机器学习学术速递[2021.6.7]

访问arxivdaily.com获取含摘要速递，更有收藏、搜索等功能，涵盖CS|物理|数学|经济|统计|金融|生物|电气领域
同步公众号:arXiv每日学术速递，欢迎关注

cs.LG 方向，今日共计126篇

Graph相关(图学习|图神经网络|图优化等)(9篇)

【1】 Graph-based Deep Learning for Communication Networks: A Survey
标题：基于图的通信网络深度学习研究综述
作者：Weiwei Jiang
机构：Department of Electronic Engineering, Tsinghua University, Beijing , China
备注：Github link: this https URL
链接：https://arxiv.org/abs/2106.02533

【2】 Graph Barlow Twins: A self-supervised representation learning framework for graphs
标题：Graph Barlow TWINS：一种图的自监督表示学习框架
作者：Piotr Bielak,Tomasz Kajdanowicz,Nitesh V. Chawla
机构：Department of Computational Intelligence, Wroclaw University of Science and Technology, Wrocław, Poland, Department of Computer Science and Engineering, University of Notre Dame, Notre Dame, IN, USA
链接：https://arxiv.org/abs/2106.02466

【3】 Proving Equivalence Between Complex Expressions Using Graph-to-Sequence Neural Models
标题：用图到序列神经模型证明复杂表达式之间的等价性
作者：Steve Kommrusch,Théo Barollet,Louis-Noël Pouchet
备注：10 pages (24 including references and appendices), 8 figures, 17 tables. arXiv admin note: substantial text overlap with arXiv:2002.06799
链接：https://arxiv.org/abs/2106.02452

【4】 A Deep Local and Global Scene-Graph Matching for Image-Text Retrieval
标题：一种用于图文检索的局部和全局深度场景图匹配算法
作者：Manh-Duy Nguyen,Binh T. Nguyen,Cathal Gurrin
机构：School of Computing, Dublin, Ireland, AISIA Research Lab, University of Science, Ho Chi Minh City, Vietnam, Vietnam National University Ho Chi Minh City, Vietnam
链接：https://arxiv.org/abs/2106.02400

【5】 Stochastic Iterative Graph Matching
标题：随机迭代图匹配
作者：Linfeng Liu,Michael C. Hughes,Soha Hassoun,Li-Ping Liu
机构： such as the total number of matched 1Department of Computer Science, Tufts University, USA 2Department of Chemical and Biological Engineering, TuftsUniversity
备注：To appear in ICML 2021
链接：https://arxiv.org/abs/2106.02206

【6】 Spatial Graph Attention and Curiosity-driven Policy for Antiviral Drug Discovery
标题：空间图形注意和好奇心驱动策略在抗病毒药物研发中的应用
作者：Yulun Wu,Nicholas Choma,Andrew Chen,Mikaela Cashman,Érica T. Prates,Manesh Shah,Verónica G. Melesse Vergara,Austin Clyde,Thomas S. Brettin,Wibe A. de Jong,Neeraj Kumar,Martha S. Head,Rick L. Stevens,Peter Nugent,Daniel A. Jacobson,James B. Brown
机构：† University of California, US Department of Energy, ¶ Universityof Tennessee, †† University of Chicago
链接：https://arxiv.org/abs/2106.02190

【7】 Counterfactual Graph Learning for Link Prediction
标题：用于链接预测的反事实图学习
作者：Tong Zhao,Gang Liu,Daheng Wang,Wenhao Yu,Meng Jiang
机构：University of Notre Dame, Notre Dame, IN
链接：https://arxiv.org/abs/2106.02172

【8】 Price graphs: Utilizing the structural information of financial time series for stock prediction
标题：价格图：利用金融时间序列的结构信息进行股票预测
作者：Junran Wu,Ke Xu,Xueyuan Chen,Shangzhe Li,Jichang Zhao
机构：State Key Lab of Software Development Environment, Beihang University, School of Mathematics Science, Beihang University, School of Economics and Management, Beihang University
链接：https://arxiv.org/abs/2106.02522

【9】 Laplacian-Based Dimensionality Reduction Including Spectral Clustering, Laplacian Eigenmap, Locality Preserving Projection, Graph Embedding, and Diffusion Map: Tutorial and Survey
标题：基于拉普拉斯的降维，包括谱聚类、拉普拉斯特征映射、保局投影、图形嵌入和扩散映射：教程和综述
作者：Benyamin Ghojogh,Ali Ghodsi,Fakhri Karray,Mark Crowley
机构：Department of Electrical and Computer Engineering, Machine Learning Laboratory, University of Waterloo, Waterloo, ON, Canada, Department of Statistics and Actuarial Science & David R. Cheriton School of Computer Science
备注：To appear as a part of an upcoming textbook on dimensionality reduction and manifold learning
链接：https://arxiv.org/abs/2106.02154

GAN|对抗|攻击|生成相关(7篇)

【1】 Forward Super-Resolution: How Can GANs Learn Hierarchical Generative Models for Real-World Distributions
标题：正向超分辨率：甘斯如何学习现实世界分布的分层生成模型
作者：Zeyuan Allen-Zhu,Yuanzhi Li
机构：Microsoft Research Redmond, Carnegie Mellon University, (version ,)
链接：https://arxiv.org/abs/2106.02619

【2】 A Procedural World Generation Framework for Systematic Evaluation of Continual Learning
标题：持续学习系统评价的过程性世界生成框架
作者：Timm Hess,Martin Mundt,Iuliia Pliushch,Visvanathan Ramesh
机构：Goethe University, Frankfurt am Main, Germany
链接：https://arxiv.org/abs/2106.02585

【3】 F-Drop&Match: GANs with a Dead Zone in the High-Frequency Domain
标题：F-Drop&Match：高频带死区的Gans
作者：Shin'ya Yamaguchi,Sekitoshi Kanai
机构：NTT
备注：Under Review
链接：https://arxiv.org/abs/2106.02343

【4】 Temporally coherent video anonymization through GAN inpainting
标题：基于GaN修复的时间相干视频匿名化
作者：Thangapavithraa Balaji,Patrick Blies,Georg Göri,Raphael Mitsch,Marcel Wasserer,Torsten Schön
机构：and Torsten Sch¨on
备注：Preprint of our FG2021 submission
链接：https://arxiv.org/abs/2106.02328

【5】 A Little Robustness Goes a Long Way: Leveraging Universal Features for Targeted Transfer Attacks
标题：一点健壮性大有用武之地：利用通用功能进行有针对性的传输攻击
作者：Jacob M. Springer,Melanie Mitchell,Garrett T. Kenyon
机构：Los Alamos National Laboratory, Los Alamos, NM, Santa Fe Institute, Santa Fe, NM
备注：25 pages, 13 figures, 3 tables
链接：https://arxiv.org/abs/2106.02105

【6】 Learning Hard Optimization Problems: A Data Generation Perspective
标题：学习困难优化问题：一种数据生成的观点
作者：James Kotary,Ferdinando Fioretto,Pascal Van Hentenryck
机构：Syracuse University, Georgia Institute of Technology
链接：https://arxiv.org/abs/2106.02601

