机器学习每日论文速递[04.08]

访问arxivdaily.com，获取带摘要的学术速递，更多学科、收藏、评论、搜索……应有尽有。同步公众号(arXiv每日学术速递)，欢迎关注

cs.LG 方向，今日共计94篇

【1】 PlasticineLab: A Soft-Body Manipulation Benchmark with Differentiable Physics
标题：Plastic ticine Lab：一种可微分物理的柔体操纵基准
作者：Zhiao Huang,Yuanming Hu,Tao Du,Siyuan Zhou,Hao Su,Joshua B. Tenenbaum,Chuang Gan
机构*：UC San Diego, Peking University, MIT BCS, CBMM, CSAIL, MIT-IBM Watson Al Lab
备注：Accepted to ICLR 2021 as a spotlight presentation. Project page: this http URL
链接：https://arxiv.org/abs/2104.03311

【2】 Regularizing Generative Adversarial Networks under Limited Data
标题：有限数据下生成性对抗性网络的正则化
作者：Hung-Yu Tseng,Lu Jiang,Ce Liu,Ming-Hsuan Yang,Weilong Yang
备注：CVPR 2021. Project Page: this https URL Code: this https URL
链接：https://arxiv.org/abs/2104.03310

【3】 Streaming Self-Training via Domain-Agnostic Unlabeled Images
标题：通过与领域无关的未标记图像进行流媒体自我训练
作者：Zhiqiu Lin,Deva Ramanan,Aayush Bansal
机构*：Carnegie Mellon University
备注：Project Page: this https URL
链接：https://arxiv.org/abs/2104.03309

【4】 Learned transform compression with optimized entropy encoding
标题：采用优化熵编码的学习型变换压缩
作者：Magda Gregorová,Marc Desaules,Alexandros Kalousis
备注：Neural Compression Workshop @ ICLR 2021
链接：https://arxiv.org/abs/2104.03305

【5】 Modern Hopfield Networks for Few- and Zero-Shot Reaction Prediction
标题：用于少射和零射反应预测的现代Hopfield网络
作者：Philipp Seidl,Philipp Renz,Natalia Dyubankova,Paulo Neves,Jonas Verhoeven,Jörg K. Wegner,Sepp Hochreiter,Günter Klambauer
备注：13 pages + 6 pages appendix
链接：https://arxiv.org/abs/2104.03279

【6】 Semi-Supervised Classification of Social Media Posts: Identifying Sex-Industry Posts to Enable Better Support for Those Experiencing Sex-Trafficking
标题：社交媒体帖子的半监督分类：识别性行业帖子，以便更好地支持性交易经验者
作者：Ellie Simonson
备注：111 pages, Thesis Supervisor: Richard Fletcher
链接：https://arxiv.org/abs/2104.03233

【7】 Evaluation of Time Series Forecasting Models for Estimation of PM2.5 Levels in Air
标题：估计空气中PM2.5水平的时间序列预测模型的评价
作者：Satvik Garg,Himanshu Jindal
机构*：Jaypee of Information Technology, Solan
备注：8 pages, This paper is accepted and presented in the IEEE 6th I2CT 2021 conference. The final version of this paper will appear in the conference proceedings
链接：https://arxiv.org/abs/2104.03226

【8】 Efficient and Accurate In-Database Machine Learning with SQL Code Generation in Python
标题：用Python生成SQL代码的高效准确的数据库内机器学习
作者：Michael Kaufmann,Gabriel Stechschulte,Anna Huber
机构*：Departement Informatik, Departement Wirtschaft, Hochschule Luzern), Hochschule luzern, Rotkreuz, Switzerland, Luzern, Switzerland
链接：https://arxiv.org/abs/2104.03224

【9】 The Use of Video Captioning for Fostering Physical Activity
标题：视频字幕在体育锻炼中的应用
作者：Soheyla Amirian,Abolfazl Farahani,Hamid R. Arabnia,Khaled Rasheed,Thiab R. Taha
机构*：dept. Computer Science, The University of Georgia, The University of georgia), Athens, GA, USA, 《haledRasheed《
链接：https://arxiv.org/abs/2104.03207

【10】 Analysis of Twitter Users' Lifestyle Choices using Joint Embedding Model
标题：基于联合嵌入模型的推特用户生活方式选择分析
作者：Tunazzina Islam,Dan Goldwasser
机构*：Purdue University, West Lafayette, Indiana , USA
备注：accepted at 15th International AAAI Conference on Web and Social Media (ICWSM-2021), 12 pages
链接：https://arxiv.org/abs/2104.03189

