0, 0), (1,0), (0,1)\}.
\]
وبالتالي لكل نقطة $P\in\mathcal S_\beta$ توجد متتالية $(d_i)\in\mathcal A^\mathbb N$ بحيث $P=\sum_{i=1}^\infty d_i/\beta^i$، وتُسمى المتتالية اللانهائية $(d_i)$ \emph{ترميزًا} للنقطة $P$. بشكل عام، قد تمتلك نقطة في $\mathcal S_\beta$ ترميزات متعددة نظرًا لأن منطقة التداخل $\mathcal O_\beta:=\bigcup_{c,d\in\mathcal A, c\ne d}f_{\beta,c}(\Delta_\beta)\cap f_{\beta,d}(\Delta_\beta)$ ذات داخلي غير فارغ، حيث $\Delta_\beta$ هو الغلاف المحدب للمجموعة $\mathcal S_\beta$. في هذا البحث نحن مهتمون بالمجموعة الثابتة
\[
\widetilde{\mathcal U}_\beta:=\left\{\sum_{i=1}^\infty \frac{d_i}{\beta^i}\in \mathcal S_\beta: \sum_{i=1}^\infty\frac{d_{n+i}}{\beta^i}\notin\mathcal O_\beta~\forall n\ge 0\right\}.
\]
عندها تمتلك كل نقطة في $\widetilde{\mathcal U}_\beta$ ترميزًا وحيدًا. نُثبت أن هناك عددًا متساميًا $\beta_c\approx 1.55263$ مرتبطًا بمتتالية ثيو-مورس، بحيث تكون لمجموعة $\widetilde{\mathcal U}_\beta$ بُعد هاوسدورف موجب إذا وفقط إذا كان $\beta>\beta_{c}$. علاوة على ذلك، عندما يكون $\beta=\beta_c$ تكون المجموعة $\widetilde{\mathcal U}_\beta$ غير معدودة ولكن بُعدها الهووسدورفي يساوي صفرًا، وعندما يكون $\beta<\beta_c$ تكون المجموعة $\widetilde{\mathcal U}_\beta$ معدودة على الأكثر. وبالتالي، نُجيب أيضًا على حدسية سيدوروف (2007). تتمثل استراتيجيتنا في استخدام التوافقيات على الكلمات استنادًا إلى الصفة المعتمدة على الترتيب المعجمي للمجموعة $\widetilde{\mathcal U}_\beta$.
|
ar
|
A new look at the structure of the electron diffusion region in collisionless magnetic reconnection is presented. The research is based on a particle-in-cell simulation of asymmetric magnetic reconnection, which include a temperature gradient across the current layer in addition to density and magnetic field gradient. We find that none of X-point, flow stagnation point, and local current density peak coincide. Current and energy balance analyses around the flow stagnation point and current density peak show consistently that current dissipation is associated with the divergence of nongyrotropic electron pressure. Furthermore, the same pressure terms, when combined with shear-type gradients of the electron flow velocity, also serve to maintain local thermal energy against convective losses. These effects are similar to those found also in symmetric magnetic reconnection. In addition, we find here significant effects related to the convection of current, which we can relate to a generalized diamagnetic drift by the nongyrotropic pressure divergence. Therefore, only part of the pressure force serves to dissipate the current density. However, the prior conclusion that the role of the reconnection electric field is to maintain the current density, which was obtained for a symmetric system, applies here as well. Finally, we discuss related features of electron distribution function in the EDR.
|
ຂໍ້ມູນໃໝ່ກ່ຽວກັບໂຄງສ້າງຂອງບໍລິເວນການແຜ່ກະຈາຍອິເລັກຕຼອນໃນການເຊື່ອມຕໍ່ແມ່ເຫຼັກທີ່ບໍ່ມີການກະທົບກັນຖືກນຳສະເໜີ. ການຄົ້ນຄວ້ານີ້ອີງໃສ່ການຈຳລອງໂດຍໃຊ້ວິທີ particle-in-cell ສຳລັບການເຊື່ອມຕໍ່ແມ່ເຫຼັກທີ່ບໍ່ສົມດຸນ, ໂດຍລວມເຖິງຄວາມເຂັ້ມຂອງອຸນຫະພູມທີ່ແຕກຕ່າງກັນຂ້າມຊັ້ນກະແສນອກຈາກຄວາມເຂັ້ມຂອງຄວາມໜາແໜ້ນ ແລະ ຄວາມເຂັ້ມຂອງເສັ້ນແຮງດູດ. ພວກເຮົາພົບວ່າຈຸດ X-point, ຈຸດຢຸດການໄຫຼ, ແລະ ຈຸດສູງສຸດຂອງກະແສໄຟຟ້າບໍ່ໄດ້ຢູ່ຮ່ວມກັນ. ການວິເຄາະດຸນຍອດກະແສ ແລະ ພະລັງງານອ້ອມຮອບຈຸດຢຸດການໄຫຼ ແລະ ຈຸດສູງສຸດຂອງກະແສສະແດງໃຫ້ເຫັນຢ່າງຕໍ່ເນື່ອງວ່າການກະຈາຍກະແສກ່ຽວຂ້ອງກັບການແຍກຕົວຂອງກົດດັນອິເລັກຕຼອນທີ່ບໍ່ມີຮູບແວ່ນ. ລວມໄປເຖິງ, ພາກສ່ວນກົດດັນດຽວກັນນີ້, ເມື່ອປະສົມກັບຄວາມເຂັ້ມຂອງຄວາມໄວການໄຫຼຂອງອິເລັກຕຼອນແບບ shear-type, ກໍຍັງເຮັດໜ້າທີ່ຮັກສາພະລັງງານຄວາມຮ້ອນທ້ອງຖິ່ນໄວ້ຕໍ່ກັບການສູນເສຍແບບກະແສ. ຜົນກະທົບເຫຼົ່ານີ້ຄ້າຍຄືກັບຜົນກະທົບທີ່ພົບໃນການເຊື່ອມຕໍ່ແມ່ເຫຼັກທີ່ສົມດຸນ. ນອກຈາກນັ້ນ, ພວກເຮົາຍັງພົບເຫັນຜົນກະທົບທີ່ສຳຄັນທີ່ກ່ຽວຂ້ອງກັບການກະແສທີ່ຖືກກະແສໄຫຼນຳໄປ, ທີ່ພວກເຮົາສາມາດເຊື່ອມໂຍງກັບການເຄື່ອນທີ່ diamagnetic ທົ່ວໄປໂດຍການແຍກຕົວຂອງກົດດັນທີ່ບໍ່ມີຮູບແວ່ນ. ສະນັ້ນ, ມີພຽງແຕ່ສ່ວນໜຶ່ງຂອງແຮງກົດດັນທີ່ເຮັດໜ້າທີ່ກະຈາຍກະແສໄຟຟ້າ. ແຕ່ຢ່າງໃດກໍຕາມ, ຂໍ້ສະຫຼຸບກ່ອນໜ້ານີ້ທີ່ວ່າບົດບາດຂອງໄຟຟ້າເຊື່ອມຕໍ່ແມ່ເຫຼັກແມ່ນເພື່ອຮັກສາກະແສໄຟຟ້າ, ທີ່ໄດ້ຮັບມາຈາກລະບົບທີ່ສົມດຸນ, ກໍຍັງນຳມາໃຊ້ກັບກໍລະນີນີ້ໄດ້. ສຸດທ້າຍ, ພວກເຮົາປຶກສາຫາລືເຖິງຄຸນລັກສະນະທີ່ກ່ຽວຂ້ອງຂອງຟັງຊັ້ນການຈຳແນກອິເລັກຕຼອນໃນ EDR.
|
lo
|
In this paper, channel estimation and data detection for multihop relaying orthogonal frequency division multiplexing (OFDM) system is investigated under time-varying channel. Different from previous works, which highly depend on the statistical information of the doubly-selective channel (DSC) and noise to deliver accurate channel estimation and data detection results, we focus on more practical scenarios with unknown channel orders and Doppler frequencies. Firstly, we integrate the multilink, multihop channel matrices into one composite channel matrix. Then, we formulate the unknown channel using generalized complex exponential basis expansion model (GCE-BEM) with a large oversampling factor to introduce channel sparsity on delay-Doppler domain. To enable the identification of nonzero entries, sparsity enhancing Gaussian distributions with Gamma hyperpriors are adopted. An iterative algorithm is developed under variational inference (VI) framework. The proposed algorithm iteratively estimate the channel, recover the unknown data using Viterbi algorithm and learn the channel and noise statistical information, using only limited number of pilot subcarrier in one OFDM symbol. Simulation results show that, without any statistical information, the performance of the proposed algorithm is very close to that of the optimal channel estimation and data detection algorithm, which requires specific information on system structure, channel tap positions, channel lengths, Doppler shifts as well as noise powers.
|
In questo articolo viene studiata la stima del canale e il rilevamento dei dati per un sistema multihop di relaying con multiplexing ortogonale per divisione di frequenza (OFDM) in presenza di canale variabile nel tempo. A differenza dei lavori precedenti, che dipendono fortemente dalle informazioni statistiche del canale doppiamente selettivo (DSC) e del rumore per ottenere risultati accurati di stima del canale e rilevamento dei dati, ci si concentra su scenari più realistici con ordini del canale e frequenze Doppler sconosciute. In primo luogo, si integrano le matrici del canale multicollegamento e multihop in un'unica matrice del canale composito. Successivamente, si formula il canale sconosciuto utilizzando il modello generalizzato di espansione in base esponenziale complessa (GCE-BEM) con un elevato fattore di sovracampionamento per introdurre una sparsità del canale nel dominio ritardo-Doppler. Al fine di consentire l'identificazione delle entrate diverse da zero, si adottano distribuzioni gaussiane di miglioramento della sparsità con iperpriori Gamma. Viene sviluppato un algoritmo iterativo nel quadro dell'inferenza variazionale (VI). L'algoritmo proposto stima iterativamente il canale, recupera i dati sconosciuti mediante l'algoritmo di Viterbi e apprende le informazioni statistiche del canale e del rumore, utilizzando solo un numero limitato di sottocarrier pilota in un simbolo OFDM. I risultati delle simulazioni mostrano che, in assenza di qualsiasi informazione statistica, le prestazioni dell'algoritmo proposto sono molto vicine a quelle dell'algoritmo ottimale di stima del canale e rilevamento dei dati, il quale richiede informazioni specifiche sulla struttura del sistema, sulle posizioni dei tap del canale, sulle lunghezze del canale, sugli spostamenti Doppler e sulle potenze del rumore.
|
it
|
Cluster formation and gas dynamics in the central regions of barred galaxies are not well understood. This paper reviews the environment of three 10^7 Msun clusters near the inner Lindblad resonance of the barred spiral NGC 1365. The morphology, mass, and flow of HI and CO gas in the spiral and barred regions are examined for evidence of the location and mechanism of cluster formation. The accretion rate is compared with the star formation rate to infer the lifetime of the starburst. The gas appears to move from inside corotation in the spiral region to looping filaments in the interbar region at a rate of ~6 Msun/yr before impacting the bar dustlane somewhere along its length. The gas in this dustlane moves inward, growing in flux as a result of the accretion to ~40 Msun/yr near the ILR. This inner rate exceeds the current nuclear star formation rate by a factor of 4, suggesting continued buildup of nuclear mass for another ~0.5 Gyr. The bar may be only 1-2 Gyr old. Extrapolating the bar flow back in time, we infer that the clusters formed in the bar dustlane outside the central dust ring at a position where an interbar filament currently impacts the lane. The ram pressure from this impact is comparable to the pressure in the bar dustlane, and both are comparable to the pressure in the massive clusters. Impact triggering is suggested. The isothermal assumption in numerical simulations seems inappropriate for the rare fraction parts of spiral and bar gas flows. The clusters have enough lower-mass counterparts to suggest they are part of a normal power law mass distribution. Gas trapping in the most massive clusters could explain their [NeII] emission, which is not evident from the lower-mass clusters nearby.
|
La formación de cúmulos y la dinámica del gas en las regiones centrales de galaxias con barra no se comprenden bien. Este artículo revisa el entorno de tres cúmulos de 10^7 Msol cerca de la resonancia de Lindblad interna de la espiral con barra NGC 1365. Se examina la morfología, la masa y el flujo de gas HI y CO en las regiones espiral y con barra en busca de evidencia sobre la ubicación y el mecanismo de formación de cúmulos. La tasa de acreción se compara con la tasa de formación estelar para inferir la duración del brote de formación estelar. El gas parece moverse desde el interior de la corotación en la región espiral hacia filamentos en bucle en la región interbarra a una tasa de ~6 Msol/año antes de impactar en alguna parte a lo largo de su longitud con el plano de polvo de la barra. El gas en este plano de polvo se mueve hacia el interior, aumentando su flujo como resultado de la acreción hasta ~40 Msol/año cerca de la ILR. Esta tasa interna supera en un factor de 4 la tasa actual de formación estelar nuclear, lo que sugiere una acumulación continua de masa nuclear durante otros ~0,5 Gyr. Es posible que la barra tenga solo 1-2 Gyr de edad. Extrapolando hacia atrás en el tiempo el flujo de la barra, inferimos que los cúmulos se formaron en el plano de polvo de la barra fuera del anillo central de polvo, en una posición donde actualmente un filamento interbarra impacta contra el plano. La presión dinámica de este impacto es comparable a la presión en el plano de polvo de la barra, y ambas son comparables a la presión en los cúmulos masivos. Se sugiere un desencadenamiento por impacto. La suposición de isoterma en simulaciones numéricas parece inapropiada para las fracciones raras de flujos de gas en regiones espirales y de barra. Los cúmulos tienen suficientes contrapartes de menor masa como para sugerir que forman parte de una distribución normal de masa según una ley de potencias. El atrapamiento de gas en los cúmulos más masivos podría explicar su emisión [NeII], que no es evidente en los cúmulos de menor masa cercanos.