【7】 Fre-GAN: Adversarial Frequency-consistent Audio Synthesis
标题：Fre-GAN：对抗性频率一致音频合成
作者：Ji-Hoon Kim,Sang-Hoon Lee,Ji-Hyun Lee,Seong-Whan Lee
机构：Department of Artificial Intelligence, Korea University, Seoul, Korea, Department of Brain and Cognitive Engineering, Korea University, Seoul, Korea
备注：Accepted paper in Interspeech 2021
链接：https://arxiv.org/abs/2106.02297

半/弱/无/有监督|不确定性|主动学习(9篇)

【1】 Self-Attention Between Datapoints: Going Beyond Individual Input-Output Pairs in Deep Learning
标题：数据点之间的自我注意：深度学习中超越个体输入-输出对
作者：Jannik Kossen,Neil Band,Clare Lyle,Aidan N. Gomez,Tom Rainforth,Yarin Gal
机构：Aidan Gomez, OATML, Department of Computer Science, University of Oxford, Department of Statistics, University of Oxford, Cohere
备注：First two authors contributed equally
链接：https://arxiv.org/abs/2106.02584

【2】 How to select and use tools? : Active Perception of Target Objects Using Multimodal Deep Learning
标题：如何选择和使用工具？：基于多模态深度学习的目标对象主动感知
作者：Namiko Saito,Tetsuya Ogata,Satoshi Funabashi,Hiroki Mori,Shigeki Sugano
备注：Best Paper Award of Cognitive Robotics in ICRA2021 IEEE Robotics and Automation Letters 2021, Proceedings of the 2021 International Conference on Robotics and Automation (ICRA 2021), 2021
链接：https://arxiv.org/abs/2106.02445

【3】 A Novel Semi-supervised Framework for Call Center Agent Malpractice Detection via Neural Feature Learning
标题：一种新的基于神经特征学习的呼叫中心座席故障检测半监督框架
作者：Şükrü Ozan,Leonardo Obinna Iheme
机构：Folkart Towers B Block,th Floor, Office , Bayraklı, ˙Izmir, Turkey, phone: +
链接：https://arxiv.org/abs/2106.02433

【4】 Online reinforcement learning with sparse rewards through an active inference capsule
标题：基于主动推理胶囊的稀疏奖励在线强化学习
作者：Alejandro Daniel Noel,Charel van Hoof,Beren Millidge
机构：Department of Cognitive Robotics, Delft University of Technology, MRC Brain Network Dynamics Unit, University of Oxford
链接：https://arxiv.org/abs/2106.02390

【5】 Fairness-Aware Unsupervised Feature Selection
标题：公平性感知的无监督特征选择
作者：Xiaoying Xing,Hongfu Liu,Chen Chen,Jundong Li
机构：Tsinghua University, Beijing, China , Brandeis University, Waltham, MA, USA , University of Virginia, Charlottesville, VA, USA
链接：https://arxiv.org/abs/2106.02216

【6】 Top-$k$ Regularization for Supervised Feature Selection
标题：有监督特征选择的Top-$k$正则化算法
作者：Xinxing Wu,Qiang Cheng
机构：University of Kentucky, Lexington, Kentucky, USA
备注：12 pages
链接：https://arxiv.org/abs/2106.02197

【7】 Self-supervised Dialogue Learning for Spoken Conversational Question Answering
标题：口语问答中的自监督对话学习
作者：Nuo Chen,Chenyu You,Yuexian Zou
机构：ADSPLAB, School of ECE, Peking University, Shenzhen, China, Department of Electrical Engineering, Yale University, CT, USA, Peng Cheng Laboratory, Shenzhen, China
备注：To Appear Interspeech 2021
链接：https://arxiv.org/abs/2106.02182

【8】 Unsupervised Learning of General-Purpose Embeddings for Code Changes
标题：代码更改通用嵌入的无监督学习
作者：Mikhail Pravilov,Egor Bogomolov,Yaroslav Golubev,Timofey Bryksin
机构：∗JetBrains Research, Saint Petersburg, Russia, §Higher School of Economics, Saint Petersburg, Russia
备注：6 pages, 2 figures
链接：https://arxiv.org/abs/2106.02087

【9】 Segmental Contrastive Predictive Coding for Unsupervised Word Segmentation
标题：分段对比预测编码在无监督分词中的应用
作者：Saurabhchand Bhati,Jesús Villalba,Piotr Żelasko,Laureano Moro-Velazquez,Najim Dehak
机构：†Center for Language and Speech Processing, ‡Human Language Technology Center of Excellence, Johns Hopkins University, Baltimore, MD, USA
链接：https://arxiv.org/abs/2106.02170

迁移|Zero/Few/One-Shot|自适应(8篇)

【1】 Entity Concept-enhanced Few-shot Relation Extraction
标题：实体概念增强型少发关系抽取
作者：Shan Yang,Yongfei Zhang,Guanglin Niu,Qinghua Zhao,Shiliang Pu
机构：Beijing Key Laboratory of Digital Media, School of Computer Science and Engineering, State Key Laboratory of Virtual Reality Technology and Systems, BeiHang University,Beijing , China, Pengcheng Laboratory, Shenzhen , China
备注：Accepted at ACL2021
链接：https://arxiv.org/abs/2106.02401

【2】 A Survey on Deep Domain Adaptation for LiDAR Perception
标题：激光雷达感知的深域自适应研究综述
作者：Larissa T. Triess,Mariella Dreissig,Christoph B. Rist,J. Marius Zöllner
机构：J. Marius Z¨ollner
备注：Accepted at IEEE Intelligent Vehicles Symposium (IV) 2021 Workshop on Autonomy at Scale. 8 pages, 5 figures
链接：https://arxiv.org/abs/2106.02377

【3】 Differentiable Dynamic Quantization with Mixed Precision and Adaptive Resolution
标题：混合精度和自适应分辨率的可微动态量化
作者：Zhang Zhaoyang,Shao Wenqi,Gu Jinwei,Wang Xiaogang,Luo Ping
机构： 1The Chinese University of Hong Kong 2SenseBrain, Ltd 3Hong Kong University
备注：Accepted by ICML 2021
链接：https://arxiv.org/abs/2106.02295

【4】 Robustifying Reinforcement Learning Policies with $\mathcal{L}_1$ Adaptive Control
作者：Yikun Cheng,Pan Zhao,Manan Gandhi,Bo Li,Evangelos Theodorou,Naira Hovakimyan
链接：https://arxiv.org/abs/2106.02249

【5】 Materials Representation and Transfer Learning for Multi-Property Prediction
标题：用于多属性预测的材料表示和转移学习
作者：Shufeng Kong,Dan Guevarra,Carla P. Gomes,John M. Gregoire
机构：)Department of Computer Science, Cornell University, Ithaca, NY, USA, )Division of Engineering and Applied Science, California Institute of Technology, Pasadena, CA
备注：accepted at the Applied Physics Reviews journal
链接：https://arxiv.org/abs/2106.02225

