检索结果-内蒙古大学图书馆

Probabilistic margins for instance reweighting in adversarial training 21

学校读者我要写书评

暂无评论

Probabilistic margins for instance reweighting in adversaria...

Proceedings of the 35th International Conference on Neural Information Processing systems

作者： Qizhou Wang Feng Liu Bo Han Tongliang Liu Chen Gong Gang Niu Mingyuan Zhou Masashi Sugiyama Department of Computer Science Hong Kong Baptist University DeSI Lab Australian Artificial Intelligence Institute University of Technology Sydney TML Lab School of Computer Science Faculty of Engineering The University of Sydney PCA Lab Key Lab of Intelligent Perception and Systems for High-Dimensional Information of MoE and Jiangsu Key Lab of Image and Video Understanding for Social Security School of Computer Science and Engineering Nanjing University of Science and Technology RIKEN Center for Advanced Intelligence Project (AIP) McCombs School of Business The University of Texas at Austin RIKEN Center for Advanced Intelligence Project (AIP) and Graduate School of Frontier Sciences The University of Tokyo

ISBN: (纸本)9781713845393

Reweighting adversarial data during training has been recently shown to improve adversarial robustness, where data closer to the current decision boundaries are regarded as more critical and given larger weights. However, existing methods measuring the closeness are not very reliable: they are discrete and can take only a few values, and they are path-dependent, i.e., they may change given the same start and end points with different attack paths. In this paper, we propose three types of probabilistic margin (PM), which are continuous and path-independent, for measuring the aforementioned closeness and reweighting adversarial data. Specifically, a PM is defined as the difference between two estimated class-posterior probabilities, e.g., such a probability of the true label minus the probability of the most confusing label given some natural data. Though different PMs capture different geometric properties, all three PMs share a negative correlation with the vulnerability of data: data with larger/smaller PMs are safer/riskier and should have smaller/larger weights. Experiments demonstrated that PMs are reliable and PM-based reweighting methods outperformed state-of-the-art counterparts.

关键词：

Finding shortest and nearly shortest path nodes in large substantially incomplete networks

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Kitsak, Maksim Ganin, Alexander Elmokashfi, Ahmed Cui, Hongzhu Eisenberg, Daniel A. Alderson, David L. Korkin, Dmitry Linkov, Igor Faculty of Electrical Engineering Mathematics and Computer Science Delft University of Technology Delft Netherlands Network Science Institute Northeastern University BostonMA02115 United States Department of Systems and Information Engineering University of Virginia CharlottesvilleVA22904 United States U.S. Army Engineer Research and Development Center Contractor Concord MA01742 United States Simula Metropolitan Center for Digital Engineering Oslo Norway Bioinformatics and Computational Biology Program Worcester Polytechnic Institute WorcesterMA01609 United States Institute for Genomic Medicine Columbia University Medical Center New YorkNY United States Operations Research Department Naval Postgraduate School MontereyCA93943 United States Data Science Program Worcester Polytechnic Institute WorcesterMA01609 United States Computer Science Department Worcester Polytechnic Institute WorcesterMA01609 United States U.S. Army Engineer Research and Development Center Environmental Laboratory Concord MA01742 United States

Dynamic processes on networks, be it information transfer in the Internet, contagious spreading in a social network, or neural signaling, take place along shortest or nearly shortest paths. Unfortunately, our maps of most large networks are substantially incomplete due to either the highly dynamic nature of networks, or high cost of network measurements, or both, rendering traditional path finding methods inefficient. We find that shortest paths in large real networks, such as the network of protein-protein interactions (PPI) and the Internet at the autonomous system (AS) level, are not random but are organized according to latent-geometric rules. If nodes of these networks are mapped to points in latent hyperbolic spaces, shortest paths in them align along geodesic curves connecting endpoint nodes. We find that this alignment is sufficiently strong to allow for the identification of shortest path nodes even in the case of substantially incomplete networks. We demonstrate the utility of latent-geometric path finding in problems of cellular pathway reconstruction and communication security. Copyright © 2022, The Authors. All rights reserved.