【11】 Deep Learning and Traffic Classification: Lessons learned from a commercial-grade dataset with hundreds of encrypted and zero-day applications
标题：深度学习和流量分类：从包含数百个加密和零日应用程序的商业级数据集中吸取的经验教训
作者：Lixuan Yang,Alessandro Finamore,Feng Jun,Dario Rossi
链接：https://arxiv.org/abs/2104.03182

【12】 Adversarial Robustness Guarantees for Gaussian Processes
标题：高斯过程的对抗性鲁棒性保证
作者：Andrea Patane,Arno Blaas,Luca Laurenti,Luca Cardelli,Stephen Roberts,Marta Kwiatkowska
机构*：University of Oxford, ARNOQROBOTS.OX.AC.UK, LUCA.CARDELLIQCS.OX.AC.UK, SJROBOROBOTS.OX. http://AC.UK
备注：Submitted for publication
链接：https://arxiv.org/abs/2104.03180

【13】 Improving Robustness of Deep Reinforcement Learning Agents: Environment Attacks based on Critic Networks
标题：提高深度强化学习代理的健壮性：基于批判网络的环境攻击
作者：Lucas Schott,Manon Césaire,Hatem Hajri,Sylvain Lamprier
机构*： Institute of Research and Technology SystemX, Palaiseau, France
链接：https://arxiv.org/abs/2104.03154

【14】 TenSEAL: A Library for Encrypted Tensor Operations Using Homomorphic Encryption
标题：TenSEAL：一个使用同态加密的加密张量运算库
作者：Ayoub Benaissa,Bilal Retiat,Bogdan Cebere,Alaa Eddine Belfedhal
备注：ICLR 2021 Workshop on Distributed and Private Machine Learning (DPML 2021)
链接：https://arxiv.org/abs/2104.03152

【15】 A matrix math facility for Power ISA(TM) processors
标题：用于Power ISA(TM)处理器的矩阵数学工具
作者：José E. Moreira,Kit Barton,Steven Battle,Peter Bergner,Ramon Bertran,Puneeth Bhat,Pedro Caldeira,David Edelsohn,Gordon Fossum,Brad Frey,Nemanja Ivanovic,Chip Kerchner,Vincent Lim,Shakti Kapoor,Tulio Machado Filho,Silvia Melitta Mueller,Brett Olsson,Satish Sadasivam,Baptiste Saleil,Bill Schmidt,Rajalakshmi Srinivasaraghavan,Shricharan Srivatsan,Brian Thompto,Andreas Wagner,Nelson Wu
链接：https://arxiv.org/abs/2104.03142

【16】 Partially-Connected Differentiable Architecture Search for Deepfake and Spoofing Detection
标题：部分连通可区分体系结构对深伪和欺骗检测的研究
作者：Wanying Ge,Michele Panariello,Jose Patino,Massimiliano Todisco,Nicholas Evans
备注：Submitted to INTERSPEECH 2021
链接：https://arxiv.org/abs/2104.03123

【17】 Scaling Scaling Laws with Board Games
标题：棋类游戏中的比例律
作者：Andrew L. Jones
链接：https://arxiv.org/abs/2104.03113

【18】 Risk-Conditioned Distributional Soft Actor-Critic for Risk-Sensitive Navigation
标题：风险条件下的分布式软行为者--风险敏感导航的批评者
作者：Jinyoung Choi,Christopher R. Dance,Jung-eun Kim,Seulbin Hwang,Kyung-sik Park
备注：ICRA 2021. For associated videos, see this https URL
链接：https://arxiv.org/abs/2104.03111

【19】 Spectral Analysis of the Neural Tangent Kernel for Deep Residual Networks
标题：深层残差网络的神经切核的谱分析
作者：Yuval Belfer,Amnon Geifman,Meirav Galun,Ronen Basri
链接：https://arxiv.org/abs/2104.03093

【20】 HumAID: Human-Annotated Disaster Incidents Data from Twitter
标题：Humaid：来自Twitter的人类注释灾难事件数据
作者：Firoj Alam,Umair Qazi,Muhammad Imran,Ferda Ofli
机构*：Qatar Computing Research Institute, HBKU, Qatar
备注：Accepted in ICWSM-2021, Twitter datasets, Textual content, Natural disasters, Crisis Informatics
链接：https://arxiv.org/abs/2104.03090