【6】 Adam in Private: Secure and Fast Training of Deep Neural Networks with Adaptive Moment Estimation
标题：ADAM私下：自适应矩估计深度神经网络的安全快速训练
作者：Nuttapong Attrapadung,Koki Hamada,Dai Ikarashi,Ryo Kikuchi,Takahiro Matsuda,Ibuki Mishina,Hiraku Morita,Jacob C. N. Schuldt
机构：∗National Institute of Advanced Industrial Science and Technology, †NTT, ‡University of St. Gallen
备注：24 pages, 13 tables
链接：https://arxiv.org/abs/2106.02203

【7】 Cross-Trajectory Representation Learning for Zero-Shot Generalization in RL
标题：RL零点泛化的交叉轨迹表示学习
作者：Bogdan Mazoure,Ahmed M. Ahmed,Patrick MacAlpine,R Devon Hjelm,Andrey Kolobov
机构：McGill University, Quebec AI Institute, Stanford University, Sony AI, Université de Montréal, Quebec AI Institute, Microsoft Research
链接：https://arxiv.org/abs/2106.02193

【8】 Adaptive Epidemic Forecasting and Community Risk Evaluation of COVID-19
标题：冠状病毒自适应流行预测与社区风险评估
作者：Vishrawas Gopalakrishnan,Sayali Navalekar,Pan Ding,Ryan Hooley,Jacob Miller,Raman Srinivasan,Ajay Deshpande,Xuan Liu,Simone Bianco,James H. Kaufman
机构：IBM, USA
备注：9 pages, 11 figures
链接：https://arxiv.org/abs/2106.02094

强化学习(4篇)

【1】 Model-agnostic and Scalable Counterfactual Explanations via Reinforcement Learning
标题：基于强化学习的模型不可知且可扩展的反事实解释
作者：Robert-Florian Samoilescu,Arnaud Van Looveren,Janis Klaise
机构：Seldon Technologies, London, UK
备注：18 pages
链接：https://arxiv.org/abs/2106.02597

【2】 RL-DARTS: Differentiable Architecture Search for Reinforcement Learning
标题：RL-DARTS：强化学习的可微结构搜索
作者：Yingjie Miao,Xingyou Song,Daiyi Peng,Summer Yue,Eugene Brevdo,Aleksandra Faust
机构：Google Research, Brain Team
备注：19 pages total, 17 figures
链接：https://arxiv.org/abs/2106.02229

【3】 Celebrating Diversity in Shared Multi-Agent Reinforcement Learning
标题：共享多智能体强化学习中的多样性庆祝
作者：Chenghao Li,Chengjie WU,Tonghan Wang,Jun Yang,Qianchuan Zhao,Chongjie Zhang
机构：IIIS, Tsinghua University
链接：https://arxiv.org/abs/2106.02195

【4】 A Consciousness-Inspired Planning Agent for Model-Based Reinforcement Learning
标题：一种基于模型强化学习的意识启发式规划Agent
作者：Mingde Zhao,Zhen Liu,Sitao Luan,Shuyuan Zhang,Doina Precup,Yoshua Bengio
机构：McGill University; ,Université de Montréal; ,DeepMind; ,Mila, ∗: Equal Contribution, †: Equal Supervision
链接：https://arxiv.org/abs/2106.02097

医学相关(2篇)

【1】 Ambulatory blood pressure monitoring versus office blood pressure measurement: Are there sex differences?
标题：动态血压监测与办公室血压测量：是否存在性别差异？
作者：Aleksandar Miladinović,Miloš Ajčević,Giulia Siveri,Laura Liguori,Lorenzo Pascazio,Agostino Accardo
机构：Institute for Maternal and Child Health – IRCCS, Burlo Garofolo, Trieste, Italy, Department of Engineering and Architecture at the University of Trieste, Trieste, Italy
链接：https://arxiv.org/abs/2106.02392

【2】 A Prospective Observational Study to Investigate Performance of a Chest X-ray Artificial Intelligence Diagnostic Support Tool Across 12 U.S. Hospitals
标题：一项前瞻性观察性研究，旨在调查美国12家医院胸部X射线人工智能诊断支持工具的性能
作者：Ju Sun,Le Peng,Taihui Li,Dyah Adila,Zach Zaiman,Genevieve B. Melton,Nicholas Ingraham,Eric Murray,Daniel Boley,Sean Switzer,John L. Burns,Kun Huang,Tadashi Allen,Scott D. Steenburg,Judy Wawira Gichoya,Erich Kummerfeld,Christopher Tignanelli
机构：Author Affiliations:, Department of Computer Science and Engineering, University of Minnesota, Minneapolis, MN, Department of Computer Science, Emory University, Atlanta, GA, Institute for Health Informatics, University of Minnesota, Minneapolis, MN
链接：https://arxiv.org/abs/2106.02118

聚类(2篇)

【1】 Fuzzy Clustering with Similarity Queries
标题：基于相似查询的模糊聚类
作者：Wasim Huleihel,Arya Mazumdar,Soumyabrata Pal
备注：47 pages, 7 figures
链接：https://arxiv.org/abs/2106.02212

【2】 Manifold-Aware Deep Clustering: Maximizing Angles between Embedding Vectors Based on Regular Simplex
标题：流形感知的深度聚类：基于正则单纯形的嵌入向量夹角最大化
作者：Keitaro Tanaka,Ryosuke Sawata,Shusuke Takahashi
机构：†Waseda University, Japan, ‡Sony Corporation, Japan
备注：Accepted by Interspeech 2021
链接：https://arxiv.org/abs/2106.02331

自动驾驶|车辆|车道检测等(1篇)

【1】 Intelligent Transportation Systems to Mitigate Road Traffic Congestion
标题：智能交通系统缓解道路交通拥堵
作者：Nizar Hamadeh,Ali Karouni,Zeinab Farhat,Hussein El Ghor,Mohamad El Ghor,Israa Katea
机构：LENS Laboratory, Lebanese University, Department of Computer and Communication Engineer, Saida, Lebanon., Research Assistant, LENS Laboratory., )
链接：https://arxiv.org/abs/2106.02315

联邦学习|隐私保护|加密(2篇)

【1】 FedCCEA : A Practical Approach of Client Contribution Evaluation for Federated Learning
标题：FedCCEA：一种实用的联合学习客户贡献评估方法
作者：Sung Kuk Shyn,Donghee Kim,Kwangsu Kim
机构：Department of Artificial Intelligence, Sungkyunkwan University, Suwon, South Korea, College of Computing
链接：https://arxiv.org/abs/2106.02310

【2】 Local Adaptivity in Federated Learning: Convergence and Consistency
标题：联合学习中的局部自适应：收敛性和一致性
作者：Jianyu Wang,Zheng Xu,Zachary Garrett,Zachary Charles,Luyang Liu,Gauri Joshi
机构：†Carnegie Mellon University, §Google Research
链接：https://arxiv.org/abs/2106.02305

推理|分析|理解|解释(10篇)

【1】 A Holistic Approach to Interpretability in Financial Lending: Models, Visualizations, and Summary-Explanations
标题：金融贷款可解释性的整体方法：模型、可视化和摘要解释
作者：Chaofan Chen,Kangcheng Lin,Cynthia Rudin,Yaron Shaposhnik,Sijia Wang,Tong Wang
机构：University of Maine, University of Illinois, Urbana-Champaign, Duke University, University of Rochester, University of Iowa
链接：https://arxiv.org/abs/2106.02605