|
es
|
In order to study the dependence of the coercive force of sintered magnets on temperature, nucleation and domain wall propagation at the grain boundary are studied as rate-determining processes of the magnetization reversal phenomena in magnets consisting of bulk hard magnetic grains contacting via grain boundaries of a soft magnetic material. These systems have been studied analytically for a continuum model at zero temperature (A. Sakuma, et al. J. Mag. Mag. Mat. {\bf 84} 52 (1990)). In the present study, the temperature dependence is studied by making use of the stochastic Landau-Lifshitz-Gilbert equation at finite temperatures. In particular, the threshold fields for nucleation and domain wall propagation are obtained as functions of ratios of magnetic interactions and anisotropies of the soft and hard magnets for various temperatures. It was found that the threshold field for domain wall propagation is robust against thermal fluctuations, while that for nucleation is fragile. The microscopic mechanisms of the observed temperature dependence are discussed.
|
A fim de estudar a dependência da força coercitiva de ímãs sinterizados com a temperatura, a nucleação e a propagação da parede de domínio no contorno de grão são analisadas como processos determinantes da velocidade nos fenômenos de inversão de magnetização em ímãs constituídos por grãos magnéticos duros em contato através de contornos de grão de um material magnético mole. Esses sistemas foram estudados analiticamente para um modelo contínuo à temperatura zero (A. Sakuma et al. J. Mag. Mag. Mat. {\bf 84} 52 (1990)). No presente estudo, a dependência com a temperatura é investigada utilizando a equação estocástica de Landau-Lifshitz-Gilbert a temperaturas finitas. Em particular, os campos limiares para nucleação e propagação da parede de domínio são obtidos como funções das razões entre interações magnéticas e anisotropias dos ímãs mole e duro para diversas temperaturas. Verificou-se que o campo limiar para a propagação da parede de domínio é robusto contra flutuações térmicas, enquanto que o campo limiar para nucleação é frágil. Os mecanismos microscópicos da dependência com a temperatura observada são discutidos.
|
pt
|
Microgrids are increasingly recognized as a key technology for the integration of distributed energy resources into the power network, allowing local clusters of load and distributed energy resources to operate autonomously. However, microgrid operation brings new challenges, especially in islanded operation as frequency and voltage control are no longer provided by large rotating machines. Instead, the power converters in the microgrid must coordinate to regulate the frequency and voltage and ensure stability. We consider the problem of designing controllers to achieve these objectives. Using passivity theory to derive decentralized stability conditions for the microgrid, we propose a control design method for grid-forming inverters. For the analysis we use higher-order models for the inverters and also advanced dynamic models for the lines with an arbitrarily large number of states. By satisfying the decentralized condition formulated, plug-and-play operation can be achieved with guaranteed stability, and performance can also be improved by incorporating this condition as a constraint in corresponding optimization problems formulated. In addition, our control design can improve the power sharing properties of the microgrid compared to previous non-droop approaches. Finally, realistic simulations confirm that the controller design improves the stability and performance of the power network.
|
Ang mga mikrogrids ay unti-unting kinikilala bilang mahalagang teknolohiya para sa pagsasama ng mga mapagkukunang distribusyong enerhiya sa network ng kuryente, na nagbibigay-daan sa lokal na mga grupo ng karga at mapagkukunang distribusyong enerhiya na kumilos nang autonomo. Gayunpaman, ang operasyon ng mikrogrid ay nagdudulot ng mga bagong hamon, lalo na sa operasyong naka-island dahil ang kontrol sa dalas at boltahe ay hindi na ibinibigay ng malalaking umiikot na makina. Sa halip, ang mga power converter sa loob ng mikrogrid ang dapat mag-koordina upang mapanatili ang dalas at boltahe at matiyak ang katatagan. Pinag-aaralan namin ang problema ng pagdidisenyo ng mga kontrolador upang makamit ang mga layuning ito. Gamit ang teorya ng passivity upang maihanap ang mga kondisyon ng desentralisadong katatagan para sa mikrogrid, iminumungkahi namin ang isang paraan sa pagdidisenyo ng kontrol para sa mga grid-forming na inverter. Para sa pagsusuri, ginagamit namin ang mga modelo ng mas mataas na orden para sa mga inverter at pati na rin ang mga napapanahong dinamikong modelo para sa mga linya na may arbitraryong malaking bilang ng mga estado. Sa pamamagitan ng pagsunod sa nabuong desentralisadong kondisyon, maaaring makamit ang plug-and-play na operasyon na may garantisadong katatagan, at maaari ring mapabuti ang pagganap sa pamamagitan ng pagsasama ng kondisyong ito bilang isang paghihigpit sa mga kaukulang problema sa pag-optimize. Bukod dito, ang aming disenyo ng kontrol ay maaaring mapabuti ang mga katangian ng pagbabahagi ng kuryente ng mikrogrid kumpara sa mga nakaraang non-droop na pamamaraan. Sa huli, ang mga realistikong simulation ay nagpapatunay na ang disenyo ng kontrolador ay mapapabuti ang katatagan at pagganap ng network ng kuryente.
|
tl
|
Governments and cities around the world are currently facing rapid growth in the use of Electric Vehicles and therewith the need for Charging Infrastructure. For these cities, the struggle remains how to further roll out charging infrastructure in the most efficient way, both in terms of cost and use. Forecasting models are not able to predict more long-term developments, and as such more complex simulation models offer opportunities to simulate various scenarios. Agent based simulation models provide insight into the effects of incentives and roll-out strategies before they are implemented in practice and thus allow for scenario testing. This paper describes the build up of an agent based model that enables policy makers to anticipate on charging infrastructure development. The model is able to simulate charging transactions of individual users and is both calibrated and validated using a dataset of charging transactions from the public charging infrastructure of the four largest cities in the Netherlands.
|
Chính phủ và các thành phố trên khắp thế giới hiện đang phải đối mặt với sự gia tăng nhanh chóng trong việc sử dụng Xe điện và do đó là nhu cầu về cơ sở hạ tầng sạc. Đối với các thành phố này, thách thức vẫn nằm ở việc làm thế nào để tiếp tục triển khai cơ sở hạ tầng sạc một cách hiệu quả nhất, cả về chi phí lẫn mức độ sử dụng. Các mô hình dự báo không thể dự đoán được các diễn biến dài hạn hơn, và do đó các mô hình mô phỏng phức tạp hơn mang lại cơ hội để mô phỏng nhiều kịch bản khác nhau. Các mô hình mô phỏng dựa trên tác nhân (agent-based) cung cấp cái nhìn sâu sắc về tác động của các chính sách khuyến khích và các chiến lược triển khai trước khi chúng được áp dụng thực tế, từ đó cho phép thử nghiệm các kịch bản. Bài báo này mô tả việc xây dựng một mô hình dựa trên tác nhân, cho phép các nhà hoạch định chính sách tiên liệu trước sự phát triển của cơ sở hạ tầng sạc. Mô hình này có khả năng mô phỏng các giao dịch sạc của từng người dùng cá nhân và được hiệu chỉnh cũng như xác thực bằng một bộ dữ liệu về các giao dịch sạc từ cơ sở hạ tầng sạc công cộng của bốn thành phố lớn nhất tại Hà Lan.
|
vi
|
Governments and cities around the world are currently facing rapid growth in the use of Electric Vehicles and therewith the need for Charging Infrastructure. For these cities, the struggle remains how to further roll out charging infrastructure in the most efficient way, both in terms of cost and use. Forecasting models are not able to predict more long-term developments, and as such more complex simulation models offer opportunities to simulate various scenarios. Agent based simulation models provide insight into the effects of incentives and roll-out strategies before they are implemented in practice and thus allow for scenario testing. This paper describes the build up of an agent based model that enables policy makers to anticipate on charging infrastructure development. The model is able to simulate charging transactions of individual users and is both calibrated and validated using a dataset of charging transactions from the public charging infrastructure of the four largest cities in the Netherlands.
|
تواجه الحكومات والمدن في جميع أنحاء العالم حاليًا نموًا سريعًا في استخدام المركبات الكهربائية، وبالتالي الحاجة إلى بنية تحتية للشحن. ولهذه المدن، تظل التحديات قائمة حول كيفية المضي قدمًا في نشر البنية التحتية للشحن بالطريقة الأكثر كفاءة من حيث التكلفة والاستخدام. لا يمكن لنموذج التنبؤ أن يتوقع التطورات على المدى الطويل، وبالتالي فإن النماذج المحاكاة الأكثر تعقيدًا توفر فرصًا لمحاكاة سيناريوهات مختلفة. تُمكّن نماذج المحاكاة القائمة على العوامل من فهم آثار الحوافز واستراتيجيات النشر قبل تنفيذها عمليًا، وبالتالي تتيح اختبار السيناريوهات. يصف هذا البحث بناء نموذج قائم على العوامل يمكن صانعي السياسات من التنبؤ بتطوير البنية التحتية للشحن. يمكن للنموذج محاكاة معاملات الشحن الخاصة بكل مستخدم على حدة، وقد تم معايرته والتحقق من صحته باستخدام مجموعة بيانات من معاملات الشحن الخاصة بالبنية التحتية العامة للشحن في أكبر أربع مدن في هولندا.
|
ar
|
Thermodynamics relies on the possibility to describe systems composed of a large number of constituents in terms of few macroscopic variables. Its foundations are rooted into the paradigm of statistical mechanics, where thermal properties originate from averaging procedures which smoothen out local details. While undoubtedly successful, elegant and formally correct, this approach carries over an operational problem: what is the precision at which such variables are inferred, when technical/practical limitations restrict our capabilities to local probing? Here we introduce the local quantum thermal susceptibility, a quantifier for the best achievable accuracy for temperature estimation via local measurements. Our method relies on basic concepts of quantum estimation theory, providing an operative strategy to address the local thermal response of arbitrary quantum systems at equilibrium. At low temperatures it highlights the local distinguishability of the ground state from the excited sub-manifolds, thus providing a method to locate quantum phase transitions.
|
열역학은 다수의 구성 요소로 이루어진 시스템을 소수의 거시적 변수들로 기술할 수 있다는 가능성에 의존한다. 그 기초는 통계역학의 패러다임에 뿌리를 두고 있으며, 여기서 열적 성질은 국소적 세부 사항을 평균화하여 매끄럽게 만드는 절차에서 비롯된다. 분명히 성공적이며 우아하고 형식적으로 정확하지만, 이러한 접근법은 실용적인 문제를 수반한다. 즉, 기술적 또는 실용적인 제약으로 인해 국소적인 측정만이 가능할 때, 이러한 변수들을 추정하는 정밀도는 어느 정도인가? 본 연구에서는 국소 양자 열적 감수도(local quantum thermal susceptibility)를 도입하는데, 이는 국소 측정을 통해 온도를 추정할 때 달성 가능한 최고 정확도를 나타내는 척도이다. 우리 방법은 양자 추정 이론의 기본 개념에 기반하여, 임의의 양자 시스템의 국소 열적 반응을 다룰 수 있는 실용적인 전략을 제공한다. 낮은 온도에서 이 척도는 기저 상태와 들뜬 부분 다양체 사이의 국소적 구분 가능성을 강조하며, 이를 통해 양자 위상 전이를 위치할 수 있는 방법을 제공한다.
|
ko
|
We present a model of worldwide crisis contagion based on the Google matrix analysis of the world trade network obtained from the UN Comtrade database. The fraction of bankrupted countries exhibits an \textit{on-off} phase transition governed by a bankruptcy threshold $\kappa$ related to the trade balance of the countries. For $\kappa>\kappa_c$, the contagion is circumscribed to less than 10\% of the countries, whereas, for $\kappa<\kappa_c$, the crisis is global with about 90\% of the countries going to bankruptcy. We measure the total cost of the crisis during the contagion process. In addition to providing contagion scenarios, our model allows to probe the structural trading dependencies between countries. For different networks extracted from the world trade exchanges of the last two decades, the global crisis comes from the Western world. In particular, the source of the global crisis is systematically the Old Continent and The Americas (mainly US and Mexico). Besides the economy of Australia, those of Asian countries, such as China, India, Indonesia, Malaysia and Thailand, are the last to fall during the contagion. Also, the four BRIC are among the most robust countries to the world trade crisis.