【21】 Trajectory Prediction with Latent Belief Energy-Based Model
标题：基于潜在信念能量模型的弹道预测
作者：Bo Pang,Tianyang Zhao,Xu Xie,Ying Nian Wu
机构*：University of California, Los Angeles(UCLA)
备注：13 pages
链接：https://arxiv.org/abs/2104.03086

【22】 Distributional Robustness Loss for Long-tail Learning
标题：长尾学习的分布稳健性损失
作者：Dvir Samuel,Gal Chechik
机构*：Bar-Ilan University, Israel, NVIDIA Research, Israel
链接：https://arxiv.org/abs/2104.03066

【23】 Differentiable Patch Selection for Image Recognition
标题：图像识别中的可微分块选择
作者：Jean-Baptiste Cordonnier,Aravindh Mahendran,Alexey Dosovitskiy,Dirk Weissenborn,Jakob Uszkoreit,Thomas Unterthiner
机构*：EPFL, Switzerland ,Google Research, Brain Team
备注：Accepted to IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2021. Code available at this https URL
链接：https://arxiv.org/abs/2104.03059

【24】 Optimizing Memory Efficiency of Graph NeuralNetworks on Edge Computing Platforms
标题：边缘计算平台上图神经网络存储效率的优化
作者：Ao Zhou,Jianlei Yang,Yeqi Gao,Tong Qiao,Yingjie Qi,Xiaoyi Wang,Yunli Chen,Pengcheng Dai,Weisheng Zhao,Chunming Hu
备注：This paper has been accepted by RTAS 2021(brief industry track), with link to publicly available code
链接：https://arxiv.org/abs/2104.03058

【25】 Few-Shot Incremental Learning with Continually Evolved Classifiers
标题：基于连续进化分类器的少概率增量学习
作者：Chi Zhang,Nan Song,Guosheng Lin,Yun Zheng,Pan Pan,Yinghui Xu
机构*： Nanyang Technological University, Singapore , Alibaba DAMO Academy
备注：Accpeted to CVPR2021
链接：https://arxiv.org/abs/2104.03047

【26】 Multimodal Continuous Visual Attention Mechanisms
标题：多通道连续视觉注意机制
作者：António Farinhas,André F. T. Martins,Pedro M. Q. Aguiar
链接：https://arxiv.org/abs/2104.03046

【27】 On-device Federated Learning with Flower
标题：与FLOWER一起进行设备上的联合学习
作者：Akhil Mathur,Daniel J. Beutel,Pedro Porto Buarque de Gusmão,Javier Fernandez-Marques,Taner Topal,Xinchi Qiu,Titouan Parcollet,Yan Gao,Nicholas D. Lane
备注：Accepted at the 2nd On-device Intelligence Workshop @ MLSys 2021. arXiv admin note: substantial text overlap with arXiv:2007.14390
链接：https://arxiv.org/abs/2104.03042

【28】 Graph-based Normalizing Flow for Human Motion Generation and Reconstruction
标题：基于图形的人体运动生成与重建的归一化流程
作者：Wenjie Yin,Hang Yin,Danica Kragic,Mårten Björkman
备注：8 pages
链接：https://arxiv.org/abs/2104.03020

【29】 Representative & Fair Synthetic Data
标题：代表性和公允的综合数据
作者：Paul Tiwald,Alexandra Ebert,Daniel T. Soukup
机构*：MOSTLY AI, Vienna, Austria, Toronto,ON, Canada
链接：https://arxiv.org/abs/2104.03007

【30】 Universal Adversarial Training with Class-Wise Perturbations
标题：具有阶段性扰动的通用对抗性训练
作者：Philipp Benz,Chaoning Zhang,Adil Karjauv,In So Kweon
机构*：Korea Advanced Institute of Science and Technology (KAIST)
备注：Accepted to ICME 2021
链接：https://arxiv.org/abs/2104.03000

【31】 Theoretically Improving Graph Neural Networks via Anonymous Walk Graph Kernels
标题：基于匿名步行图核的图神经网络的理论改进
作者：Qingqing Long,Yilun Jin,Yi Wu,Guojie Song
机构*：Key Laboratory of Machine Perception (Ministry of The Hong Kong University of Science and Technology, Education), Peking University, China, Hong Kong SAR, China, Peking University,China, Key Laboratory of Machine Perception(Ministry of
备注：11 pages
链接：https://arxiv.org/abs/2104.02995