关键词： Proteins

Retrospective for the Dynamic Sensorium Competition for predicting large-scale mouse primary visual cortex activity from videos

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Turishcheva, Polina Fahey, Paul G. Vystrčilová, Michaela Hansel, Laura Froebe, Rachel Ponder, Kayla Qiu, Yongrong Willeke, Konstantin F. Bashiri, Mohammad Baikulov, Ruslan Zhu, Yu Ma, Lei Yu, Shan Huang, Tiejun Li, Bryan M. De Wulf, Wolf Kudryashova, Nina Hennig, Matthias H. Rochefort, Nathalie L. Onken, Arno Wang, Eric Ding, Zhiwei Tolias, Andreas S. Sinz, Fabian H. Ecker, Alexander S. Institute of Computer Science University of Göttingen Campus Institute Data Science Germany Department of Neuroscience Center for Neuroscience and Artificial Intelligence Baylor College of Medicine HoustonTX United States Department of Ophthalmology Byers Eye Institute Stanford University School of Medicine StanfordCA United States Stanford Bio-X Stanford University StanfordCA United States Wu Tsai Neurosciences Institute Stanford University StanfordCA United States International Max Planck Research School for Intelligent Systems Tübingen Germany Institute for Bioinformatics and Medical Informatics Tübingen University Germany lRomul Russia Institute of Automation Chinese Academy of Sciences China Beijing Academy of Artificial Intelligence China The Alan Turing Institute United Kingdom School of Informatics University of Edinburgh United Kingdom Centre for Discovery Brain Sciences University of Edinburgh United Kingdom Simons Initiative for the Developing Brain University of Edinburgh United Kingdom Department of Electrical Engineering Stanford University StanfordCA United States Max Planck Institute for Dynamics and Self-Organization Göttingen Germany

Understanding how biological visual systems process information is challenging because of the nonlinear relationship between visual input and neuronal responses. Artificial neural networks allow computational neuroscientists to create predictive models that connect biological and machine vision. Machine learning has benefited tremendously from benchmarks that compare different model on the same task under standardized conditions. However, there was no standardized benchmark to identify state-of-the-art dynamic models of the mouse visual system. To address this gap, we established the SENSORIUM 2023 Benchmark Competition with dynamic input, featuring a new large-scale dataset from the primary visual cortex of ten mice. This dataset includes responses from 78,853 neurons to 2 hours of dynamic stimuli per neuron, together with the behavioral measurements such as running speed, pupil dilation, and eye movements. The competition ranked models in two tracks based on predictive performance for neuronal responses on a held-out test set: one focusing on predicting in-domain natural stimuli and another on out-of-distribution (OOD) stimuli to assess model generalization. As part of the NeurIPS 2023 competition track, we received more than 160 model submissions from 22 teams. Several new architectures for predictive models were proposed, and the winning teams improved the previous state-of-the-art model by 50%. Access to the dataset as well as the benchmarking infrastructure will remain online at ***. © 2024, CC BY-NC-ND.

关键词： Neurons

MOST DISCRIMINATIVE STIMULI FOR FUNCTIONAL CELL TYPE CLUSTERING

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Burg, Max F. Zenkel, Thomas Vystrčilová, Michaela Oesterle, Jonathan Höfling, Larissa Willeke, Konstantin F. Lause, Jan Müller, Sarah Fahey, Paul G. Ding, Zhiwei Restivo, Kelli Sridhar, Shashwat Gollisch, Tim Berens, Philipp Tolias, Andreas S. Euler, Thomas Bethge, Matthias Ecker, Alexander S. International Max Planck Research School for Intelligent Systems Tübingen Germany Institute of Computer Science University of Göttingen Campus Institute Data Science Germany Tübingen AI Center University of Tübingen Germany Institute of Ophthalmic Research University of Tübingen Germany Centre for Integrative Neuroscience University of Tübingen Germany Institute for Bioinformatics and Medical Informatics Tübingen University Germany Hertie Institute for AI in Brain Health University of Tübingen Germany Department of Neuroscience Baylor College of Medicine HoustonTX United States Center for Neuroscience and Artificial Intelligence Baylor College of Medicine HoustonTX United States University Medical Center Göttingen Department of Ophthalmology Germany Bernstein Center for Computational Neuroscience Göttingen Germany University of Göttingen Germany Department of Electrical and Computer Engineering Rice University HoustonTX United States Max Planck Institute for Dynamics and Self-Organization Göttingen Germany

Identifying cell types and understanding their functional properties is crucial for unraveling the mechanisms underlying perception and cognition. In the retina, functional types can be identified by carefully selected stimuli, but this requires expert domain knowledge and biases the procedure towards previously known cell types. In the visual cortex, it is still unknown what functional types exist and how to identify them. Thus, for unbiased identification of the functional cell types in retina and visual cortex, new approaches are needed. Here we propose an optimization-based clustering approach using deep predictive models to obtain functional clusters of neurons using Most Discriminative Stimuli (MDS). Our approach alternates between stimulus optimization with cluster reassignment akin to an expectation-maximization algorithm. The algorithm recovers functional clusters in mouse retina, marmoset retina and macaque visual area V4. This demonstrates that our approach can successfully find discriminative stimuli across species, stages of the visual system and recording techniques. The resulting most discriminative stimuli can be used to assign functional cell types fast and on the fly, without the need to train complex predictive models or show a large natural scene dataset, paving the way for experiments that were previously limited by experimental time. Crucially, MDS are interpretable: they visualize the distinctive stimulus patterns that most unambiguously identify a specific type of neuron. © 2023, CC BY-NC-SA.