|
Apresentamos um modelo de contágio de crise mundial baseado na análise da matriz Google da rede de comércio mundial obtida a partir da base de dados UN Comtrade. A fração de países falidos exibe uma transição de fase \textit{liga-desliga} governada por um limiar de falência $\kappa$ relacionado ao saldo comercial dos países. Para $\kappa>\kappa_c$, o contágio restringe-se a menos de 10\% dos países, enquanto, para $\kappa<\kappa_c$, a crise é global, com cerca de 90\% dos países entrando em falência. Medimos o custo total da crise durante o processo de contágio. Além de fornecer cenários de contágio, nosso modelo permite investigar as dependências comerciais estruturais entre os países. Para diferentes redes extraídas das trocas comerciais mundiais das últimas duas décadas, a crise global tem origem no mundo ocidental. Em particular, a fonte da crise global é sistematicamente o Velho Continente e as Américas (principalmente EUA e México). Além da economia da Austrália, as dos países asiáticos, como China, Índia, Indonésia, Malásia e Tailândia, são as últimas a colapsar durante o contágio. Além disso, os quatro países do BRIC estão entre os mais resistentes à crise do comércio mundial.
|
pt
|
OSD PSE is the Indonesian Government Certification Authority (CA) for National e-Procurement System and later named OSD PSE G2. It has a unique hierarchical structure under the OSD Lemsaneg. As an Issuing CA, the OSD PSE G2 publishes and guarantee the quality of the Certificate Policy and Certification Practice Statement (CP-CPS) in order to gain the PKI user trustworthy. In this article, we analyze the CP-CPS version 1.0 that published by OSD PSE G2. For this purpose, we apply the methodology of PKI Assessment Guidelines (PAG). The quality assessment of this CP-CPS, including its compliance to the related reference/standard, namely: CP OSD Lemsaneg v.1.1; RFC 3647; and CA Business Practice Disclosure Principle on Trust Service Principles and Criteria for Certification Authorities (BPDP-TSPCCA) version 2.0. We finally found that the CP-CPS version 1.0 does not comply with related standard and reference. Hence, the CP-CPS need to be updated following the current condition of OSD PSE G2.
|
A OSD PSE é a Autoridade Certificadora (AC) do Governo Indonésio para o Sistema Nacional de Licitações Eletrônicas, posteriormente denominada OSD PSE G2. Possui uma estrutura hierárquica única sob a OSD Lemsaneg. Como uma AC emissora, a OSD PSE G2 publica e garante a qualidade da Política de Certificação e da Declaração de Práticas de Certificação (PC-DPC) com o objetivo de obter a confiança dos usuários da infraestrutura de chaves públicas (ICP). Neste artigo, analisamos a versão 1.0 da PC-DPC publicada pela OSD PSE G2. Para tanto, aplicamos a metodologia das Diretrizes de Avaliação de ICP (PAG). A avaliação da qualidade desta PC-DPC inclui sua conformidade com as referências/normas relacionadas, a saber: PC OSD Lemsaneg v.1.1; RFC 3647; e Princípios de Divulgação das Práticas Comerciais de AC sobre Princípios e Critérios de Serviços Confiáveis para Autoridades Certificadoras (BPDP-TSPCCA) versão 2.0. Concluímos, por fim, que a PC-DPC versão 1.0 não está em conformidade com as normas e referências relacionadas. Assim, a PC-DPC precisa ser atualizada de acordo com a situação atual da OSD PSE G2.
|
pt
|
Super-resolution fluorescence microscopy, with a resolution beyond the diffraction limit of light, has become an indispensable tool to directly visualize biological structures in living cells at a nanometer-scale resolution. Despite advances in high-density super-resolution fluorescent techniques, existing methods still have bottlenecks, including extremely long execution time, artificial thinning and thickening of structures, and lack of ability to capture latent structures. Here we propose a novel deep learning guided Bayesian inference approach, DLBI, for the time-series analysis of high-density fluorescent images. Our method combines the strength of deep learning and statistical inference, where deep learning captures the underlying distribution of the fluorophores that are consistent with the observed time-series fluorescent images by exploring local features and correlation along time-axis, and statistical inference further refines the ultrastructure extracted by deep learning and endues physical meaning to the final image. Comprehensive experimental results on both real and simulated datasets demonstrate that our method provides more accurate and realistic local patch and large-field reconstruction than the state-of-the-art method, the 3B analysis, while our method is more than two orders of magnitude faster. The main program is available at https://github.com/lykaust15/DLBI
|
Kırınım sınırının ötesinde bir çözünürlüğe sahip olan süper çözünürlüklü floresans mikroskopi, canlı hücrelerdeki biyolojik yapıları nanometre ölçeğinde doğrudan görselleştirmek için vazgeçilmez bir araç haline gelmiştir. Yüksek yoğunluklu süper çözünürlüklü floresan tekniklerinde ilerlemeler kaydedilmesine rağmen mevcut yöntemler hâlâ aşırı uzun çalışma süresi, yapıların yapay olarak incelmesi ve kalınlaşması ile gizli yapıları yakalayamama gibi darboğazlara sahiptir. Bu çalışmada, yüksek yoğunluklu floresan görüntülerin zaman serisi analizi için yeni bir derin öğrenme destekli Bayes çıkarımı yaklaşımı, DLBI, öneriyoruz. Yöntemimiz derin öğrenmenin ve istatistiksel çıkarımın gücünü birleştirir; derin öğrenme, yerel özellikleri ve zaman ekseni boyunca korelasyonu inceleyerek gözlemlenen zaman serisi floresan görüntülerine uygun floroforların temel dağılımını yakalar, istatistiksel çıkarım ise derin öğrenme tarafından çıkarılan ultrastruktuру daha da iyileştirir ve nihai görüntüye fiziksel anlam kazandırır. Gerçek ve simüle edilmiş veri kümeleri üzerinde yapılan kapsamlı deneysel sonuçlar, önerilen yöntemimizin 3B analizi olarak bilinen en gelişmiş yönteme kıyasla daha doğru ve gerçekçi yerel alan ve geniş alanlı yeniden inşaa sağladığını göstermektedir. Ayrıca yöntemimiz en az iki kat daha hızlıdır. Ana program şu adresten ulaşılabilir: https://github.com/lykaust15/DLBI
|
tr
|
Recognition of human actions and associated interactions with objects and the environment is an important problem in computer vision due to its potential applications in a variety of domains. The most versatile methods can generalize to various environments and deal with cluttered backgrounds, occlusions, and viewpoint variations. Among them, methods based on graph convolutional networks that extract features from the skeleton have demonstrated promising performance. In this paper, we propose a novel Spatio-Temporal Pyramid Graph Convolutional Network (ST-PGN) for online action recognition for ergonomic risk assessment that enables the use of features from all levels of the skeleton feature hierarchy. The proposed algorithm outperforms state-of-art action recognition algorithms tested on two public benchmark datasets typically used for postural assessment (TUM and UW-IOM). We also introduce a pipeline to enhance postural assessment methods with online action recognition techniques. Finally, the proposed algorithm is integrated with a traditional ergonomic risk index (REBA) to demonstrate the potential value for assessment of musculoskeletal disorders in occupational safety.
|
ການຮັບຮູ້ການກະທຳຂອງມະນຸດ ແລະ ການມີສ່ວນຮ່ວມທີ່ກ່ຽວຂ້ອງກັບວັດຖຸ ແລະ ສິ່ງແວດລ້ອມ ແມ່ນບັນຫາສຳຄັນໃນດ້ານວິໄສທັດຂອງຄອມພິວເຕີ ເນື່ອງຈາກການນຳໃຊ້ທີ່ມີ»ຄວາມເປັນໄປໄດ້ໃນດ້ານຕ່າງໆ. ວິທີການທີ່ມີຄວາມຫຼາກຫຼາຍທີ່ສຸດສາມາດນຳໃຊ້ໄດ້ໃນສິ່ງແວດລ້ອມຕ່າງໆ ແລະ ຈັດການກັບພື້ນຖານທີ່ມີຄວາມສັບສົນ, ການບັງກັ້ນ ແລະ ການປ່ຽນແປງມຸມມອງ. ໃນຈຳນວນນີ້, ວິທີການທີ່ອີງໃສ່ເຄືອຂ່າຍກຣາຟຄອນໂວລູຊັ່ນ (graph convolutional networks) ທີ່ສະກັດເອົາຄຸນລັກສະນະຈາກຂະໜານ (skeleton) ໄດ້ສະແດງໃຫ້ເຫັນເຖິງຜົນງານທີ່ດີ. ໃນບົດຄວາມນີ້, ພວກເຮົາຂໍເສີມເຕີມເຄືອຂ່າຍກຣາຟຄອນໂວລູຊັ່ນຊັ້ນພື້ນທີ່-ເວລາ (Spatio-Temporal Pyramid Graph Convolutional Network - ST-PGN) ໃໝ່ ສຳລັບການຮັບຮູ້ການກະທຳແບບອອນໄລນ໌ ສຳລັບການປະເມີນຄວາມສ່ຽງດ້ານໂອໂລຈິກ (ergonomic risk assessment) ທີ່ເຮັດໃຫ້ສາມາດນຳໃຊ້ຄຸນລັກສະນະຈາກທຸກລະດັບຂອງຊັ້ນຂອງຄຸນລັກສະນະຂະໜານ. ອະລະກິດທີ່ສະເໜີນີ້ໃຫ້ຜົນງານດີກວ່າອະລະກິດການຮັບຮູ້ການກະທຳລະດັບສູງສຸດທີ່ຖືກທົດສອບໃນຊຸດຂໍ້ມູນສອງຊຸດທີ່ເປັນທີ່ນິຍົມໃຊ້ສຳລັບການປະເມີນທ່າທີ (TUM ແລະ UW-IOM). ພວກເຮົາຍັງນຳສະເໜີເສັ້ນທາງການປະມວນຜົນເພື່ອປັບປຸງວິທີການປະເມີນທ່າທີ ດ້ວຍເຕັກນິກການຮັບຮູ້ການກະທຳແບບອອນໄລນ໌. ສຸດທ້າຍ, ອະລະກິດທີ່ສະເໜີນີ້ຖືກຜະສົມເຂົ້າກັບດັດສະນີຄວາມສ່ຽງດ້ານໂອໂລຈິກແບບດັ້ງເດີມ (REBA) ເພື່ອສະແດງໃຫ້ເຫັນເຖິງຄຸນຄ່າທີ່ເປັນໄປໄດ້ໃນການປະເມີນອຸບັດຕິເຫດຂອງລະບົບກ້າມເນື້ອ ແລະ ໂລກຂໍ້ເຂົ່າໃນຄວາມປອດໄພດ້ານອາຊີບ.
|
lo
|
We have detected, for the first time, Cepheid variables in the Sculptor Group spiral galaxy NGC 7793. From wide-field images obtained in the optical V and I bands on 56 nights in 2003-2005, we have discovered 17 long-period (24-62 days) Cepheids whose periods and mean magnitudes define tight period-luminosity relations. We use the (V-I) Wesenheit index to determine a reddening-free true distance modulus to NGC 7793 of 27.68 +- 0.05 mag (internal error) +- 0.08 mag (systematic error). The comparison of the reddened distance moduli in V and I with the one derived from the Wesenheit magnitude indicates that the Cepheids in NGC 7793 are affected by an average total reddening of E(B-V)=0.08 mag, 0.06 of which is produced inside the host galaxy. As in the earlier Cepheid studies of the Araucaria Project, the reported distance is tied to an assumed LMC distance modulus of 18.50. The quoted systematic uncertainty takes into account effects like blending and possible inhomogeneous filling of the Cepheid instability strip on the derived distance. The reported distance value does not depend on the (unknown) metallicity of the Cepheids according to recent theoretical and empirical results. Our Cepheid distance is shorter, but within the errors consistent with the distance to NGC 7793 determined earlier with the TRGB and Tully-Fisher methods. The NGC 7793 distance of 3.4 Mpc is almost identical to the one our project had found from Cepheid variables for NGC 247, another spiral member of the Sculptor Group located close to NGC 7793 on the sky. Two other conspicuous spiral galaxies in the Sculptor Group, NGC 55 and NGC 300, are much nearer (1.9 Mpc), confirming the picture of a very elongated structure of the Sculptor Group in the line of sight put forward by Jerjen et al. and others.