【32】 Plinius: Secure and Persistent Machine Learning Model Training
标题：普利纽斯：安全持久的机器学习模型训练
作者：Peterson Yuhala,Pascal Felber,Valerio Schiavoni,Alain Tchana
链接：https://arxiv.org/abs/2104.02987

【33】 Reinforcement Learning with a Disentangled Universal Value Function for Item Recommendation
标题：基于解缠普适价值函数的强化学习项目推荐
作者：Kai Wang,Zhene Zou,Qilin Deng,Runze Wu,Jianrong Tao,Changjie Fan,Liang Chen,Peng Cui
机构*：IFuxi AI Lab, NetEase Inc., Hangzhou, China, School of Data and Computer Science, Sun Yat-sen University, China, Tsinghua University, China
备注：9 pages, 4 figures, to be published in Proceedings of the AAAI Conference on Artificial Intelligence 2021
链接：https://arxiv.org/abs/2104.02981

【34】 Synthetic training data generation for deep learning based quality inspection
标题：基于深度学习的质量检测综合训练数据生成
作者：Pierre Gutierrez,Maria Luschkova,Antoine Cordier,Mustafa Shukor,Mona Schappert,Tim Dahmen
机构*：Scortex, rue Berbier du Mets, Paris, France, German Research Center for Artificial Intelligence (DFKI), Saarbrucken, Germany
备注：8 pages, 4 figures, to be published in QCAV 2021 conference, proceedings will by published by SPIE
链接：https://arxiv.org/abs/2104.02980

【35】 The art of defense: letting networks fool the attacker
标题：防御艺术：让网络愚弄攻击者
作者：Jinlai Zhang,Binbin Liu,Lyvjie Chen,Bo Ouyang,Jihong Zhu,Minchi Kuang,Houqing Wang,Yanmei Meng
机构*： Guangxi University ,Tsinghua University
链接：https://arxiv.org/abs/2104.02963

【36】 DyGCN: Dynamic Graph Embedding with Graph Convolutional Network
标题：DyGCN：基于图卷积网络的动态图嵌入
作者：Zeyu Cui,Zekun Li,Shu Wu,Xiaoyu Zhang,Qiang Liu,Liang Wang,Mengmeng Ai
备注：21 pages, 5 figures, submitted to TOIS
链接：https://arxiv.org/abs/2104.02962

【37】 The Emergence of Abstract and Episodic Neurons in Episodic Meta-RL
标题：抽象性和幕幕性神经元在发作性Meta-RL中的出现
作者：Badr AlKhamissi,Muhammad ElNokrashy,Michael Spranger
机构*：Muhammad EINokrashy, Sony CSL, Microsoft EGDC, Tokyo, Japan, Cairo, Egypt, muelnokr [at] microsoft. com, michael. spranger [at] sony. com
备注：This work was accepted at the Learning to Learn Workshop (ICLR 2021)
链接：https://arxiv.org/abs/2104.02959

【38】 Level-Set Curvature Neural Networks: A Hybrid Approach
标题：水平集曲率神经网络：一种混合方法
作者：Luis Ángel Larios-Cárdenas,Frederic Gibou
备注：Submitted to Journal of Computational Physics
链接：https://arxiv.org/abs/2104.02951

【39】 Distantly Supervised Transformers For E-Commerce Product QA
标题：用于电子商务产品QA的远程监控Transformer
作者：Happy Mittal,Aniket Chakrabarti,Belhassen Bayar,Animesh Anant Sharma,Nikhil Rasiwasia
机构*：India Machine Learning, Digital Video, Amazon, Community Shopping, rasiwasicamazon. com
备注：NAACL 2021 (10 pages)
链接：https://arxiv.org/abs/2104.02947

【40】 Deep Interpretable Models of Theory of Mind For Human-Agent Teaming
标题：人-智能体组合心理理论的深层可解释模型
作者：Ini Oguntola,Dana Hughes,Katia Sycara
链接：https://arxiv.org/abs/2104.02938