关键词： Maximum principle

Time Series Forecasting of New Cases and New Deaths Rate for COVID-19 using Deep Learning Methods

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Ayoobi, Nooshin Sharifrazi, Danial Alizadehsani, Roohallah Shoeibi, Afshin Gorriz, Juan M. Moosaei, Hossein Khosravi, Abbas Nahavandi, Saeid Chofreh, Abdoulmohammad Gholamzadeh Goni, Feybi Ariani Klemeš, Jiří Jaromír Mosavi, Amir Department of Mathematics Savitribai Phule Pune University Pune411007 India Department of Computer Engineering School of Technical and Engineering Shiraz Branch Islamic Azad University Shiraz Iran Institute for Intelligent Systems Research and Innovation Deakin University GeelongVIC3216 Australia Computer Engineering Department Ferdowsi University of Mashhad Mashhad Iran Faculty of Electrical and Computer Engineering Biomedical Data Acquisition Lab K. N. Toosi University of Technology Tehran Iran Department of Signal Theory Networking and Communications Universidad de Granada Spain Department of Mathematics Faculty of Science University of Bojnord Iran Sustainable Process Integration Laboratory - SPIL NETME Centre Faculty of Mechanical Engineering Brno University of Technology - VUT Brno Technická 2896/2 Brno616 69 Czech Republic Department of Management Faculty of Business and Management Brno University of Technology VUT Brno Kolejní 2906/4 Brno612 00 Czech Republic Faculty of Civil Engineering Technische Universität Dresden Dresden01069 Germany John von Neumann Faculty of Informatics Obuda University Budapest1034 Hungary School of Economics and Business Norwegian University of Life Sciences 1430 Norway

The first known case of Coronavirus disease 2019 (COVID-19) was identified in December 2019. It has spread worldwide, leading to an ongoing pandemic, imposed restrictions and costs to many countries. Predicting the number of new cases and deaths during this period can be a useful step in predicting the costs and facilities required in the future. The purpose of this study is to predict new cases and deaths rate one, three and seven-day ahead during the next 100 days. The motivation for predicting every n days (instead of just every day) is the investigation of the possibility of computational cost reduction and still achieving reasonable performance. Such a scenario may be encountered real-time forecasting of time series. Six different deep learning methods are examined on the data adopted from the WHO website. Three methods are LSTM, Convolutional LSTM, and GRU. The bidirectional extension is then considered for each method to forecast the rate of new cases and new deaths in Australia and Iran countries. This study is novel as it carries out a comprehensive evaluation of the aforementioned three deep learning methods and their bidirectional extensions to perform prediction on COVID-19 new cases and new death rate time series. To the best of our knowledge, this is the first time that Bi-GRU and Bi-Conv-LSTM models are used for prediction on COVID-19 new cases and new deaths time series. The evaluation of the methods is presented in the form of graphs and Friedman statistical test. The results show that the bidirectional models have lower errors than other models. Several error evaluation metrics are presented to compare all models, and finally, the superiority of bidirectional methods is determined. This research could be useful for organizations working against COVID-19 and determining their long-term plans. Copyright © 2021, The Authors. All rights reserved.

关键词： Coronavirus

Waveform optimization with multiple performance metrics for broadband joint communication and radar sensing

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Ni, Zhitong Andrew Zhang, J. Yang, Kai Huang, Xiaojing Tsiftsis, Theodoros A. School of Information and Electronics Beijing Institute of Technology Beijing100081 China Global Big Data Technologies Centre University of Technology SydneyNSW2007 Australia Institute of Physical Internet school of Intelligent Systems Science and Engineering Jinan University Zhuhai Campus Zhuhai519070 China

Joint communication and radar sensing (JCAS) integrates communication and radar/radio sensing into one system, sharing one transmitted signal. In this paper, we investigate JCAS waveform optimization underlying communication signals, where a base station detects radar targets and communicates with mobile users simultaneously. We first develop individual novel waveform optimization problems for communications and sensing, respectively. For communications, we propose a novel lower bound of sum rate by integrating multi-user interference and effective channel gain into one metric that simplifies the optimization of the sum rate. For radar sensing, we consider optimizing one of two metrics, the mutual information or the Cramér-Rao bound. Then, we formulate the JCAS problem by optimizing the communication metric under different constraints of the radar metric, and we obtain both closed-form solutions and iterative solutions to the non-convex JCAS optimization problem. Numerical results are provided and verify the proposed optimization solutions. © 2020, CC-BY.