|
Detectamos, pela primeira vez, variáveis Cefeidas na galáxia espiral NGC 7793 do Grupo Escultor. A partir de imagens de campo amplo obtidas nas bandas ópticas V e I em 56 noites entre 2003 e 2005, descobrimos 17 Cefeidas de longo período (24-62 dias), cujos períodos e magnitudes médias definem relações período-luminosidade bem definidas. Utilizamos o índice Wesenheit (V-I) para determinar um módulo de distância verdadeiro livre de avermelhamento para NGC 7793 de 27,68 ± 0,05 mag (erro interno) ± 0,08 mag (erro sistemático). A comparação dos módulos de distância avermelhados nas bandas V e I com aquele derivado da magnitude Wesenheit indica que as Cefeidas em NGC 7793 são afetadas por um avermelhamento total médio de E(B-V) = 0,08 mag, dos quais 0,06 são produzidos no interior da galáxia hospedeira. Como nos estudos anteriores de Cefeidas do Projeto Araucária, a distância relatada está vinculada ao módulo de distância assumido de 18,50 para a Nuvem Grande de Magalhães. A incerteza sistemática citada leva em conta efeitos como mistura de luz (blending) e possível preenchimento não homogêneo da faixa de instabilidade das Cefeidas na distância derivada. O valor de distância relatado não depende da metalicidade (desconhecida) das Cefeidas, de acordo com resultados teóricos e empíricos recentes. Nossa distância baseada em Cefeidas é menor, mas compatível dentro dos erros com a distância a NGC 7793 determinada anteriormente pelos métodos TRGB e Tully-Fisher. A distância de 3,4 Mpc para NGC 7793 é quase idêntica à encontrada pelo nosso projeto a partir de variáveis Cefeidas para NGC 247, outra galáxia espiral membro do Grupo Escultor localizada perto de NGC 7793 no céu. Outras duas galáxias espirais proeminentes no Grupo Escultor, NGC 55 e NGC 300, estão muito mais próximas (1,9 Mpc), confirmando o quadro de uma estrutura altamente alongada do Grupo Escultor ao longo da linha de visada, proposto por Jerjen et al. e outros.
|
pt
|
We report the results of an analysis of the redshift power spectrum $P^S(k,\mu)$ in three typical Cold Dark Matter (CDM) cosmological models, where $\mu$ is the cosine of the angle between the wave vector and the line-of-sight. Two distinct biased tracers derived from the primordial density peaks of Bardeen et al. and the cluster-underweight model of Jing, Mo, & B\"orner are considered in addition to the pure dark matter models. Based on a large set of high resolution simulations, we have measured the redshift power spectrum for the three tracers from the linear to the nonlinear regime. We investigate the validity of the relation - guessed from linear theory - in the nonlinear regime $$ P^S(k,\mu)=P^R(k)[1+\beta\mu^2]^2D(k,\mu,\sigma_{12}(k)), $$ where $P^R(k)$ is the real space power spectrum, and $\beta$ equals $\Omega_0^{0.6}/b_l$. The damping function $D$ which should generally depend on $k$, $\mu$, and $\sigma_{12}(k)$, is found to be a function of only one variable $k\mu\sigma_{12}(k)$. This scaling behavior extends into the nonlinear regime, while $D$ can be accurately expressed as a Lorentz function - well known from linear theory - for values $D > 0.1$. The difference between $\sigma_{12}(k)$ and the pairwise velocity dispersion defined by the 3-D peculiar velocity of the simulations (taking $r=1/k$) is about 15%. Therefore $\sigma_{12}(k)$ is a good indicator of the pairwise velocity dispersion. The exact functional form of $D$ depends on the cosmological model and on the bias scheme. We have given an accurate fitting formula for the functional form of $D$ for the models studied.
|
Hlásíme výsledky analýzy výkonového spektra červeného posuvu $P^S(k,\mu)$ ve třech typických kosmologických modelech studené temné hmoty (CDM), kde $\mu$ je kosinus úhlu mezi vlnovým vektorem a směrem pohledu. Kromě čistých modelů temné hmoty uvažujeme dva různé nesymetrické trasy odvozené od prvořadých hustotních maxim Bardeena et al. a modelu chudého na shluky Jinga, Mo a B\"ornera. Na základě rozsáhlé sady simulací s vysokým rozlišením jsme změřili výkonové spektrum červeného posuvu pro tři trasy od lineárního až po nelineární režim. Zkoumáme platnost vztahu – odhadnutého z lineární teorie – v nelineárním režimu $$ P^S(k,\mu)=P^R(k)[1+\beta\mu^2]^2D(k,\mu,\sigma_{12}(k)), $$ kde $P^R(k)$ je výkonové spektrum v reálném prostoru a $\beta$ se rovná $\Omega_0^{0.6}/b_l$. Zjistili jsme, že tlumicí funkce $D$, která obecně závisí na $k$, $\mu$ a $\sigma_{12}(k)$, je funkcí pouze jedné proměnné $k\mu\sigma_{12}(k)$. Toto škálovací chování přetrvává i do nelineárního režimu, zatímco $D$ lze přesně vyjádřit jako Lorentzovu funkci – dobře známou z lineární teorie – pro hodnoty $D > 0.1$. Rozdíl mezi $\sigma_{12}(k)$ a disperzí párové rychlosti definovanou 3D vlastní rychlostí simulací (s $r=1/k$) činí přibližně 15 %. Proto $\sigma_{12}(k)$ je dobrým ukazatelem disperze párové rychlosti. Přesný tvar funkce $D$ závisí na kosmologickém modelu a schématu nesymetrie. Pro studované modely jsme uvedli přesný aproximační vzorec pro funkční tvar $D$.
|
cs
|
To improve the system throughput, this paper proposes a network non-orthogonal multiple access (N-NOMA) technique for the uplink coordinated multi-point transmission (CoMP). In the considered scenario, multiple base stations collaborate with each other to serve a single user, referred to as the CoMP user, which is the same as for conventional CoMP. However, unlike conventional CoMP, each base station in N-NOMA opportunistically serves an extra user, referred to as the NOMA user, while serving the CoMP user at the same bandwidth. The CoMP user is typically located far from the base stations, whereas users close to the base stations are scheduled as NOMA users. Hence, the channel conditions of the two kind of users are very distinctive, which facilitates the implementation of NOMA. Compared to the conventional orthogonal multiple access based CoMP scheme, where multiple base stations serve a single CoMP user only, the proposed N-NOMA scheme can support larger connectivity by serving the extra NOMA users, and improve the spectral efficiency by avoiding the CoMP user solely occupying the spectrum. A stochastic geometry approach is applied to model the considered N-NOMA scenario as a Poisson cluster process, based on which closed-form analytical expressions for outage probabilities and ergodic rates are obtained. Numerical results are presented to show the accuracy of the analytical results and also demonstrate the superior performance of the proposed N-NOMA scheme.
|
Za účelem zlepšení propustnosti systému navrhuje tento článek techniku sítě s neortogonálním víceuzivatelským přístupem (N-NOMA) pro vícebodový přenos (CoMP) v odchozím směru. V uvažovaném scénáři více základnových stanic spolupracuje navzájem při obsluze jednoho uživatele, označovaného jako uživatel CoMP, což je stejné jako u klasického CoMP. Na rozdíl od klasického CoMP však každá základnová stanice v N-NOMA příležitostně obsluhuje dodatečného uživatele, označovaného jako uživatel NOMA, a to současně se zpracováním uživatele CoMP ve stejném frekvenčním pásmu. Uživatel CoMP je typicky umístěn daleko od základnových stanic, zatímco uživatelé blízko základnových stanic jsou plánováni jako uživatelé NOMA. Podmínky kanálu u těchto dvou typů uživatelů jsou proto velmi odlišné, což usnadňuje implementaci NOMA. Ve srovnání s klasickým CoMP schématem založeným na ortogonálním víceuzivatelském přístupu, kde více základnových stanic obsluhuje pouze jednoho uživatele CoMP, navržené schéma N-NOMA umožňuje větší konektivitu díky obsluze dodatečných uživatelů NOMA a zlepšuje spektrální efektivitu tím, že uživatel CoMP neobsadí spektrum výhradně. Pro modelování uvažovaného scénáře N-NOMA je použit přístup založený na stochastické geometrii, který modeluje situaci jako Poissonův shlukový proces, na jehož základě jsou odvozeny analytické výrazy uzavřené formy pro pravděpodobnosti výpadků a ergodické rychlosti. Jsou prezentovány numerické výsledky, které ukazují přesnost analytických výsledků a rovněž demonstrují lepší výkon navrženého schématu N-NOMA.
|
cs
|
Given $\beta\in(1,2)$ the fat Sierpinski gasket $\mathcal S_\beta$ is the self-similar set in $\mathbb R^2$ generated by the iterated function system (IFS) \[ f_{\beta,d}(x)=\frac{x+d}{\beta},\quad d\in\mathcal A:=\{(0, 0), (1,0), (0,1)\}. \] Then for each point $P\in\mathcal S_\beta$ there exists a sequence $(d_i)\in\mathcal A^\mathbb N$ such that $P=\sum_{i=1}^\infty d_i/\beta^i$, and the infinite sequence $(d_i)$ is called a \emph{coding} of $P$. In general, a point in $\mathcal S_\beta$ may have multiple codings since the overlap region $\mathcal O_\beta:=\bigcup_{c,d\in\mathcal A, c\ne d}f_{\beta,c}(\Delta_\beta)\cap f_{\beta,d}(\Delta_\beta)$ has non-empty interior, where $\Delta_\beta$ is the convex hull of $\mathcal S_\beta$. In this paper we are interested in the invariant set \[ \widetilde{\mathcal U}_\beta:=\left\{\sum_{i=1}^\infty \frac{d_i}{\beta^i}\in \mathcal S_\beta: \sum_{i=1}^\infty\frac{d_{n+i}}{\beta^i}\notin\mathcal O_\beta~\forall n\ge 0\right\}. \] Then each point in $ \widetilde{\mathcal U}_\beta$ has a unique coding. We show that there is a transcendental number $\beta_c\approx 1.55263$ related to the Thue-Morse sequence, such that $\widetilde{\mathcal U}_\beta$ has positive Hausdorff dimension if and only if $\beta>\beta_{c}$. Furthermore, for $\beta=\beta_c$ the set $\widetilde{\mathcal U}_\beta$ is uncountable but has zero Hausdorff dimension, and for $\beta<\beta_c$ the set $\widetilde{\mathcal U}_\beta$ is at most countable. Consequently, we also answer a conjecture of Sidorov (2007). Our strategy is using combinatorics on words based on the lexicographical characterization of $\widetilde{\mathcal U}_\beta$.
|
Dado $\beta\in(1,2)$, el tapete de Sierpinski gordo $\mathcal S_\beta$ es el conjunto auto-similar en $\mathbb R^2$ generado por el sistema de funciones iteradas (IFS) \[ f_{\beta,d}(x)=\frac{x+d}{\beta},\quad d\in\mathcal A:=\{(0, 0), (1,0), (0,1)\}. \] Entonces, para cada punto $P\in\mathcal S_\beta$ existe una sucesión $(d_i)\in\mathcal A^\mathbb N$ tal que $P=\sum_{i=1}^\infty d_i/\beta^i$, y la sucesión infinita $(d_i)$ se denomina una \emph{codificación} de $P$. En general, un punto en $\mathcal S_\beta$ puede tener múltiples codificaciones, ya que la región de superposición $\mathcal O_\beta:=\bigcup_{c,d\in\mathcal A, c\ne d}f_{\beta,c}(\Delta_\beta)\cap f_{\beta,d}(\Delta_\beta)$ tiene interior no vacío, donde $\Delta_\beta$ es la envolvente convexa de $\mathcal S_\beta$. En este artículo estamos interesados en el conjunto invariante \[ \widetilde{\mathcal U}_\beta:=\left\{\sum_{i=1}^\infty \frac{d_i}{\beta^i}\in \mathcal S_\beta: \sum_{i=1}^\infty\frac{d_{n+i}}{\beta^i}\notin\mathcal O_\beta~\forall n\ge 0\right\}. \] Entonces, cada punto en $ \widetilde{\mathcal U}_\beta$ tiene una única codificación. Mostramos que existe un número trascendente $\beta_c\approx 1.55263$ relacionado con la secuencia de Thue-Morse, tal que $\widetilde{\mathcal U}_\beta$ tiene dimensión de Hausdorff positiva si y solo si $\beta>\beta_{c}$. Además, para $\beta=\beta_c$ el conjunto $\widetilde{\mathcal U}_\beta$ es no numerable pero tiene dimensión de Hausdorff cero, y para $\beta<\beta_c$ el conjunto $\widetilde{\mathcal U}_\beta$ es a lo sumo numerable. En consecuencia, también respondemos una conjetura de Sidorov (2007). Nuestra estrategia consiste en usar combinatoria de palabras basada en la caracterización lexicográfica de $\widetilde{\mathcal U}_\beta$.
|
es
|
Recently, many works have tried to augment the performance of Chinese named entity recognition (NER) using word lexicons. As a representative, Lattice-LSTM (Zhang and Yang, 2018) has achieved new benchmark results on several public Chinese NER datasets. However, Lattice-LSTM has a complex model architecture. This limits its application in many industrial areas where real-time NER responses are needed. In this work, we propose a simple but effective method for incorporating the word lexicon into the character representations. This method avoids designing a complicated sequence modeling architecture, and for any neural NER model, it requires only subtle adjustment of the character representation layer to introduce the lexicon information. Experimental studies on four benchmark Chinese NER datasets show that our method achieves an inference speed up to 6.15 times faster than those of state-ofthe-art methods, along with a better performance. The experimental results also show that the proposed method can be easily incorporated with pre-trained models like BERT.