【2】 Stochastic gradient descent with noise of machine learning type. Part II: Continuous time analysis
标题：具有机器学习型噪声的随机梯度下降。第二部分：连续时间分析
作者：Stephan Wojtowytsch
链接：https://arxiv.org/abs/2106.02588

【3】 Generative Text Modeling through Short Run Inference
标题：基于短期推理的生成性文本建模
作者：Bo Pang,Erik Nijkamp,Tian Han,Ying Nian Wu
机构：UCLA, Stevens Institute of Technology
备注：10 pages
链接：https://arxiv.org/abs/2106.02513

【4】 Homological Time Series Analysis of Sensor Signals from Power Plants
标题：电厂传感器信号的同源时间序列分析
作者：Luciano Melodia,Richard Lenz
机构：Professorship for Evolutionary Data Management, Friedrich-Alexander University Erlangen-N¨urnberg, Erlangen, Deutschland
链接：https://arxiv.org/abs/2106.02493

【5】 Evaluation of Local Model-Agnostic Explanations Using Ground Truth
标题：使用基本事实对局部模型不可知性解释的评估
作者：Amir Hossein Akhavan Rahnama,Judith Butepage,Pierre Geurts,Henrik Bostrom
机构：KTH Royal Institute of Technology
备注：Submitted on May 28 2021, 13 pages, 4 Figures
链接：https://arxiv.org/abs/2106.02488

【6】 Can convolutional ResNets approximately preserve input distances? A frequency analysis perspective
标题：卷积ResNet可以近似保持输入距离吗？频率分析视角
作者：Lewis Smith,Joost van Amersfoort,Haiwen Huang,Stephen Roberts,Yarin Gal
机构：OATML Group, Department of Computer Science, University of Oxford, Oxford, United Kingdom, Machine Learning Research Group, Department of Engineering Science
备注：Main paper 10 pages including references, appendix 10 pages. 7 figures and 6 tables including appendix
链接：https://arxiv.org/abs/2106.02469

【7】 Ukiyo-e Analysis and Creativity with Attribute and Geometry Annotation
标题：带属性和几何标注的浮世绘分析与创意
作者：Yingtao Tian,Tarin Clanuwat,Chikahiko Suzuki,Asanobu Kitamoto
机构：Google Brain, Tokyo, Japan, ROIS-DS Center for, Open Data in the Humanities, NII
链接：https://arxiv.org/abs/2106.02267

【8】 Analysis of the robustness of NMF algorithms
标题：NMF算法的鲁棒性分析
作者：Alex Díaz,Damian Steele
机构：Dan Dinh Nguyen
备注：16 pages
链接：https://arxiv.org/abs/2106.02213

【9】 The Signed Cumulative Distribution Transform for 1-D Signal Analysis and Classification
标题：一维信号分析与分类的符号累积分布变换
作者：Akram Aldroubi,Rocio Diaz Martin,Ivan Medri,Gustavo K. Rohde,Sumati Thareja
链接：https://arxiv.org/abs/2106.02146

【10】 Finding and Fixing Spurious Patterns with Explanations
标题：使用解释查找和修复虚假模式
作者：Gregory Plumb,Marco Tulio Ribeiro,Ameet Talwalkar
机构：Microsoft Research, CMU, Determined AI
链接：https://arxiv.org/abs/2106.02112

检测相关(4篇)

【1】 Principled change point detection via representation learning
标题：基于表征学习的原则性变化点检测
作者：Evgenia Romanenkova,Alexey Zaytsev,Ramil Zainulin,Matvey Morozov
机构：Skoltech
链接：https://arxiv.org/abs/2106.02602

【2】 Hallucination In Object Detection -- A Study In Visual Part Verification
标题：物体检测中的幻觉--视觉部分验证研究
作者：Osman Semih Kayhan,Bart Vredebregt,Jan C. van Gemert
机构：⋆Computer Vision Lab, Delft University of Technology and §Aiir Innovations
备注：ICIP 2021
链接：https://arxiv.org/abs/2106.02523

【3】 DOCTOR: A Simple Method for Detecting Misclassification Errors
标题：医生：一种检测错误分类错误的简单方法
作者：Federica Granese,Marco Romanelli,Daniele Gorla,Catuscia Palamidessi,Pablo Piantanida
机构：Inria, École Polytechnique, IPP, Sapienza University of Rome, Laboratoire des Signaux et Systèmes (L,S), CentraleSupélec CNRS Université Paris Saclay
链接：https://arxiv.org/abs/2106.02395

【4】 Towards Equal Gender Representation in the Annotations of Toxic Language Detection
标题：论“有毒语言检测”注释中的性别平等表征
作者：Elizabeth Excell,Noura Al Moubayed
机构：Department of Computer Science, Durham University, Durham, UK
备注：Paper is accepted at GeBNLP2021 workshop at ACL-IJCNLP 2021
链接：https://arxiv.org/abs/2106.02183

分类|识别(3篇)

【1】 Musical Prosody-Driven Emotion Classification: Interpreting Vocalists Portrayal of Emotions Through Machine Learning
标题：音乐韵律驱动的情感分类：通过机器学习解读歌唱家对情感的刻画
作者：Farris Nicholas,Model Brian,Savery Richard,Weinberg Gil
机构： Georgia Institute of Technology
链接：https://arxiv.org/abs/2106.02556

【2】 Event Classification with Multi-step Machine Learning
标题：基于多步机器学习的事件分类
作者：Masahiko Saito,Tomoe Kishimoto,Yuya Kaneta,Taichi Itoh,Yoshiaki Umeda,Junichi Tanaka,Yutaro Iiyama,Ryu Sawada,Koji Terashi
机构：International Center for Elementary Particle Physics, The University of Tokyo,-,-, Hongo, Bunkyo, BrainPad Inc.,-,-, Shirokanedai, Minato, Tokyo, Japan, Institute for AI and Beyond, The University of Tokyo,-,-, Hongo, Bunkyo, Tokyo, Japan
链接：https://arxiv.org/abs/2106.02301

【3】 Minimum Word Error Rate Training with Language Model Fusion for End-to-End Speech Recognition
标题：基于语言模型融合的端到端语音识别最小错误率训练
作者：Zhong Meng,Yu Wu,Naoyuki Kanda,Liang Lu,Xie Chen,Guoli Ye,Eric Sun,Jinyu Li,Yifan Gong
机构：Microsoft Corporation, Redmond, WA, USA
备注：5 pages, Interspeech 2021
链接：https://arxiv.org/abs/2106.02302

表征(2篇)

【1】 Light Field Networks: Neural Scene Representations with Single-Evaluation Rendering
标题：光场网络：单次评估渲染的神经场景表示
作者：Vincent Sitzmann,Semon Rezchikov,William T. Freeman,Joshua B. Tenenbaum,Fredo Durand
机构：Frédo Durand, MIT CSAIL, Columbia University, NSF IAFI, MIT BCS, NSF CBMM, vsitzmann.github.iolfns
备注：First two authors contributed equally. Project website: this https URL
链接：https://arxiv.org/abs/2106.02634