【41】 TSception: Capturing Temporal Dynamics and Spatial Asymmetry from EEG for Emotion Recognition
标题：TScept：从脑电信号中捕捉时间动力学和空间不对称性用于情感识别
作者：Yi Ding,Neethu Robinson,Qiuhao Zeng,Cuntai Guan
链接：https://arxiv.org/abs/2104.02935

【42】 A Question-answering Based Framework for Relation Extraction Validation
标题：一种基于问答的关系抽取验证框架
作者：Jiayang Cheng,Haiyun Jiang,Deqing Yang,Yanghua Xiao
机构*：Equal contribution to this work.
链接：https://arxiv.org/abs/2104.02934

【43】 Bootstrapping Your Own Positive Sample: Contrastive Learning With Electronic Health Record Data
标题：引导你自己的阳性样本：与电子健康记录数据的对比学习
作者：Tingyi Wanyan,Jing Zhang,Ying Ding,Ariful Azad,Zhangyang Wang,Benjamin S Glicksberg
机构*：TIWANYANQIU.EDU, Indiana University, Bloomington, IN,USA, ZHANG-JINGQRUC.EDU.CN, Renmin University of china, University of Texas at Austin, Austin, Texas, USA, AZADQIU.EDU, Bloomington, IN, USA, ATLASWANG QUTEXAS. EDU, Benjamin Glicksberg
链接：https://arxiv.org/abs/2104.02932

【44】 Pretrained equivariant features improve unsupervised landmark discovery
标题：预先训练的等变特征改进了无监督地标发现
作者：Rahul Rahaman,Atin Ghosh,Alexandre H. Thiery
机构*：National University of Singapore
链接：https://arxiv.org/abs/2104.02925

【45】 Sparse Oblique Decision Trees: A Tool to Understand and Manipulate Neural Net Features
标题：稀疏斜决策树：一种理解和处理神经网络特征的工具
作者：Suryabhan Singh Hada,Miguel Á. Carreira-Perpiñán,Arman Zharmagambetov
机构*：Dept. of Computer Science Engineering, University of California, Merced
链接：https://arxiv.org/abs/2104.02922

【46】 Recognizing and Verifying Mathematical Equations using Multiplicative Differential Neural Units
标题：利用乘性微分神经元识别和验证数学方程
作者：Ankur Mali,Alexander Ororbia,Daniel Kifer,C. Lee Giles
机构*：t The Pennsylvania State University, University Park, PA , Rochester Institute of Technology, Rochester, NY
链接：https://arxiv.org/abs/2104.02899

【47】 On the Critical Role of Conventions in Adaptive Human-AI Collaboration
标题：论惯例在适应性人类-人工智能协作中的关键作用
作者：Andy Shih,Arjun Sawhney,Jovana Kondic,Stefano Ermon,Dorsa Sadigh
机构*：Stanford University, Princeton University
备注：9th International Conference on Learning Representations (ICLR 2021)
链接：https://arxiv.org/abs/2104.02871

【48】 Quasi-Newton Quasi-Monte Carlo for variational Bayes
标题：变分贝叶斯的拟牛顿拟蒙特卡罗
作者：Sifan Liu,Art B. Owen
机构*：Stanford University
链接：https://arxiv.org/abs/2104.02865

【49】 The Value of Planning for Infinite-Horizon Model Predictive Control
标题：无穷远景模型预测控制的规划价值
作者：Nathan Hatch,Byron Boots
备注：7 pages, 8 figures. To appear in the proceedings of the International Conference on Robotics and Automation (ICRA) 2021
链接：https://arxiv.org/abs/2104.02863

【50】 Finite-Sample Analysis for Two Time-scale Non-linear TDC with General Smooth Function Approximation
标题：两时间尺度非线性TDC的广义光滑函数逼近有限样本分析
作者：Yue Wang,Shaofeng Zou,Yi Zhou
机构*：University at Buffalo, Buffalo, NY, USA, University of Utah, Salt Lake City, Utah, USA
链接：https://arxiv.org/abs/2104.02836

【51】 IndoFashion : Apparel Classification for Indian Ethnic Clothes
标题：IndoFashion：印度少数民族服装分类
作者：Pranjal Singh Rajput,Shivangi Aneja
机构*：Delft University of Technology, Technical University of Munich, SAREE, LEHENGA WOMEN KURTA DUPATTA, GOWN NEHRU JACKET SHERWANI MEN KURTA, LEGGINGS SALWAR, BLOUSE, PALAZZO DHOTI PANTS PETTICOAT WOMEN MOJARI
链接：https://arxiv.org/abs/2104.02830