|
ໃນໄລຍະມໍ້ານີ້, ມີຫຼາຍງານທີ່ພະຍາຍາມປັບປຸງປະສິດທິພາບຂອງການຮູ້ຈັກຊື່ເອກະລັກຈີນ (NER) ໂດຍໃຊ້ພจนານຸກົມຄຳ. ເປັນຕົວແທນ, Lattice-LSTM (Zhang ແລະ Yang, 2018) ໄດ້ບັນລຸຜົນໄດ້ຮັບໃໝ່ໃນຫຼາຍຊຸດຂໍ້ມູນສາທາລະນະດ້ານ NER ຂອງຈີນ. ແຕ່ຢ່າງໃດກໍຕາມ, Lattice-LSTM ມີໂຄງສ້າງຮູບແບບທີ່ສັບສົນ. ສິ່ງນີ້ຈຳກັດການນຳໃຊ້ຂອງມັນໃນເຂດອຸດສາຫະກຳຫຼາຍດ້ານທີ່ຕ້ອງການການຕອບສະໜອງ NER ໃນທັນທີ. ໃນງານນີ້, ພວກເຮົາຂໍສະເໜີວິທີການງ່າຍດາຍແຕ່ມີປະສິດທິພາບສຳລັບການນຳເອົາຂໍ້ມູນພจนານຸກົມຄຳເຂົ້າໃນການແທດແທງຕົວອັກສອນ. ວິທີການນີ້ຫຼີກລ່ຽງການອອກແບບໂຄງສ້າງການຈຳລອງລຳດັບທີ່ສັບສົນ, ແລະ ສຳລັບຮູບແບບ NER ທີ່ເປັນເຄືອຂ່າຍປັນຍາປະດິດໃດໆ, ມັນຕ້ອງການພຽງແຕ່ການປັບປຸງເລັກນ້ອຍໃນຊັ້ນການແທດແທງຕົວອັກສອນເພື່ອນຳເອົາຂໍ້ມູນພจนານຸກົມເຂົ້າມາ. ການສຶກສາທາງການທົດລອງໃນສີ່ຊຸດຂໍ້ມູນສາທາລະນະດ້ານ NER ຂອງຈີນສະແດງໃຫ້ເຫັນວ່າວິທີການຂອງພວກເຮົາບັນລຸຄວາມໄວໃນການຄາດຄະເນໄດ້ໄວຂຶ້ນເຖິງ 6.15 ເທົ່າ ປຽບທຽບກັບວິທີການທີ່ດີທີ່ສຸດໃນປັດຈຸບັນ, ພ້ອມທັງມີປະສິດທິພາບດີຂຶ້ນ. ຜົນການທົດລອງຍັງສະແດງໃຫ້ເຫັນວ່າວິທີການທີ່ສະເໜີນີ້ສາມາດຖືກນຳມານຳໃຊ້ຮ່ວມກັບຮູບແບບທີ່ຖືກຝຶກມາແລ້ວໄດ້ຢ່າງງ່າຍດາຍ, ເຊັ່ນ BERT.
|
lo
|
We prove the computational weakness of a model of tile assembly that has so far resisted many attempts of formal analysis or positive constructions. Specifically, we prove that, in Winfree's abstract Tile Assembly Model, when restricted to use only noncooperative bindings, any long enough path that can grow in all terminal assemblies is pumpable, meaning that this path can be extended into an infinite, ultimately periodic path. This result can be seen as a geometric generalization of the pumping lemma of finite state automata, and closes the question of what can be computed deterministically in this model. Moreover, this question has motivated the development of a new method called visible glues. We believe that this method can also be used to tackle other long-standing problems in computational geometry, in relation for instance with self-avoiding paths. Tile assembly (including non-cooperative tile assembly) was originally introduced by Winfree and Rothemund in STOC 2000 to understand how to program shapes. The non-cooperative variant, also known as temperature 1 tile assembly, is the model where tiles are allowed to bind as soon as they match on one side, whereas in cooperative tile assembly, some tiles need to match on several sides in order to bind. In this work, we prove that only very simple shapes can indeed be programmed, whereas exactly one known result (SODA 2014) showed a restriction on the assemblies general non-cooperative self-assembly could achieve, without any implication on its computational expressiveness. With non-square tiles (like polyominos, SODA 2015), other recent works have shown that the model quickly becomes computationally powerful.
|
نُثبت الضعف الحسابي لنموذج تجميع البلاط الذي ظل حتى الآن يقاوم العديد من محاولات التحليل الرسمي أو البناءات الإيجابية. على وجه التحديد، نُثبت أنه في النموذج المجرد لتجميع البلاط لـ وينفري، عندما يُقتصر على استخدام الارتباطات غير التعاونية فقط، فإن أي مسار طويل بما يكفي يمكن أن ينمو في جميع التجميعات النهائية يكون قابلاً للتضخيم، بمعنى أن هذا المسار يمكن توسيعه ليصبح مسارًا غير منتهٍ ودوريًا في النهاية. يمكن اعتبار هذه النتيجة تعميمًا هندسيًا لبُعد الضخ لآلات الحالة المنتهية، وتسد مسألة ما يمكن حسابه بشكل حتمي في هذا النموذج. علاوة على ذلك، حفّزت هذه المسألة تطوير طريقة جديدة تُعرف باللصقات المرئية. ونعتقد أن هذه الطريقة يمكن أيضًا استخدامها لمعالجة مسائل مستعصية أخرى في الهندسة الحسابية، فيما يتعلق مثلًا بالمسارات ذاتية التجنّب. تم تقديم نموذج تجميع البلاط (بما في ذلك التجميع غير التعاوني للبلاط) في الأصل من قِبل وينفري وروثيموند في مؤتمر STOC 2000 لفهم كيفية برمجة الأشكال. أما الشكل غير التعاوني، والمعروف أيضًا باسم تجميع البلاط عند درجة الحرارة 1، فهو النموذج الذي تُسمح فيه للبلاطات بالارتباط بمجرد تطابقها من جانب واحد، في حين أن التجميع التعاوني للبلاط يتطلب من بعض البلاطات أن تتطابق على عدة جوانب كي ترتبط. في هذا العمل، نُثبت أنه لا يمكن برمجة أشكال بسيطة جدًا فقط فعليًا، في حين أظهرت نتيجة واحدة فقط معروفة (SODA 2014) وجود قيد على التجميعات التي يمكن أن يحققها التجميع الذاتي غير التعاوني بشكل عام، دون أي تأثير على قابليته التعبيرية الحسابية. ومع البلاط غير المربع (مثل البوليومينوات، SODA 2015)، أظهرت أعمال حديثة أخرى أن النموذج يصبح سريعًا قويًا من الناحية الحسابية.
|
ar
|
In order to study the dependence of the coercive force of sintered magnets on temperature, nucleation and domain wall propagation at the grain boundary are studied as rate-determining processes of the magnetization reversal phenomena in magnets consisting of bulk hard magnetic grains contacting via grain boundaries of a soft magnetic material. These systems have been studied analytically for a continuum model at zero temperature (A. Sakuma, et al. J. Mag. Mag. Mat. {\bf 84} 52 (1990)). In the present study, the temperature dependence is studied by making use of the stochastic Landau-Lifshitz-Gilbert equation at finite temperatures. In particular, the threshold fields for nucleation and domain wall propagation are obtained as functions of ratios of magnetic interactions and anisotropies of the soft and hard magnets for various temperatures. It was found that the threshold field for domain wall propagation is robust against thermal fluctuations, while that for nucleation is fragile. The microscopic mechanisms of the observed temperature dependence are discussed.
|
Upang mapag-aralan ang pagkagugol ng coercive force ng mga sintered magnet sa temperatura, ang nucleation at paggalaw ng domain wall sa hangganan ng grano ay pinag-aralan bilang mga prosesong nakapangangasiwa sa bilis ng mga pangyayari ng magnetization reversal sa mga magnet na binubuo ng maselang mga grano ng matitigas na magnet na kumokontak sa pamamagitan ng mga hangganan ng grano ng isang malambot na magnetic na materyal. Ang mga sistemang ito ay napag-aralan nang analitikal para sa isang continuum model sa serong temperatura (A. Sakuma, at iba pa. J. Mag. Mag. Mat. {\bf 84} 52 (1990)). Sa kasalukuyang pag-aaral, ang pag-aaral ng pagkagugol sa temperatura ay isinagawa gamit ang stochastic Landau-Lifshitz-Gilbert equation sa mga may hanggang temperatura. Partikular, ang mga threshold field para sa nucleation at paggalaw ng domain wall ay natukoy bilang mga tungkulin ng mga rasyo ng magnetic interactions at anisotropies ng malambot at matitigas na mga magnet para sa iba't ibang temperatura. Natuklasan na ang threshold field para sa paggalaw ng domain wall ay matibay laban sa thermal fluctuations, samantalang ang naman para sa nucleation ay mahina. Ang mga mikroskopikong mekanismo ng obserbong pagkagugol sa temperatura ay tinalakay.
|
tl
|
A novel concept of orbital launch system in which all stages are reusable is presented. The first two stages called Midpoint Delivery System (MPDS) deliver the next stages to a midpoint. A midpoint is defined by an altitude of 100 $km$ to 120 $km$ and horizontal velocity of 2.8 $km/s$ to 3.2 $km/s$. MPDS stages decelerate in the atmosphere and perform vertical landing on barges. These stages can be reused daily for many years. The payload is delivered from the midpoint to a 400 $km$ Low Earth Orbit by one or two stage rocket called Midpoint to Orbit Delivery System (MPTO). All of MPTO engines are delivered to LEO. These engines do not return to Earth themselves. They are returned to Earth in packs of 50 to 100 by a Reentry Vehicle. Overall, the fully and multiply reusable launch system should deliver payload to LEO for \$300 to \$400 per $kg$
|
পুনর্ব্যবহারযোগ্য সমস্ত পর্যায়ের একটি কক্ষপথীয় উৎক্ষেপণ ব্যবস্থার একটি নতুন ধারণা উপস্থাপন করা হয়েছে। প্রথম দুটি পর্যায়, যাদের মাঝামাঝি বিতরণ ব্যবস্থা (MPDS) বলা হয়, পরবর্তী পর্যায়গুলিকে একটি মাঝামাঝি বিন্দুতে পৌঁছে দেয়। একটি মাঝামাঝি বিন্দু সংজ্ঞায়িত করা হয় 100 $km$ থেকে 120 $km$ উচ্চতা এবং 2.8 $km/s$ থেকে 3.2 $km/s$ আনুভূমিক বেগ দ্বারা। MPDS এর পর্যায়গুলি বায়ুমণ্ডলে মন্থর গতি হ্রাস করে এবং ঘাটের উপর উল্লম্বভাবে অবতরণ করে। এই পর্যায়গুলি বহু বছর ধরে প্রতিদিন পুনর্ব্যবহার করা যেতে পারে। মাঝামাঝি বিন্দু থেকে 400 $km$ কম পৃথিবী কক্ষপথে এক বা দুই পর্যায়ের রকেট দ্বারা পে-লোড পৌঁছে দেওয়া হয়, যাকে মাঝামাঝি থেকে কক্ষপথে বিতরণ ব্যবস্থা (MPTO) বলা হয়। MPTO-এর সমস্ত ইঞ্জিন LEO-এ পৌঁছে যায়। এই ইঞ্জিনগুলি নিজেরাই পৃথিবীতে ফিরে আসে না। এগুলি 50 থেকে 100 পর্যন্ত গুচ্ছে একটি পুনঃপ্রবেশ যান দ্বারা পৃথিবীতে ফিরিয়ে আনা হয়। মোটের উপর, সম্পূর্ণ এবং বহুগুণ পুনর্ব্যবহারযোগ্য উৎক্ষেপণ ব্যবস্থা $kg$ প্রতি \$300 থেকে \$400 খরচে LEO-এ পে-লোড পৌঁছে দেবে।
|
bn
|
Technological advances have made wireless sensors cheap and reliable enough to be brought into industrial use. A major challenge arises from the fact that wireless channels introduce random packet dropouts. Power control and coding are key enabling technologies in wireless communications to ensure efficient communications. In the present work, we examine the role of power control and coding for Kalman filtering over wireless correlated channels. Two estimation architectures are considered: In the first, the sensors send their measurements directly to a single gateway. In the second scheme, wireless relay nodes provide additional links. The gateway decides on the coding scheme and the transmitter power levels of the wireless nodes. The decision process is carried out on-line and adapts to varying channel conditions in order to improve the trade-off between state estimation accuracy and energy expenditure. In combination with predictive power control, we investigate the use of multiple-description coding, zero-error coding and network coding and provide sufficient conditions for the expectation of the estimation error covariance matrix to be bounded. Numerical results suggest that the proposed method may lead to energy savings of around 50 %, when compared to an alternative scheme, wherein transmission power levels and bit-rates are governed by simple logic. In particular, zero-error coding is preferable at time instances with high channel gains, whereas multiple-description coding is superior for time instances with low gains. When channels between the sensors and the gateway are in deep fades, network coding improves estimation accuracy significantly without sacrificing energy efficiency.