【2】 Learning Slice-Aware Representations with Mixture of Attentions
标题：混合注意力学习切片感知表征
作者：Cheng Wang,Sungjin Lee,Sunghyun Park,Han Li,Young-Bum Kim,Ruhi Sarikaya
机构：Amazon Alexa AI
备注：Findings of the ACL: ACL-IJCNLP 2021
链接：https://arxiv.org/abs/2106.02363

优化|敛散性(5篇)

【1】 Optimal Rates of (Locally) Differentially Private Heavy-tailed Multi-Armed Bandits
标题：(局部)差分私用重尾多臂抢劫机的最优费率
作者：Youming Tao,Yulian Wu,Peng Zhao,Di Wang
机构：School of Computer Science, Shandong University, CEMSE, KAUST, Department of Computer Science, Nanjing University
链接：https://arxiv.org/abs/2106.02575

【2】 Debiasing a First-order Heuristic for Approximate Bi-level Optimization
标题：近似双层优化的一阶启发式去偏
作者：Valerii Likhosherstov,Xingyou Song,Krzysztof Choromanski,Jared Davis,Adrian Weller
机构： 20 18;Equal contribution 1University of Cambridge 2Google Re-search, Brain Team 3Columbia University 4Deepmind 5StanfordUniversity 6The Alan Turing Institute
备注：Proceedings of the 38th International Conference on Machine Learning, PMLR 139, 2021. arXiv admin note: text overlap with arXiv:2006.03631
链接：https://arxiv.org/abs/2106.02487

【3】 A Learning-based Optimal Market Bidding Strategy for Price-Maker Energy Storage
标题：基于学习的发电商储能最优市场竞价策略
作者：Mathilde D. Badoual,Scott J. Moura
机构： Controls andApplications Lab (eCAL) at University of California
备注：Presented at the 2021 American Control Conference (ACC), New Orleans, USA, May 25-28, 2021
链接：https://arxiv.org/abs/2106.02396

【4】 Semi-Empirical Objective Functions for MCMC Proposal Optimization
标题：MCMC方案优选的半经验目标函数
作者：Chris Cannella,Vahid Tarokh
机构：Department of ECE, Duke University, Durham, NC
备注：26 pages, 15 tables, 16 figures
链接：https://arxiv.org/abs/2106.02104

【5】 Double Descent Optimization Pattern and Aliasing: Caveats of Noisy Labels
标题：双下降优化模式和混叠：噪声标签的注意事项
作者：Florian Dubost,Khaled Kamal Saab,Erin Hong,Daniel Yang Fu,Max Pike,Siddharth Sharma,Siyi Tang,Nandita Bhaskhar,Christopher Lee-Messer,Daniel Rubin
机构：Stanford University, Khaled K. Saab, Daniel Y. Fu
链接：https://arxiv.org/abs/2106.02100

预测|估计(2篇)

【1】 Influence Estimation and Maximization via Neural Mean-Field Dynamics
标题：基于神经平均场动力学的影响估计与最大化
作者：Shushan He,Hongyuan Zha,Xiaojing Ye
机构：†School of Data Science, Shenzhen Research Institute of Big Data, The Chinese University of Hong Kong
备注：26 pages, 6 figures. arXiv admin note: text overlap with arXiv:2006.09449
链接：https://arxiv.org/abs/2106.02608

【2】 Deep Switching State Space Model (DS$^3$M) for Nonlinear Time Series Forecasting with Regime Switching
标题：带状态切换的非线性时间序列预测的深度切换状态空间模型(DS$^3$M)
作者：Xiuqin Xu,Ying Chen
机构：Integrative Sciences and Engineering Programme, NUS Graduate School, Institute of Data Science, National University of Singapore, Department of Mathematics, National University of Singapore, Risk Management Institute, National University of Singapore, Singapore
链接：https://arxiv.org/abs/2106.02329

其他神经网络|深度学习|模型|建模(18篇)

【1】 MERLOT: Multimodal Neural Script Knowledge Models
标题：Merlot：多模态神经脚本知识模型
作者：Rowan Zellers,Ximing Lu,Jack Hessel,Youngjae Yu,Jae Sung Park,Jize Cao,Ali Farhadi,Yejin Choi
机构：♠Paul G. Allen School of Computer Science & Engineering, University of Washington, ♥Allen Institute for Artificial Intelligence
备注：project page at this https URL
链接：https://arxiv.org/abs/2106.02636

【2】 Consensus Multiplicative Weights Update: Learning to Learn using Projector-based Game Signatures
标题：共识乘性权重更新：学习使用基于投影仪的游戏签名
作者：Nelson Vadori,Rahul Savani,Thomas Spooner,Sumitra Ganesh
机构：J.P. Morgan AI Research, Dept. of Computer Science, University of Liverpool
链接：https://arxiv.org/abs/2106.02615

【3】 Defending Democracy: Using Deep Learning to Identify and Prevent Misinformation
标题：捍卫民主：使用深度学习来识别和防止错误信息
作者：Anusua Trivedi,Alyssa Suhm,Prathamesh Mahankal,Subhiksha Mukuntharaj,Meghana D. Parab,Malvika Mohan,Meredith Berger,Arathi Sethumadhavan,Ashish Jaiman,Rahul Dodhia
机构：the date of receipt and acceptance should be inserted later
链接：https://arxiv.org/abs/2106.02607

【4】 Improve the Interpretability of Attention: A Fast, Accurate, and Interpretable High-Resolution Attention Model
标题：提高注意的可解释性：一种快速、准确、可解释的高分辨率注意模型
作者：Tristan Gomez,Suiyi Ling,Thomas Fréour,Harold Mouchère
机构：Université de Nantes, CHU Nantes, Inserm, CRTI, Inserm UMR , F-, Nantes, France, Harold Mouchére
链接：https://arxiv.org/abs/2106.02566

【5】 Transferable and Distributed User Association Policies for 5G and Beyond Networks
标题：适用于5G及以上网络的可转移和分布式用户关联策略
作者：Mohamed Sana,Nicola di Pietro,Emilio Calvanese Strinati
机构：CEA-Leti, Universit´e Grenoble Alpes, F-, Grenoble, France, Athonet, via Ca del Luogo ,, Bolzano Vicentino (VI), Italy
链接：https://arxiv.org/abs/2106.02540

【6】 Nara: Learning Network-Aware Resource Allocation Algorithms for Cloud Data Centres
标题：NARA：面向云数据中心的学习网络感知资源分配算法
作者：Zacharaya Shabka,Georgios Zervas
机构：Optical Networks Group, University College London
备注：9 content pages, 10 appendix pages, 8 tables, 9 figures
链接：https://arxiv.org/abs/2106.02412

【7】 Strategyproof Learning: Building Trustworthy User-Generated Datasets
标题：防策略学习：构建值得信赖的用户生成的数据集
作者：Sadegh Farhadkhani,Rachid Guerraoui,Lê-Nguyên Hoang
机构：IC School, EPFL, Lausanne, Switzerland
备注：31 pages
链接：https://arxiv.org/abs/2106.02398