关键词： Optimization

DeepKAF: A Heterogeneous CBR & Deep Learning Approach for NLP Prototyping

学校读者我要写书评

暂无评论

DeepKAF: A Heterogeneous CBR & Deep Learning Approach for NL...

IEEE International Symposium on Innovations in intelligent systems and Applications (INISTA)

作者： Kareem Amin Stelios Kapetanakis Nikolaos Polatidis Klaus-Dieter Althoff Andreas Dengel Smart Data and Knowledge Services German Research Center for Artificial Intelligence Technische Universität Kaiserslautern Kaiserslautern Germany School of Computing Engineering and Mathematics University of Brighton Brighton United Kingdom Intelligent Information Systems Lab Institute of Computer Science University of Hildesheim Hildesheim Germany

ISBN: (数字)9781728167992

ISBN: (纸本)9781728168005

With widespread modernization, digitization and transformations of most of industries, Artificial Intelligence (AI) has become the key enabler in that modernization journey. AI offers substantial capabilities to solve new problems and optimise existing solutions specialising on specific problems and learning from different domains. AI solutions can be either explainable or black box ones with the latter being urged to improve since they cannot trust. Case-based Reasoning (CBR) is an explainable AI approach where solutions are provided along with relevant explanations in terms of why a solution was selected. However, CBR, like most other explainable approaches, has several limitations in terms of scalability, large data volumes, domain complexity, that reduce its ability to scale any CBR system in industrial applications. In this paper, we provide a heterogeneous CBR framework - DeepKAF where we combine CBR paradigm with Deep Learning architectures to solve complicated Natural Language Processing (NLP) problems (eg. mixed language and grammatically incorrect text).DeepKAF is built based on continuous research in the area of Deep Learning and CBR. DeepKAF has been implemented and used across different domains, test use cases and research models as an ensemble deep learning and CBR Architecture.

关键词： Machine learning Training Neural networks Natural language processing Cognition Buildings Task analysis

From ancient times to modern: realizing the power of data visualization in healthcare and medicine

学校读者我要写书评

暂无评论

Big data Analytics 2020年第1期5卷 1-7页

作者： Elgendi, Mohamed Howard, Newton Hussain, Amir Menon, Carlo Ward, Rabab School of Electrical and Computer Engineering University of British Columbia Vancouver Canada Faculty of Medicine University of British Columbia Vancouver Canada School of Mechatronic Systems Engineering Simon Fraser University Burnaby Canada Nuffield Department of Surgical Sciences University of Oxford Oxford UK Data Science and Cyber Analytics Research Group Edinburgh Napier University Edinburgh UK

data visuals (scientific images) display and express various amounts and types of information, and, as the saying goes,“an image is worth 1,000 words.” Based on a review of two studies, a new estimation of how many words an image is actually worth was calculated in an attempt to quantify the complicated biological process of image perception. The result revealed that an image is actually worth more than 30,000 words. This new value estimation provides insight into the power of images. Given that figures, graphs, and data visualizations are types of images commonly used in research and publications, every produced figure is important and must be carefully considered during the publication process.

关键词：

Hysteresis Control of the Pseudo Boost PFC Converter

学校读者我要写书评

暂无评论

Hysteresis Control of the Pseudo Boost PFC Converter

IEEE International Symposium on Industrial Electronics (ISIE)

作者： Aleksandra Lekić-Vervoort Milovan Majstorović Leposava Ristić Dušan Stipanović Intelligent Electrical Power Grids (IEPG) Electrical Sustainable Energy (ESE) Faculty of Electrical Engineering Mathematics and Computer Science TU Delft Delft Netherlands The Department of Power Converters and Drive Systems University of Belgrade Belgrade Serbia Coordinated Science Laboratory University of Illinois Urbana IL USA

ISBN: (数字)9781728156354

ISBN: (纸本)9781728156361

In this paper, a control design for the the Pseudo Boost PFC converter with the application in the charger for electric vehicles, is formulated and verified. The rectifier ends up being energy efficient because due to containing smaller number of switches. The converter is controlled using a method involving the minimum Lyapunov function derivative and a hysteresis. This control provides an averaged value of the converter sine input current, and thus, decreasing the THD value. Theoretical results are complemented with the corresponding simulations and experimental waveforms.

关键词： Hysteresis Switches Lyapunov methods Inductors Mathematical model Electric vehicles