|
Kemajuan teknologi telah menjadikan sensor tanpa wayar murah dan cukup boleh dipercayai untuk digunakan dalam industri. Cabaran utama timbul daripada hakikat bahawa saluran tanpa wayar memperkenalkan kehilangan paket secara rawak. Kawalan kuasa dan penyulitan merupakan teknologi utama yang membolehkan komunikasi tanpa wayar berjalan secara efisien. Dalam kajian ini, kami mengkaji peranan kawalan kuasa dan penyulitan untuk penapisan Kalman melalui saluran berkorelasi tanpa wayar. Dua seni bina anggaran dipertimbangkan: Pertama, sensor menghantar ukuran mereka secara langsung ke satu gerbang tunggal. Dalam skema kedua, nod relai tanpa wayar menyediakan pautan tambahan. Gerbang membuat keputusan mengenai skema penyulitan dan tahap kuasa pemancar bagi nod tanpa wayar tersebut. Proses keputusan ini dilaksanakan secara dalam-talian dan menyesuaikan diri dengan keadaan saluran yang berubah bagi memperbaiki imbangan antara ketepatan anggaran keadaan dan perbelanjaan tenaga. Seiring dengan kawalan kuasa berjadual, kami mengkaji penggunaan penyulitan penerangan berganda, penyulitan ralat-sifar dan penyulitan rangkaian, serta memberikan syarat mencukupi bagi jangkaan matriks kovarians ralat anggaran untuk kekal terbatas. Keputusan berangka menunjukkan bahawa kaedah yang dicadangkan boleh membawa kepada penjimatan tenaga sekitar 50%, berbanding dengan skema alternatif di mana tahap kuasa pemancaran dan kadar bit dikawal oleh logik yang mudah. Secara khusus, penyulitan ralat-sifar lebih digemari pada saat-saat dengan gandaan saluran yang tinggi, manakala penyulitan penerangan berganda lebih unggul pada saat-saat dengan gandaan rendah. Apabila saluran antara sensor dan gerbang mengalami kerosotan teruk, penyulitan rangkaian meningkatkan ketepatan anggaran secara ketara tanpa mengorbankan kecekapan tenaga.
|
ms
|
We have detected, for the first time, Cepheid variables in the Sculptor Group spiral galaxy NGC 7793. From wide-field images obtained in the optical V and I bands on 56 nights in 2003-2005, we have discovered 17 long-period (24-62 days) Cepheids whose periods and mean magnitudes define tight period-luminosity relations. We use the (V-I) Wesenheit index to determine a reddening-free true distance modulus to NGC 7793 of 27.68 +- 0.05 mag (internal error) +- 0.08 mag (systematic error). The comparison of the reddened distance moduli in V and I with the one derived from the Wesenheit magnitude indicates that the Cepheids in NGC 7793 are affected by an average total reddening of E(B-V)=0.08 mag, 0.06 of which is produced inside the host galaxy. As in the earlier Cepheid studies of the Araucaria Project, the reported distance is tied to an assumed LMC distance modulus of 18.50. The quoted systematic uncertainty takes into account effects like blending and possible inhomogeneous filling of the Cepheid instability strip on the derived distance. The reported distance value does not depend on the (unknown) metallicity of the Cepheids according to recent theoretical and empirical results. Our Cepheid distance is shorter, but within the errors consistent with the distance to NGC 7793 determined earlier with the TRGB and Tully-Fisher methods. The NGC 7793 distance of 3.4 Mpc is almost identical to the one our project had found from Cepheid variables for NGC 247, another spiral member of the Sculptor Group located close to NGC 7793 on the sky. Two other conspicuous spiral galaxies in the Sculptor Group, NGC 55 and NGC 300, are much nearer (1.9 Mpc), confirming the picture of a very elongated structure of the Sculptor Group in the line of sight put forward by Jerjen et al. and others.
|
Hemos detectado, por primera vez, variables Cefeidas en la galaxia espiral NGC 7793 del Grupo Escultor. A partir de imágenes de campo amplio obtenidas en los bandas ópticos V e I en 56 noches entre 2003 y 2005, hemos descubierto 17 Cefeidas de largo período (24-62 días) cuyos períodos y magnitudes medias definen relaciones período-luminosidad muy ajustadas. Utilizamos el índice Wesenheit (V-I) para determinar un módulo de distancia verdadero libre de enrojecimiento hacia NGC 7793 de 27,68 ± 0,05 mag (error interno) ± 0,08 mag (error sistemático). La comparación de los módulos de distancia enrojecidos en V e I con el derivado a partir de la magnitud Wesenheit indica que las Cefeidas en NGC 7793 están afectadas por un enrojecimiento total promedio de E(B-V) = 0,08 mag, de los cuales 0,06 se producen dentro de la galaxia huésped. Como en estudios anteriores de Cefeidas del Proyecto Araucaria, la distancia informada está referida a un módulo de distancia asumido de 18,50 para la Gran Nube de Magallanes. La incertidumbre sistemática citada tiene en cuenta efectos como la mezcla de luz (blending) y el posible relleno no homogéneo de la banda de inestabilidad de las Cefeidas sobre la distancia derivada. El valor de distancia informado no depende de la metalicidad (desconocida) de las Cefeidas, de acuerdo con resultados teóricos y empíricos recientes. Nuestra distancia basada en Cefeidas es más corta, pero compatible dentro de los errores con la distancia a NGC 7793 determinada previamente mediante los métodos TRGB y de Tully-Fisher. La distancia a NGC 7793 de 3,4 Mpc es casi idéntica a la que nuestro proyecto había encontrado a partir de variables Cefeidas para NGC 247, otra galaxia espiral miembro del Grupo Escultor ubicada cerca de NGC 7793 en el cielo. Otras dos galaxias espirales prominentes del Grupo Escultor, NGC 55 y NGC 300, están mucho más cerca (1,9 Mpc), confirmando el panorama de una estructura muy alargada del Grupo Escultor a lo largo de la línea de visión, propuesto por Jerjen et al. y otros.
|
es
|
We prove the computational weakness of a model of tile assembly that has so far resisted many attempts of formal analysis or positive constructions. Specifically, we prove that, in Winfree's abstract Tile Assembly Model, when restricted to use only noncooperative bindings, any long enough path that can grow in all terminal assemblies is pumpable, meaning that this path can be extended into an infinite, ultimately periodic path. This result can be seen as a geometric generalization of the pumping lemma of finite state automata, and closes the question of what can be computed deterministically in this model. Moreover, this question has motivated the development of a new method called visible glues. We believe that this method can also be used to tackle other long-standing problems in computational geometry, in relation for instance with self-avoiding paths. Tile assembly (including non-cooperative tile assembly) was originally introduced by Winfree and Rothemund in STOC 2000 to understand how to program shapes. The non-cooperative variant, also known as temperature 1 tile assembly, is the model where tiles are allowed to bind as soon as they match on one side, whereas in cooperative tile assembly, some tiles need to match on several sides in order to bind. In this work, we prove that only very simple shapes can indeed be programmed, whereas exactly one known result (SODA 2014) showed a restriction on the assemblies general non-cooperative self-assembly could achieve, without any implication on its computational expressiveness. With non-square tiles (like polyominos, SODA 2015), other recent works have shown that the model quickly becomes computationally powerful.
|
Мы доказываем вычислительную слабость модели сборки плиток, которая до сих пор противостояла многим попыткам формального анализа или построения положительных результатов. А именно, мы доказываем, что в абстрактной модели сборки плиток Винфри, когда она ограничена использованием только некооперативных связываний, любой достаточно длинный путь, который может расти во всех терминальных сборках, является прокачиваемым, то есть этот путь можно расширить до бесконечного, в конечном счёте периодического пути. Этот результат можно рассматривать как геометрическое обобщение леммы о прокачке для конечных автоматов и он закрывает вопрос о том, что может быть вычислено детерминированно в данной модели. Более того, этот вопрос стимулировал разработку нового метода, названного видимыми клеевыми соединениями. Мы полагаем, что данный метод может быть также использован для решения других давних проблем в вычислительной геометрии, в частности, связанных с самонепересекающимися путями. Модель сборки плиток (включая некооперативную сборку плиток) была впервые предложена Винфри и Ротемундом на конференции STOC 2000 для понимания того, как программировать формы. Некооперативный вариант, также известный как сборка плиток при температуре 1, — это модель, в которой плитки могут присоединяться, как только совпадают по одной стороне, тогда как в кооперативной сборке некоторые плитки должны совпадать по нескольким сторонам, чтобы присоединиться. В данной работе мы доказываем, что на самом деле можно запрограммировать лишь очень простые формы, тогда как ранее был известен лишь один результат (SODA 2014), показывающий ограничение на структуры, достижимые общей некооперативной самосборкой, не имеющий следствий относительно её вычислительной выразительности. При использовании непрямоугольных плиток (например, полимино, SODA 2015) другие недавние работы показали, что модель быстро становится вычислительно мощной.
|
ru
|
This paper investigates the problem of finding a fixed point for a global nonexpansive operator under time-varying communication graphs in real Hilbert spaces, where the global operator is separable and composed of an aggregate sum of local nonexpansive operators. Each local operator is only privately accessible to each agent, and all agents constitute a network. To seek a fixed point of the global operator, it is indispensable for agents to exchange local information and update their solution cooperatively. To solve the problem, two algorithms are developed, called distributed Krasnosel'ski\u{\i}-Mann (D-KM) and distributed block-coordinate Krasnosel'ski\u{\i}-Mann (D-BKM) iterations, for which the D-BKM iteration is a block-coordinate version of the D-KM iteration in the sense of randomly choosing and computing only one block-coordinate of local operators at each time for each agent. It is shown that the proposed two algorithms can both converge weakly to a fixed point of the global operator. Meanwhile, the designed algorithms are applied to recover the classical distributed gradient descent (DGD) algorithm, devise a new block-coordinate DGD algorithm, handle a distributed shortest distance problem in the Hilbert space for the first time, and solve linear algebraic equations in a novel distributed approach. Finally, the theoretical results are corroborated by a few numerical examples.
|
Bài báo này nghiên cứu bài toán tìm điểm bất động của một toán tử không dãn toàn cục trong không gian Hilbert thực dưới đồ thị truyền thông biến đổi theo thời gian, trong đó toán tử toàn cục là phân tách được và bao gồm tổng tổng hợp các toán tử không dãn cục bộ. Mỗi toán tử cục bộ chỉ được truy cập riêng bởi từng tác nhân, và tất cả các tác nhân tạo thành một mạng lưới. Để tìm điểm bất động của toán tử toàn cục, việc trao đổi thông tin cục bộ và cập nhật nghiệm một cách hợp tác giữa các tác nhân là điều thiết yếu. Để giải quyết bài toán này, hai thuật toán được đề xuất, gọi là phép lặp Krasnosel'ski\u{\i}-Mann phân tán (D-KM) và phép lặp Krasnosel'ski\u{\i}-Mann phân tán theo tọa độ khối (D-BKM), trong đó phép lặp D-BKM là phiên bản theo tọa độ khối của D-KM, theo nghĩa tại mỗi thời điểm, mỗi tác nhân chỉ chọn ngẫu nhiên và tính toán một tọa độ khối duy nhất của toán tử cục bộ. Kết quả cho thấy cả hai thuật toán đề xuất đều hội tụ yếu đến một điểm bất động của toán tử toàn cục. Đồng thời, các thuật toán được thiết kế áp dụng để khôi phục lại thuật toán giảm gradient phân tán cổ điển (DGD), đề xuất một thuật toán DGD theo tọa độ khối mới, giải quyết lần đầu tiên bài toán khoảng cách ngắn nhất phân tán trong không gian Hilbert, và giải hệ phương trình đại số tuyến tính theo một cách tiếp cận phân tán mới mẻ. Cuối cùng, các kết quả lý thuyết được kiểm chứng bằng một số ví dụ số.
|
vi
|
This paper considers a downlink ultra-dense heterogeneous cloud radio access network (H-CRAN) which guarantees seamless coverage and can provide high date rates. In order to reduce channel state information (CSI) feedback overhead, incomplete inter-cluster CSI is considered, i.e., each remote radio head (RRH) or macro base station (MBS) only measures the CSI from user equipments (UEs) in its serving cluster. To reduce pilot consumption, pilot reuse among UEs is assumed, resulting in imperfect intra-cluster CSI. A two-stage optimization problem is then formulated. In the first stage, a pilot scheduling algorithm is proposed to minimize the sum mean square error (MSE) of all channel estimates. Specifically, the minimum number of required pilots along with a feasible pilot allocation solution are first determined by applying the Dsatur algorithm, and adjustments based on the defined level of pilot contamination are then carried out for further improvement. Based on the pilot allocation result obtained in the first stage, the second stage aims to maximize the sum spectral efficiency (SE) of the network by optimizing the beam-vectors. Due to incomplete inter-cluster CSI and imperfect intra-cluster CSI, an explicit expression of each UE's achievable rate is unavailable. Hence, a lower bound on the achievable rate is derived based on Jensen's inequality, and an alternative robust transmission design (RTD) algorithm along with its distributed realization are then proposed to maximize the derived tight lower bound. Simulation results show that compared with existing algorithms, the system performance can be greatly improved by the proposed algorithms in terms of both sum MSE and sum SE.