【52】 Ecole: A Library for Learning Inside MILP Solvers
标题：Ecole：一个学习MILP解算器的图书馆
作者：Antoine Prouvost,Justin Dumouchelle,Maxime Gasse,Didier Chételat,Andrea Lodi
机构*：Canada Excellence Research Chair in Data Science for Decision Making, Ecole, Polytechnique de Montreal, Mila, Quebec Artificial Intelligence Institute
链接：https://arxiv.org/abs/2104.02828

【53】 Low-Regret Active learning
标题：低悔主动学习
作者：Cenk Baykal,Lucas Liebenwein,Dan Feldman,Daniela Rus
链接：https://arxiv.org/abs/2104.02822

【54】 Towards measuring fairness in AI: the Casual Conversations dataset
标题：走向衡量人工智能中的公平性：非正式对话数据集
作者：Caner Hazirbas,Joanna Bitton,Brian Dolhansky,Jacqueline Pan,Albert Gordo,Cristian Canton Ferrer
链接：https://arxiv.org/abs/2104.02821

【55】 C2CL: Contact to Contactless Fingerprint Matching
标题：C2CL：接触式到非接触式指纹匹配
作者：Steven A. Grosz,Joshua J. Engelsma,Anil K. Jain
机构*：Fingerprint Acquisition, Self-capture, Segmentation Enhancement Warp Scale, Contact-based, Contactless, Operator-Capture, Minutiae Matching, ∽, Legacy Database, Cosine, Similarity, Texture Matching
链接：https://arxiv.org/abs/2104.02811

【56】 NeuMIP: Multi-Resolution Neural Materials
标题：NeuMIP：多分辨神经材料
作者：Alexandr Kuznetsov,Krishna Mullia,Zexiang Xu,Miloš Hašan,Ravi Ramamoorthi
链接：https://arxiv.org/abs/2104.02789

【57】 Safe-by-Repair: A Convex Optimization Approach for Repairing Unsafe Two-Level Lattice Neural Network Controllers
标题：修复安全：修复不安全的两级格型神经网络控制器的凸优化方法
作者：Ulices Santa Cruz,James Ferlez,Yasser Shoukry
链接：https://arxiv.org/abs/2104.02788

【58】 Autoencoder-based Representation Learning from Heterogeneous Multivariate Time Series Data of Mechatronic Systems
标题：基于自动编码器的机电系统异构多变量时间序列表示学习
作者：K. -P. Kortmann,M. Fehsenfeld,M. Wielitzka
备注：A later version of this paper in german language was submitted to VDI Mechatronic Tagung 2021 and will be published in the conference proceedings
链接：https://arxiv.org/abs/2104.02784

【59】 Bayesian adversarial multi-node bandit for optimal smart grid protection against cyber attacks
标题：贝叶斯对抗性多节点强盗在智能电网最优网络攻击防护中的应用
作者：Jianyu Xu,Bin Liu,Huadong Mo,Daoyi Dong
备注：None
链接：https://arxiv.org/abs/2104.02774

【60】 The Power of Subsampling in Submodular Maximization
标题：子模最大化中的次采样功率
作者：Christopher Harshaw,Ehsan Kazemi,Moran Feldman,Amin Karbasi
备注：arXiv admin note: text overlap with arXiv:1802.07098
链接：https://arxiv.org/abs/2104.02772

【61】 Efficient transfer learning for NLP with ELECTRA
标题：基于ELECTRA的NLP的高效迁移学习
作者：François Mercier
备注：Submission for ML Reproducibility Challenge 2020
链接：https://arxiv.org/abs/2104.02756

【62】 Machine Learning-Driven Virtual Bidding with Electricity Market Efficiency Analysis
标题：基于机器学习的虚拟竞价与电力市场效率分析
作者：Yinglun Li,Nanpeng Yu,Wei Wang
链接：https://arxiv.org/abs/2104.02754

【63】 Communication-Efficient Agnostic Federated Averaging
标题：通信效率不可知的联合平均
作者：Jae Ro,Mingqing Chen,Rajiv Mathews,Mehryar Mohri,Ananda Theertha Suresh
链接：https://arxiv.org/abs/2104.02748