【8】 Subdivision-Based Mesh Convolution Networks
标题：基于细分的网格卷积网络
作者：Shi-Min Hu,Zheng-Ning Liu,Meng-Hao Guo,Jun-Xiong Cai,Jiahui Huang,Tai-Jiang Mu,Ralph R. Martin
机构：remesh, convs &, pooling, global, Feature Vector, Classification, Segmentation, Correspondence, Retrieval, dog, dinosaur
备注：Codes are available in this https URL
链接：https://arxiv.org/abs/2106.02285

【9】 Regularization and Reparameterization Avoid Vanishing Gradients in Sigmoid-Type Networks
标题：Sigmoid网络中的正则化和再参数化避免梯度消失
作者：Leni Ven,Johannes Lederer
机构：University of Waterloo, Ruhr-University Bochum
链接：https://arxiv.org/abs/2106.02260

【10】 Deep Contextual Learners for Protein Networks
标题：蛋白质网络中的深度上下文学习器
作者：Michelle M. Li,Marinka Zitnik
机构： 20 20; 1Harvard University 2Broad Institute of MIT and Har-vard 3Harvard Data Science
链接：https://arxiv.org/abs/2106.02246

【11】 Discovery of Causal Additive Models in the Presence of Unobserved Variables
标题：存在不可观测变量的因果加法模型的发现
作者：Takashi Nicholas Maeda,Shohei Shimizu
机构：RIKEN Center for Advanced Intelligence Project, Shiga University
备注：This is an extended version of the UAI 2021 paper entitled "Causal Additive Models with Unobserved Variables"
链接：https://arxiv.org/abs/2106.02234

【12】 Fluctuation-dissipation Type Theorem in Stochastic Linear Learning
标题：随机线性学习中的涨落-耗散型定理
作者：Manhyung Han,Jeonghyeok Park,Taewoong Lee,Jung Hoon Han
机构：School of Electrical Engineering, Seoul National University, Seoul, Korea∗, Cambridge University, Jesus College, Jesus Ln, Cambridge CB,BL, UK†, Harvard College, Harvard University, Cambridge, MA , United States‡
链接：https://arxiv.org/abs/2106.02220

【13】 Enabling Lightweight Fine-tuning for Pre-trained Language Model Compression based on Matrix Product Operators
标题：基于矩阵乘积运算符实现对预训练语言模型压缩的轻量级微调
作者：Peiyu Liu,Ze-Feng Gao,Wayne Xin Zhao,Z. Y. Xie,Zhong-Yi Lu,Ji-Rong Wen
机构：Gaoling School of Artificial Intelligence, Renmin University of China, Department of Physics, Renmin University of China, School of Information, Renmin University of China, Beijing Key Laboratory of Big Data Management and Analysis Methods
备注：Accepted by ACL 2021 main conference
链接：https://arxiv.org/abs/2106.02205

【14】 Privately Learning Mixtures of Axis-Aligned Gaussians
标题：私下学习轴对齐的高斯混合
作者：Ishaq Aden-Ali,Hassan Ashtiani,Christopher Liaw
备注：21 pages
链接：https://arxiv.org/abs/2106.02162

【15】 Beyond Target Networks: Improving Deep $Q$-learning with Functional Regularization
标题：超越目标网络：用功能正规化改善深度$Q$学习
作者：Alexandre Piché,Joseph Marino,Gian Maria Marconi,Christopher Pal,Mohammad Emtiyaz Khan
机构：Mila, Université de Montréal, California Institute of Technology, RIKEN Center for Advanced Intelligence Project, Chris J. Pal, Mila, Polytechnique Montréal, ServiceNow
链接：https://arxiv.org/abs/2106.02613

【16】 Neural Network Surrogate Models for Absorptivity and Emissivity Spectra of Multiple Elements
标题：多元素吸收光谱和发射率光谱的神经网络代理模型
作者：Michael D. Vander Wal,Ryan G. McClarren,Kelli D. Humbird
机构：University of Notre Dame - Department of Aerospace and Mechanical Engineering, University of Notre Dame Notre Dame Indiana USA , Lawrence Livermore National Laboratory , East Ave Livermore California USA
备注：Elsevier Review Format, Double Spaced, 26 pages, 10 figures, 5 tables Michael D. Vander Wal: conceptualization, investigation, writing - original draft, writing - editing and review. Ryan G. McClarren - conceptualization, writing - editing and review. Kelli D. Humbird: conceptualization, writing - editing and review
链接：https://arxiv.org/abs/2106.02528

【17】 PCA Initialization for Approximate Message Passing in Rotationally Invariant Models
标题：旋转不变模型中近似消息传递的PCA初始化
作者：Marco Mondelli,Ramji Venkataramanan
备注：70 pages, 2 figures
链接：https://arxiv.org/abs/2106.02356

【18】 Robust Learning via Persistency of Excitation
标题：基于激励持续性的鲁棒学习
作者：Kaustubh Sridhar,Oleg Sokolsky,Insup Lee,James Weimer
机构：University of Pennsylvania
链接：https://arxiv.org/abs/2106.02078

其他(38篇)

【1】 ViViT: Curvature access through the generalized Gauss-Newton's low-rank structure
标题：Vivit：通过广义高斯-牛顿低阶结构的曲率访问
作者：Felix Dangel,Lukas Tatzel,Philipp Hennig
机构：Tübingen, Germany, University of Tübingen &, MPI for Intelligent Systems
备注：Main text: 11 pages, 3 figures; Supplements: 14 pages, 10 figures, 2 tables
链接：https://arxiv.org/abs/2106.02624

【2】 Be Considerate: Objectives, Side Effects, and Deciding How to Act
标题：体贴：目标、副作用和决定如何行动
作者：Parand Alizadeh Alamdari,Toryn Q. Klassen,Rodrigo Toro Icarte,Sheila A. McIlraith
机构：Department of Computer Science, University of Toronto, Toronto, Canada, Vector Institute, Toronto, Canada, † Schwartz Reisman Institute for Technology and Society, Toronto, Canada
链接：https://arxiv.org/abs/2106.02617

【3】 Sigma-Delta and Distributed Noise-Shaping Quantization Methods for Random Fourier Features
标题：随机傅立叶特征的Sigma-Delta和分布式噪声整形量化方法
作者：Jinjie Zhang,Alexander Cloninger,Rayan Saab
机构：Department of Mathematics, Halıcıo˘glu Data Science Institute, University of California San Diego
链接：https://arxiv.org/abs/2106.02614

【4】 Collection and harmonization of system logs and prototypal Analytics services with the Elastic (ELK) suite at the INFN-CNAF computing centre
标题：利用INFN-CNAF计算中心的Elastic(ELK)套件收集和协调系统日志和原型分析服务
作者：Tommaso Diotalevi,Antonio Falabella,Barbara Martelli,Diego Michelotto,Lucia Morganti,Daniele Bonacorsi,Luca Giommi,Simone Rossi Tisbeni
机构： DiotaleviUniversity of Bologna and INFN, Rossi TisbeniUniversity of Bologna and INFN
备注：Submitted to proceedings of International Symposium on Grids & Clouds 2019 (ISGC2019)
链接：https://arxiv.org/abs/2106.02612