|
本文研究了一种下行链路超密集异构云无线接入网络（H-CRAN），该网络可保证无缝覆盖并提供高速数据速率。为了降低信道状态信息（CSI）反馈开销，本文考虑不完整的跨簇CSI，即每个远程射频头端（RRH）或宏基站（MBS）仅测量其服务簇内用户设备（UE）的CSI。为减少导频开销，假设UE之间进行导频复用，从而导致簇内CSI不准确。随后，本文建立了一个两阶段优化问题。在第一阶段，提出一种导频调度算法，以最小化所有信道估计的均方误差（MSE）之和。具体而言，首先利用Dsatur算法确定所需的最少导频数量及一种可行的导频分配方案，然后根据定义的导频污染程度进行调整，以进一步优化性能。在第二阶段，基于第一阶段获得的导频分配结果，通过优化波束向量以最大化网络的总频谱效率（SE）。由于存在跨簇CSI不完整和簇内CSI不准确的情况，无法获得每个UE可达速率的显式表达式。因此，基于Jensen不等式推导出可达速率的一个下界，并提出一种替代性的鲁棒传输设计（RTD）算法及其分布式实现方法，以最大化该紧致的下界。仿真结果表明，与现有算法相比，所提出的算法在总MSE和总SE方面均能显著提升系统性能。
|
zh
|
Successful applications of reinforcement learning in real-world problems often require dealing with partially observable states. It is in general very challenging to construct and infer hidden states as they often depend on the agent's entire interaction history and may require substantial domain knowledge. In this work, we investigate a deep-learning approach to learning the representation of states in partially observable tasks, with minimal prior knowledge of the domain. In particular, we propose a new family of hybrid models that combines the strength of both supervised learning (SL) and reinforcement learning (RL), trained in a joint fashion: The SL component can be a recurrent neural networks (RNN) or its long short-term memory (LSTM) version, which is equipped with the desired property of being able to capture long-term dependency on history, thus providing an effective way of learning the representation of hidden states. The RL component is a deep Q-network (DQN) that learns to optimize the control for maximizing long-term rewards. Extensive experiments in a direct mailing campaign problem demonstrate the effectiveness and advantages of the proposed approach, which performs the best among a set of previous state-of-the-art methods.
|
Le applicazioni di successo dell'apprendimento per rinforzo in problemi del mondo reale richiedono spesso di affrontare stati parzialmente osservabili. In generale, è molto difficile costruire e inferire stati nascosti, poiché spesso dipendono dall'intera storia di interazione dell'agente e possono richiedere una conoscenza approfondita del dominio. In questo lavoro, esploriamo un approccio basato sul deep learning per apprendere la rappresentazione degli stati in compiti parzialmente osservabili, con un minimo di conoscenza a priori del dominio. In particolare, proponiamo una nuova famiglia di modelli ibridi che combina i punti di forza sia dell'apprendimento supervisionato (SL) che dell'apprendimento per rinforzo (RL), addestrati congiuntamente: il componente SL può essere una rete neurale ricorrente (RNN) o la sua versione long short-term memory (LSTM), dotata della proprietà desiderata di essere in grado di catturare dipendenze a lungo termine dalla storia, fornendo così un modo efficace per apprendere la rappresentazione degli stati nascosti. Il componente RL è una deep Q-network (DQN) che apprende a ottimizzare il controllo per massimizzare le ricompense a lungo termine. Esperimenti approfonditi in un problema di campagna di posta diretta dimostrano l'efficacia e i vantaggi dell'approccio proposto, che ottiene prestazioni migliori rispetto a un insieme di metodi all'avanguardia precedentemente esistenti.
|
it
|
A new network with super approximation power is introduced. This network is built with Floor ($\lfloor x\rfloor$) or ReLU ($\max\{0,x\}$) activation function in each neuron and hence we call such networks Floor-ReLU networks. For any hyper-parameters $N\in\mathbb{N}^+$ and $L\in\mathbb{N}^+$, it is shown that Floor-ReLU networks with width $\max\{d,\, 5N+13\}$ and depth $64dL+3$ can uniformly approximate a H\"older function $f$ on $[0,1]^d$ with an approximation error $3\lambda d^{\alpha/2}N^{-\alpha\sqrt{L}}$, where $\alpha \in(0,1]$ and $\lambda$ are the H\"older order and constant, respectively. More generally for an arbitrary continuous function $f$ on $[0,1]^d$ with a modulus of continuity $\omega_f(\cdot)$, the constructive approximation rate is $\omega_f(\sqrt{d}\,N^{-\sqrt{L}})+2\omega_f(\sqrt{d}){N^{-\sqrt{L}}}$. As a consequence, this new class of networks overcomes the curse of dimensionality in approximation power when the variation of $\omega_f(r)$ as $r\to 0$ is moderate (e.g., $\omega_f(r) \lesssim r^\alpha$ for H\"older continuous functions), since the major term to be considered in our approximation rate is essentially $\sqrt{d}$ times a function of $N$ and $L$ independent of $d$ within the modulus of continuity.
|
Diperkenalkan sebuah jaringan baru dengan kemampuan aproksimasi yang sangat tinggi. Jaringan ini dibangun menggunakan fungsi aktivasi Floor ($\lfloor x\rfloor$) atau ReLU ($\max\{0,x\}$) pada setiap neuron, sehingga kita menyebut jaringan semacam ini sebagai jaringan Floor-ReLU. Untuk sebarang hiperparameter $N\in\mathbb{N}^+$ dan $L\in\mathbb{N}^+$, ditunjukkan bahwa jaringan Floor-ReLU dengan lebar $\max\{d,\, 5N+13\}$ dan kedalaman $64dL+3$ dapat mengaproksimasi secara seragam suatu fungsi H\"older $f$ pada $[0,1]^d$ dengan galat aproksimasi sebesar $3\lambda d^{\alpha/2}N^{-\alpha\sqrt{L}}$, di mana $\alpha \in(0,1]$ dan $\lambda$ masing-masing merupakan orde dan konstanta H\"older. Secara lebih umum, untuk sebarang fungsi kontinu $f$ pada $[0,1]^d$ dengan modulus kontinuitas $\omega_f(\cdot)$, laju aproksimasi konstruktifnya adalah $\omega_f(\sqrt{d}\,N^{-\sqrt{L}})+2\omega_f(\sqrt{d}){N^{-\sqrt{L}}}$. Akibatnya, kelas jaringan baru ini mengatasi kurva dimensi dalam kemampuan aproksimasi ketika variasi $\omega_f(r)$ saat $r\to 0$ moderat (misalnya, $\omega_f(r) \lesssim r^\alpha$ untuk fungsi yang kontinu H\"older), karena suku utama yang perlu dipertimbangkan dalam laju aproksimasi kita pada dasarnya adalah $\sqrt{d}$ dikalikan suatu fungsi dari $N$ dan $L$ yang bebas dari $d$ di dalam modulus kontinuitas.
|
id
|
We present a multiwavelength study of the infrared dark cloud MSXDC G034.43+00.24. Dust emission, traced by millimeter/submillimeter images obtained with the IRAM, JCMT, and CSO telescopes, reveals three compact cores within this infrared dark cloud with masses of 170--800 Msun and sizes < 0.5 pc. Spitzer 3.6-8.0 um images show slightly extended emission toward these cores, with a spectral enhancement at 4.5 um that probably arises from shocked H2. In addition, the broad line widths (Delta V ~ 10 km/s) of HCN (4-3), and CS (3-2), and the detection of SiO (2-1), observed with the JCMT and IRAM telescopes, also indicate active star formation. Spitzer 24 um images reveal that each of these cores contains a bright, unresolved continuum source; these sources are most likely embedded protostars. Their millimeter to mid-IR continuum spectral energy distributions reveal very high luminosities, 9000-32,000 Lsun. Because such large luminosities cannot arise from low-mass protostars, MSXDC G034.43+00.24 is actively forming massive (~ 10 Msun) stars.
|
Prezentujeme vícevlnovou studii infračerveně tmavé mračna MSXDC G034.43+00.24. Emise prachu, sledovaná pomocí milimetrových a submilimetrových snímků získaných dalekohledy IRAM, JCMT a CSO, odhaluje tři kompaktní jádra uvnitř tohoto infračerveně tmavého mračna s hmotnostmi 170–800 hmotností Slunce a rozměry menšími než 0,5 pc. Snímky ze Spitzera v rozsahu 3,6–8,0 μm ukazují mírně rozšířenou emisi směrem k těmto jádrům, s pozorovaným zvýšením spektrální intenzity na 4,5 μm, které pravděpodobně pochází z rázově excitovaného H2. Dále široké šířky spektrálních čar (ΔV ~ 10 km/s) HCN (4-3) a CS (3-2) a detekce SiO (2-1), pozorované pomocí dalekohledů JCMT a IRAM, také naznačují probíhající hvězdotvorbu. Snímky ze Spitzera na vlnové délce 24 μm odhalují, že každé z těchto jader obsahuje jasný, nerozlišený kontinuální zdroj; tyto zdroje jsou s největší pravděpodobností vnořené protostary. Jejich spektrální rozdělení energie kontinua v milimetrovém až středním infračerveném oboru ukazuje velmi vysoké zářivé výkony, 9000–32 000 hmotností Slunce. Protože takto velké zářivé výkony nemohou pocházet od nízkomasivních protostar, mračno MSXDC G034.43+00.24 právě vytváří hmotné hvězdy (~ 10 hmotností Slunce).
|
cs
|
Internet-scale distributed systems such as content delivery networks (CDNs) operate hundreds of thousands of servers deployed in thousands of data center locations around the globe. Since the energy costs of operating such a large IT infrastructure are a significant fraction of the total operating costs, we argue for redesigning CDNs to incorporate energy optimizations as a first-order principle. We propose techniques to turn off CDN servers during periods of low load while seeking to balance three key design goals: maximize energy reduction, minimize the impact on client-perceived service availability (SLAs), and limit the frequency of on-off server transitions to reduce wear-and-tear and its impact on hardware reliability. We propose an optimal offline algorithm and an online algorithm to extract energy savings both at the level of local load balancing within a data center and global load balancing across data centers. We evaluate our algorithms using real production workload traces from a large commercial CDN. Our results show that it is possible to reduce the energy consumption of a CDN by more than 55% while ensuring a high level of availability that meets customer SLA requirements and incurring an average of one on-off transition per server per day. Further, we show that keeping even 10% of the servers as hot spares helps absorb load spikes due to global flash crowds with little impact on availability SLAs. Finally, we show that redistributing load across proximal data centers can enhance service availability significantly, but has only a modest impact on energy savings.