【5】 Inferring Granger Causality from Irregularly Sampled Time Series
标题：从不规则抽样时间序列推断Granger因果关系
作者：Song Wei,Yao Xie,Christopher S. Josef,Rishikesan Kamaleswaran
机构：School of Industrial and Systems Engineering, Georgia Institute of Technology., Department of Surgery, Emory University School of Medicine., Department of Biomedical Informatics, Emory University School of Medicine.
备注：33 pages, 10 figures, 4 tables
链接：https://arxiv.org/abs/2106.02600

【6】 AI Driven Road Maintenance Inspection
标题：人工智能驱动的道路维修检查
作者：Ratnajit Mukherjee,Haris Iqbal,Shabbir Marzban,Ahmed Badar,Terence Brouns,Shruthi Gowda,Elahe Arani,Bahram Zonooz
机构：Advanced Research Lab, Navinfo Europe, Eindhoven, Netherlands
备注：accepted at 27th ITS World Congress, 2021
链接：https://arxiv.org/abs/2106.02567

【7】 Do Syntactic Probes Probe Syntax? Experiments with Jabberwocky Probing
标题：语法探测器探测语法吗？Jabberwocky探测的实验
作者：Rowan Hall Maudslay,Ryan Cotterell
机构：University of Cambridge, ETH Zürich
链接：https://arxiv.org/abs/2106.02559

【8】 Fair Exploration via Axiomatic Bargaining
标题：公理讨价还价的公平探索
作者：Jackie Baek,Vivek F. Farias
机构：Operations Research Center, Massachusetts Institute of Technology, Sloan School of Management
链接：https://arxiv.org/abs/2106.02553

【9】 Active Covering
标题：主动覆盖
作者：Heinrich Jiang,Afshin Rostamizadeh
备注：ICML 2021
链接：https://arxiv.org/abs/2106.02552

【10】 Detect the Interactions that Matter in Matter: Geometric Attention for Many-Body Systems
标题：探测物质中重要的相互作用：多体系统的几何注意
作者：Thorben Frank,Stefan Chmiela
机构：Machine Learning Group, Technische Universität Berlin, Berlin, Germany
链接：https://arxiv.org/abs/2106.02549

【11】 Contracting Neural-Newton Solver
标题：压缩神经牛顿解算器
作者：Samuel Chevalier,Jochen Stiasny,Spyros Chatzivasileiadis
机构：Technical University of Denmark, Kgs. Lyngby, Denmark
链接：https://arxiv.org/abs/2106.02543

【12】 Distributional Sliced Embedding Discrepancy for Incomparable Distributions
标题：不可比分布的分布分片嵌入差异
作者：Mokhtar Z. Alaya,Gilles Gasso,Maxime Berar,Alain Rakotomamonjy
机构：LMAC EA , Université de Technologie de Compiègne, INSA-Rouen, Université de Rouen Normandie, Maxime Bérar, Criteo AI Lab, & LITIS EA , Université de Rouen Normandie
链接：https://arxiv.org/abs/2106.02542

【13】 CAFLOW: Conditional Autoregressive Flows
标题：CAFLOW：条件自回归流
作者：Georgios Batzolis,Marcello Carioni,Christian Etmann,Soroosh Afyouni,Zoe Kourtzi,Carola Bibiane Schönlieb
机构：DAMTP, University of Cambridge, Cambridge CB,WA, Department of Psychology, Cambridge CB,EB, Carola-Bibiane Schönlieb
链接：https://arxiv.org/abs/2106.02531

【14】 CLIP: A Dataset for Extracting Action Items for Physicians from Hospital Discharge Notes
标题：CLIP：用于从医院放电笔记中提取医生操作项的数据集
作者：James Mullenbach,Yada Pruksachatkun,Sean Adler,Jennifer Seale,Jordan Swartz,T. Greg McKelvey,Hui Dai,Yi Yang,David Sontag
机构： ASAPP, Amazon Alexa AI, CUNY Graduate Center, United States Digital Service, MIT
备注：ACL 2021
链接：https://arxiv.org/abs/2106.02524

【15】 Neural Architecture Search via Bregman Iterations
标题：基于Bregman迭代的神经结构搜索
作者：Leon Bungert,Tim Roith,Daniel Tenbrinck,Martin Burger
机构：Department Mathematics, Friedrich-Alexander University Erlangen-Nürnberg, Germany
链接：https://arxiv.org/abs/2106.02479

【16】 Stochastic Multi-Armed Bandits with Unrestricted Delay Distributions
标题：无限制时滞分布的随机多臂带
作者：Tal Lancewicki,Shahar Segal,Tomer Koren,Yishay Mansour
备注：33 pages, 5 figures, ICML 2021
链接：https://arxiv.org/abs/2106.02436

【17】 An Intelligent Resource Reservation for Crowdsourced Live Video Streaming Applications in Geo-Distributed Cloud Environment
标题：地理分布云环境下众包视频直播应用的智能资源预留
作者：Emna Baccour,Fatima Haouari,Aiman Erbad,Amr Mohamed,Kashif Bilal,Mohsen Guizani,Mounir Hamdi
机构： Hamad Bin Khalifa University, Qatar University, ⋆Comsats Institute of Information Technology
备注：Published in IEEE systems journal
链接：https://arxiv.org/abs/2106.02420

【18】 Multitask Online Mirror Descent
标题：多任务在线镜像下降
作者：Nicolò Cesa-Bianchi,Pierre Laforgue,Andrea Paudice,Massimiliano Pontil
机构： Università degli Studi di Milano, Milan, Italy, Istituto Italiano di Tecnologia, Genova, Italy, University College London, London, United Kingdom
链接：https://arxiv.org/abs/2106.02393

【19】 How Good Is NLP? A Sober Look at NLP Tasks through the Lens of Social Impact
标题：NLP有多好？透过社会影响的镜头冷静看待NLP任务
作者：Zhijing Jin,Geeticka Chauhan,Brian Tse,Mrinmaya Sachan,Rada Mihalcea
机构：Max Planck Institute for Intelligent Systems, T¨ubingen, Germany, MIT, Oxford, ETH Z¨urich, University of Michigan
备注：ACL 2021 Findings
链接：https://arxiv.org/abs/2106.02359

【20】 Adiabatic Quantum Feature Selection for Sparse Linear Regression
标题：稀疏线性回归的绝热量子特征选择
作者：Surya Sai Teja Desu,P. K. Srijith,M. V. Panduranga Rao,Naveen Sivadasan
机构：Indian Institute of Technology Hyderabad
备注：8 pages, 2 tables
链接：https://arxiv.org/abs/2106.02357

【21】 SAND-mask: An Enhanced Gradient Masking Strategy for the Discovery of Invariances in Domain Generalization
标题：沙子掩模：一种发现区域综合不变性的增强梯度掩蔽策略
作者：Soroosh Shahtalebi,Jean-Christophe Gagnon-Audet,Touraj Laleh,Mojtaba Faramarzi,Kartik Ahuja,Irina Rish
机构：Mila - Quebec AI Institute, Canada, Université de Montréal, Département d’Informatique et Recherche Opérationelle, Montreal, Canada
链接：https://arxiv.org/abs/2106.02266