|
အင်တာနက်အရွယ်အစားရှိ ဖြန့်ကျက်ထားသောစနစ်များဖြစ်သည့် အကြောင်းအရာ ဖြန့်ဖြူးပေးသည့် ကွန်ရက်များ (CDN) တို့သည် ကမ္ဘာတစ်ဝှမ်းရှိ ဒေတာစင်တာ တည်နေရာ ထောင်ပေါင်းများစွာတွင် စက်ရုံထောင်နှင့်ချီ ထားရှိထားကြသည်။ ထိုကဲ့သို့သော IT အခြေခံအဆောက်အအုံကြီးမားမှု၏ စွမ်းအင်ကုန်ကျစရိတ်များသည် စုစုပေါင်းလည်ပတ်မှုကုန်ကျစရိတ်၏ သိသိသာသာ အပိုင်းအစဖြစ်နေသောကြောင့် စွမ်းအင် အကျိုးရှိမှုများကို ဦးစားပေးသော မူလအချက်တစ်ခုအဖြစ် ထည့်သွင်း၍ CDN များကို ပြန်လည်ဒီဇိုင်းဆွဲရန် ကျွန်ုပ်တို့ အကြံပြုပါသည်။ မျှော်မှန်းထားသော ဝန်ပမာဏ နည်းပါးသည့် အချိန်များအတွင်း စက်ရုံများကို ပိတ်ထားရန် နည်းလမ်းများကို ကျွန်ုပ်တို့ အကြံပြုပါသည်။ ထိုသို့ပြုလုပ်ရာတွင် စွမ်းအင်ကို အများဆုံးလျှော့ချရန်၊ ဝန်ဆောင်မှု ရရှိနိုင်မှု (SLA) ကို ဖောက်သည်များက မည်သို့မြင်မြင် ထိခိုက်မှုကို အနည်းဆုံးဖြစ်စေရန်နှင့် စက်ရုံများ၏ ဖွင့်ပိတ် လုပ်ဆောင်မှု အကြိမ်ရေကို ကန့်သတ်၍ စက်ပစ္စည်း၏ ယုံကြည်စိတ်ချရမှုအပေါ် သက်ရောက်မှုကို လျော့နည်းစေရန် အတွက် အဓိက ဒီဇိုင်းရည်မှန်းချက် သုံးခုကို မျှတစွာ ထားရှိရန် ရည်ရွယ်ပါသည်။ ဒေတာစင်တာအတွင်း ဒေသတွင်း ဝန်ချိန်ညှိမှုအဆင့်နှင့် ဒေတာစင်တာများကို ကျော်လွန်သော ကမ္ဘာ့ဝန်ချိန်ညှိမှုအဆင့်တို့တွင် စွမ်းအင်ချွေတာမှုများကို ရယူရန် အကောင်းဆုံး offline အယ်လဂိုရီသမ်နှင့် online အယ်လဂိုရီသမ်ကို ကျွန်ုပ်တို့ အကြံပြုပါသည်။ ကုန်ပစ္စည်းထုတ်လုပ်မှု၏ အမှန်တကယ် လုပ်ငန်းဝန်ပမာဏ ခြေရာများကို ကုန်သည်ကြီးတစ်ခု၏ CDN မှ အသုံးပြု၍ ကျွန်ုပ်တို့၏ အယ်လဂိုရီသမ်များကို ကျွန်ုပ်တို့ အကဲဖြတ်ပါသည်။ ကျွန်ုပ်တို့၏ ရလဒ်များအရ စက်ရုံတစ်ခုလျှင် တစ်နေ့လျှင် ပျမ်းမျှ ဖွင့်ပိတ်ပြောင်းလဲမှု တစ်ကြိမ်သာ ကျော်လွန်ပြီး ဖောက်သည်များ၏ SLA လိုအပ်ချက်များကို ပြည့်မီစေမည့် အဆင့်မြင့် ရရှိနိုင်မှုကို သေချာစေရန်ဖြင့် CDN ၏ စွမ်းအင်စားသုံးမှုကို ၅၅% ထက်ပို၍ လျှော့ချနိုင်ကြောင်း ပြသပါသည်။ ထို့အပြင် စက်ရုံများ၏ ၁၀% ကို ပူးတွဲအစားထိုးစက်ရုံများအဖြစ် ထားရှိခြင်းဖြင့် ကမ္ဘာ့အဆင့် လူအုပ်ကြီးများ ရုတ်တရက် ဝင်ရောက်လာခြင်းကြောင့် ဖြစ်ပေါ်လာသော ဝန်ပမာဏ တိုးမြင့်မှုကို ရရှိနိုင်မှု SLA များအပေါ် သက်ရောက်မှု နည်းပါးစွာဖြင့် စုပ်ယူနိုင်ကြောင်း ပြသပါသည်။ နောက်ဆုံးတွင် နီးစပ်သော ဒေတာစင်တာများအကြား ဝန်ပမာဏကို ပြန်လည်ဖြန့်ဖြူးခြင်းဖြင့် ဝန်ဆောင်မှုရရှိနိုင်မှုကို သိသိသာသာ မြှင့်တင်နိုင်သော်လည်း စွမ်းအင်ချွေတာမှုအပေါ် သက်ရောက်မှုမှာ သိပ်မများကြောင်း ပြသပါသည်။
|
my
|
Self-assembly of colloidal particles due to elastic interactions in nematic liquid crystals promises tunable composite materials and can be guided by exploiting surface functionalization, geometric shape and topology, though these means of controlling self-assembly remain limited. Here, we realize low-symmetry achiral and chiral elastic colloids in the nematic liquid crystals using colloidal polygonal concave and convex prisms. We show that the controlled pinning of disclinations at the prisms edges alters the symmetry of director distortions around the prisms and their orientation with respect to the far-field director. The controlled localization of the disclinations at the prism's edges significantly influences anisotropy of the diffusion properties of prisms dispersed in liquid crystals and allows one to modify their self-assembly. We show that elastic interactions between polygonal prisms can be switched between repulsive and attractive just by controlled re-pinning the disclinations at different edges using laser tweezers. Our findings demonstrate that elastic interactions between colloidal particles dispersed in nematic liquid crystals are sensitive to the topologically equivalent but geometrically rich controlled configurations of the particle-induced defects.
|
تعد عملية التجميع الذاتي للجسيمات الغروية الناتجة عن التفاعلات المرنة في البلورات السائلة النيماتية واعدةً في إنتاج مواد مركبة قابلة للضبط، ويمكن توجيه هذه العملية باستغلال وظائف السطح والشكل الهندسي والطوبولوجيا، على الرغم من أن هذه الوسائل للتحكم في التجميع الذاتي تظل محدودة. في هذا السياق، نحقق جسيمات غروية مرنة غير متماثلة وذات يدائية في البلورات السائلة النيماتية باستخدام مخاريط غروية مضلعة مقعرة ومحدبة. نُظهر أن تثبيت العيوب الخطية بشكل مُتحكم به عند حواف المخاريط يُغير من تناظر تشوهات الموجه المحيط بالمخاريط، وكذلك من اتجاهها بالنسبة للموجه البعيد. إن التموضع المُتحكم به للعيوب الخطية عند حواف المخروط يؤثر بشكل كبير على تباين خواص الانتشار للمخاريط المعلقة في البلورات السائلة، ويسمح بتعديل تجميعها الذاتي. نُظهر أن التفاعلات المرنة بين المخاريط المضلعة يمكن تحويلها من تنافرية إلى جاذبية، فقط من خلال إعادة تثبيت مُتحكم بها للعيوب الخطية عند حواف مختلفة باستخدام ملاقط الليزر. تُظهر نتائجنا أن التفاعلات المرنة بين الجسيمات الغروية المعلقة في البلورات السائلة النيماتية تكون حساسة للتكوينات الطوبولوجية المكافئة ولكن الغنية من الناحية الهندسية للعيوب الناتجة عن الجسيمات.
|
ar
|
Given $\beta\in(1,2)$ the fat Sierpinski gasket $\mathcal S_\beta$ is the self-similar set in $\mathbb R^2$ generated by the iterated function system (IFS) \[ f_{\beta,d}(x)=\frac{x+d}{\beta},\quad d\in\mathcal A:=\{(0, 0), (1,0), (0,1)\}. \] Then for each point $P\in\mathcal S_\beta$ there exists a sequence $(d_i)\in\mathcal A^\mathbb N$ such that $P=\sum_{i=1}^\infty d_i/\beta^i$, and the infinite sequence $(d_i)$ is called a \emph{coding} of $P$. In general, a point in $\mathcal S_\beta$ may have multiple codings since the overlap region $\mathcal O_\beta:=\bigcup_{c,d\in\mathcal A, c\ne d}f_{\beta,c}(\Delta_\beta)\cap f_{\beta,d}(\Delta_\beta)$ has non-empty interior, where $\Delta_\beta$ is the convex hull of $\mathcal S_\beta$. In this paper we are interested in the invariant set \[ \widetilde{\mathcal U}_\beta:=\left\{\sum_{i=1}^\infty \frac{d_i}{\beta^i}\in \mathcal S_\beta: \sum_{i=1}^\infty\frac{d_{n+i}}{\beta^i}\notin\mathcal O_\beta~\forall n\ge 0\right\}. \] Then each point in $ \widetilde{\mathcal U}_\beta$ has a unique coding. We show that there is a transcendental number $\beta_c\approx 1.55263$ related to the Thue-Morse sequence, such that $\widetilde{\mathcal U}_\beta$ has positive Hausdorff dimension if and only if $\beta>\beta_{c}$. Furthermore, for $\beta=\beta_c$ the set $\widetilde{\mathcal U}_\beta$ is uncountable but has zero Hausdorff dimension, and for $\beta<\beta_c$ the set $\widetilde{\mathcal U}_\beta$ is at most countable. Consequently, we also answer a conjecture of Sidorov (2007). Our strategy is using combinatorics on words based on the lexicographical characterization of $\widetilde{\mathcal U}_\beta$.
|
주어진 $\beta\in(1,2)$에 대하여 두꺼운 시어핀스키 개스크 $\mathcal S_\beta$란 반복함수계(IFS)
\[
f_{\beta,d}(x)=\frac{x+d}{\beta},\quad d\in\mathcal A:=\{(0, 0), (1,0), (0,1)\}
\]
에 의해 생성되는 $\mathbb R^2$ 안의 자기유사 집합이다. 그러면 각 점 $P\in\mathcal S_\beta$에 대하여 $P=\sum_{i=1}^\infty d_i/\beta^i$를 만족하는 수열 $(d_i)\in\mathcal A^\mathbb N$이 존재하며, 이 무한수열 $(d_i)$를 $P$의 \emph{부호화(coding)}라 한다. 일반적으로, $\mathcal S_\beta$의 점은 여러 부호화를 가질 수 있는데, 이는 중첩 영역 $\mathcal O_\beta:=\bigcup_{c,d\in\mathcal A, c\ne d}f_{\beta,c}(\Delta_\beta)\cap f_{\beta,d}(\Delta_\beta)$의 내부가 공집합이 아니기 때문이다. 여기서 $\Delta_\beta$는 $\mathcal S_\beta$의 볼록 껍질(convex hull)이다. 본 논문에서는 다음의 불변 집합에 관심을 둔다:
\[
\widetilde{\mathcal U}_\beta:=\left\{\sum_{i=1}^\infty \frac{d_i}{\beta^i}\in \mathcal S_\beta: \sum_{i=1}^\infty\frac{d_{n+i}}{\beta^i}\notin\mathcal O_\beta~\forall n\ge 0\right\}.
\]
그러면 $\widetilde{\mathcal U}_\beta$의 각 점은 유일한 부호화를 가진다. 우리는 트루-모스 수열과 관련된 초월수 $\beta_c\approx 1.55263$가 존재하여, $\widetilde{\mathcal U}_\beta$의 하우스도르프 차원이 양수가 될 필요충분조건이 $\beta>\beta_{c}$임을 보인다. 더 나아가, $\beta=\beta_c$일 때 집합 $\widetilde{\mathcal U}_\beta$는 비가산적이지만 하우스도르프 차원은 0이며, $\beta<\beta_c$일 때 집합 $\widetilde{\mathcal U}_\beta$는 최대 가산적이다. 따라서 우리는 또한 시도로프(Sidorov, 2007)의 추측에 대한 답을 제시한다. 우리의 전략은 $\widetilde{\mathcal U}_\beta$의 사전식 특성화(lexicographical characterization)에 기반한 단어 조합론을 사용하는 것이다.
|
ko
|
Using synthetic lattices of laser-coupled atomic momentum modes, we experimentally realize a recently proposed family of nearest-neighbor tight-binding models having quasiperiodic site energy modulation that host an exact mobility edge protected by a duality symmetry. These one-dimensional tight-binding models can be viewed as a generalization of the well-known Aubry-Andr\'{e} (AA) model, with an energy-dependent self duality condition that constitutes an analytical mobility edge relation. By adiabatically preparing the lowest and highest energy eigenstates of this model system and performing microscopic measurements of their participation ratio, we track the evolution of the mobility edge as the energy-dependent density of states is modified by the model's tuning parameter. Our results show strong deviations from single-particle predictions, consistent with attractive interactions causing both enhanced localization of the lowest energy state due to self-trapping and inhibited localization of the highest energy state due to screening. This study paves the way for quantitative studies of interaction effects on self duality induced mobility edges.
|
لیزر سے جڑے ایٹمی مومینٹم ماڈلز کے مصنوعی لیٹس کا استعمال کرتے ہوئے، ہم تجرباتی طور پر ایک حال ہی میں تجویز کردہ قریب ترین پڑوسی ٹائٹ بائنڈنگ ماڈلز کے خاندان کو عملی شکل دیتے ہیں جن میں عینیت کی حفاظت یافتہ حرکت پذیری کی حد کو برقرار رکھنے والی کوازی-پیریئوڈک سائٹ توانائی کی ماڈولیشن ہوتی ہے۔ یہ ایک بعدی ٹائٹ بائنڈنگ ماڈلز اچھی طرح جانے جانے والے آبری اینڈرے ماڈل کی توسیع کے طور پر دیکھے جا سکتے ہیں، جن میں توانائی پر منحصر خود-دوقطبیت کی شرط تجزیاتی حرکت پذیری کی حد کے تعلق کی تشکیل کرتی ہے۔ اس ماڈل نظام کی توانائی کی سب سے کم اور سب سے زیادہ ویژہ حالتیں مٰہدود وقت میں تیار کر کے اور ان کے شمولیت کے تناسب کے مائیکرو اقدار کو ناپ کر، ہم حرکت پذیری کی حد کی تبدیلی کا جائزہ لیتے ہیں جب توانائی پر منحصر حالت کی کثافت کو ماڈل کی ٹیوننگ پیرامیٹر کے ذریعے تبدیل کیا جاتا ہے۔ ہمارے نتائج واحد ذرہ کی پیش گوئیوں سے واضح طور پر مختلف ہیں، جو کھینچنے والی تعاملات کی وجہ سے نچلی ترین توانائی کی حالت میں خود پھنسنے کی وجہ سے اضافی مقامیکاری اور سب سے زیادہ توانائی والی حالت میں اسکریننگ کی وجہ سے مقامیکاری میں رکاوٹ کی عکاسی کرتے ہیں۔ اس مطالعہ سے خود-دوقطبیت کی وجہ سے پیدا ہونے والی حرکت پذیری کی حد پر تعاملات کے اثرات کے مقداری مطالعات کی راہ ہموار ہوتی ہے۔
|
ur
|
End of preview. Expand
in Data Studio
README.md exists but content is empty.
- Downloads last month
- 9