【22】 Tractable Regularization of Probabilistic Circuits
标题：概率电路的易处理正则化
作者：Anji Liu,Guy Van den Broeck
机构：Computer Science Department, University of California, Los Angeles, USA
链接：https://arxiv.org/abs/2106.02264

【23】 Visual Question Rewriting for Increasing Response Rate
标题：提高答复率的可视化问题重写
作者：Jiayi Wei,Xilian Li,Yi Zhang,Xin Wang
机构：University of California, Santa Cruz, USA, Xin Eric Wang
链接：https://arxiv.org/abs/2106.02257

【24】 Disentangling Dense Multi-Cable Knots
标题：解开密集的多缆结
作者：Vainavi Viswanath,Jennifer Grannen,Priya Sundaresan,Brijen Thananjeyan,Ashwin Balakrishna,Ellen Novoseller,Jeffrey Ichnowski,Michael Laskey,Joseph E. Gonzalez,Ken Goldberg
机构： 1AUTOLAB at the University of California, Berkeley 2Toyota Research Instituteequal contributionFig
备注：First three authors contributed equally
链接：https://arxiv.org/abs/2106.02252

【25】 A Closer Look at the Worst-case Behavior of Multi-armed Bandit Algorithms
标题：仔细研究多臂Bandit算法的最坏情况行为
作者：Anand Kalvit,Assaf Zeevi
机构：Columbia University
链接：https://arxiv.org/abs/2106.02126

【26】 Neural Collapse Under MSE Loss: Proximity to and Dynamics on the Central Path
标题：MSE缺失下的神经崩溃：中枢通路的近似性和动力学
作者：X. Y. Han,Vardan Papyan,David L. Donoho
机构：University of Toronto, Stanford University
备注：Appendix contains [Section A] empirical experiments, [Sections B-D] discussions and proofs of theoretical results, and [Section E] survey of related works examining Neural Collapse
链接：https://arxiv.org/abs/2106.02073

【27】 The Earth Mover's Pinball Loss: Quantiles for Histogram-Valued Regression
标题：推土机的弹球损失：直方图值回归的分位数
作者：Florian List
机构： and 1The University of Sydney, Sydney Institute for Astronomy, School of Physics
备注：ICML 2021. The code is available at this https URL
链接：https://arxiv.org/abs/2106.02051

【28】 Heterogeneous Noisy Short Signal Camouflage in Multi-Domain Environment Decision-Making
标题：多域环境决策中的异质含噪短信号伪装
作者：Piyush K. Sharma
机构：Army Research Laboratory, Adelphi, MD , USA
备注：Published at: http://www.ibai-publishing.org/journal/issue_massdata/2020_september/massdata_11_1_3_26.php. arXiv admin note: substantial text overlap with arXiv:2106.01497
链接：https://arxiv.org/abs/2106.02044

【29】 Fundamental tradeoffs between memorization and robustness in random features and neural tangent regimes
标题：随机特征和神经切线区域中记忆和鲁棒性之间的基本权衡
作者：Elvis Dohmatob
机构：Criteo AI Lab, (Now at Facebook)
链接：https://arxiv.org/abs/2106.02630

【30】 Extreme sparsity gives rise to functional specialization
标题：极端稀疏导致功能专门化
作者：Gabriel Béna,Dan F. M. Goodman
机构：Department of Electrical and Electronic Engineering, Imperial College London, London, UK
备注：12 pages, 4 figures, Preprint (submitted to Neurips 2021)
链接：https://arxiv.org/abs/2106.02626

【31】 Quantum Perceptron Revisited: Computational-Statistical Tradeoffs
标题：量子感知器的再认识：计算与统计的权衡
作者：Mathieu Roget,Giuseppe Di Molfetta,Hachem Kadri
机构：Aix-Marseille Université, Université de Toulon, CNRS, LIS, Marseille & École Normale Superieure de Lyon, France, CNRS, LIS, CaNa group, Marseille, France, CNRS, LIS, QARMA group, Marseille, France
备注：14 pages, 4 figures
链接：https://arxiv.org/abs/2106.02496

【32】 Teaching keyword spotters to spot new keywords with limited examples
标题：教关键词鉴赏员用有限的例子发现新的关键词
作者：Abhijeet Awasthi,Kevin Kilgour,Hassan Rom
机构：Google Research, Switzerland, Indian Institute of Technology Bombay, India
备注：In INTERSPEECH 2021
链接：https://arxiv.org/abs/2106.02443

【33】 COLD: Concurrent Loads Disaggregator for Non-Intrusive Load Monitoring
标题：COLD：用于非侵入式负载监控的并发负载解聚器
作者：Ilia Kamyshev,Dmitrii Kriukov,Elena Gryazina
机构：Skolkovo Institute of Science and Technology
链接：https://arxiv.org/abs/2106.02352

【34】 Provably Strict Generalisation Benefit for Invariance in Kernel Methods
标题：核方法中不变性的可证明严格泛化效益
作者：Bryn Elesedy
机构：University of Oxford
链接：https://arxiv.org/abs/2106.02346

【35】 Out-of-Distribution Generalization in Kernel Regression
标题：核回归中的非分布泛化
作者：Abdulkadir Canatar,Blake Bordelon,Cengiz Pehlevan
机构：Department of Physics, Harvard University, Cambridge, MA , Center for Brain Science, Harvard University, Cambridge, MA , John A. Paulson School of Engineering and Applied Sciences, Harvard University, Cambridge
链接：https://arxiv.org/abs/2106.02261

【36】 Influence of cognitive, geographical, and collaborative proximity on knowledge production of Canadian nanotechnology
标题：认知、地理和协作邻近性对加拿大纳米技术知识生产的影响
作者：Elva Luz Crespo Neira,Ashkan Ebadi,Catherine Beaudry,Andrea Schiffauerova
机构： Concordia Institute for Information Systems Engineering (CIISE), Concordia University, Ste-Catherine Street West, Montréal, Québec H,G ,W, Canada, Department of Mathematical and Industrial Engineering, École Polytechnique de Montréal
备注：21 pages, 4 figures
链接：https://arxiv.org/abs/2106.02110

【37】 Shape-Preserving Dimensionality Reduction : An Algorithm and Measures of Topological Equivalence
标题：保形降维：一种拓扑等价的算法和度量
作者：Byeongsu Yu,Kisung You
机构：Dept. of Mathematics, Texas A&M University, Dept. of ACMS, University of Notre Dame
备注：18 pages, 2 figures
链接：https://arxiv.org/abs/2106.02096

【38】 Solving Schrödinger Bridges via Maximum Likelihood
标题：用最大似然法求解薛定谔桥
作者：Francisco Vargas,Pierre Thodoroff,Neil D. Lawrence,Austen Lamacraft
机构：Department of Computer Science, Cambridge University, Department of Physics
备注：9 pages + appendix (total 28 pages)
链接：https://arxiv.org/abs/2106.02081

机器翻译，仅供参考

访问arxivdaily.com获取含摘要速递，更有收藏、搜索等功能，涵盖CS|物理|数学|经济|统计|金融|生物|电气领域
同步公众号:arXiv每日学术速递，欢迎关注