【64】 Proof of the Theory-to-Practice Gap in Deep Learning via Sampling Complexity bounds for Neural Network Approximation Spaces
标题：用神经网络逼近空间的采样复杂度界证明深度学习中理论与实践的差距
作者：Philipp Grohs,Felix Voigtlaender
链接：https://arxiv.org/abs/2104.02746

【65】 InverseForm: A Loss Function for Structured Boundary-Aware Segmentation
标题：InverseForm：一种结构化边界感知分割的损失函数
作者：Shubhankar Borse,Ying Wang,Yizhe Zhang,Fatih Porikli
机构*：Qualcomm AI Research
备注：Accepted to CVPR 2021 as an oral presentation
链接：https://arxiv.org/abs/2104.02745

【66】 Approximate Robust NMPC using Reinforcement Learning
标题：基于强化学习的近似鲁棒NMPC
作者：Hossein Nejatbakhsh Esfahani,Arash Bahari Kordabad,Sebastien Gros
备注：This paper has been accepted to 2021 European Control Conference (ECC)
链接：https://arxiv.org/abs/2104.02743

【67】 Conditional physics informed neural networks
标题：条件物理知情神经网络
作者：Alexander Kovacs,Lukas Exl,Alexander Kornell,Johann Fischbacher,Markus Hovorka,Markus Gusenbauer,Leoni Breth,Harald Oezelt,Masao Yano,Noritsugu Sakuma,Akihito Kinoshita,Tetsuya Shoji,Akira Kato,Thomas Schrefl
机构*：Christian Doppler Laboratory for Magnet design through physics informed machine learning, Viktor Kaplan-StraBe, E, Wiener Neustadt, austria, Danube University Krems, Viktor Kaplan-StraBe ,E, Wiener, University of Vienna Oskar-Morgenstern-Platz , Vienna, austria
备注：18 pages, 11 figures
链接：https://arxiv.org/abs/2104.02741

【68】 Creativity and Machine Learning: a Survey
标题：创造力与机器学习：综述
作者：Giorgio Franceschelli,Mirco Musolesi
机构*：and Alma Mater Studiorum Universita di Bologna, Italy
备注：25 pages, 3 figures, 2 tables
链接：https://arxiv.org/abs/2104.02726

【69】 Minimax Estimation of Linear Functions of Eigenvectors in the Face of Small Eigen-Gaps
标题：小特征间隙情况下特征向量线性函数的极小极大估计
作者：Gen Li,Changxiao Cai,Yuantao Gu,H. Vincent Poor,Yuxin Chen
链接：https://arxiv.org/abs/2104.03298

【70】 DoubleML -- An Object-Oriented Implementation of Double Machine Learning in Python
标题：DoubleML--Python中双机器学习的面向对象实现
作者：Philipp Bach,Victor Chernozhukov,Malte S. Kurz,Martin Spindler
机构*：PHILIPP. BACHQUNI-HAMBURG.DE, MALTE. SIMON. KURZQUNI-HAMBURG.DE, MARTIN. SPINDLERQUNI-HAMBURG.DE, Hamburg Business School, University of Hamburg, Moorweidenstrase , Hamburg, Ger-, many, Massachussets Institute of, Technology, Memorial Drive, Cambridge, MA , USA
备注：6 pages, 2 figures
链接：https://arxiv.org/abs/2104.03220

【71】 TB-Net: A Tailored, Self-Attention Deep Convolutional Neural Network Design for Detection of Tuberculosis Cases from Chest X-ray Images
标题：TB-NET：一种量身定制的自关注深度卷积神经网络胸片结核病例检测设计
作者：Alexander Wong,James Ren Hou Lee,Hadi Rahmat-Khah,Ali Sabri,Amer Alaref
机构*：Waterloo Artificial Intelligence Institute, Canada, DarwinAI Corp., Canada, Niagara Health, McMaster University, Canada, Northern Ontario School of Medicine, Canada
备注：9 pages
链接：https://arxiv.org/abs/2104.03165

【72】 Prediction with Missing Data
标题：缺失数据情况下的预测
作者：Dimitris Bertsimas,Arthur Delarue,Jean Pauphilet
机构*：Sloan School of Management, Massachusetts Institute of Technology, Cambridge, MA, USA., ORCID:,-,-,-, Operations Research Center, Massachusetts Institute of Technology, Cambridge, MA, USA, ORCD:,-,-, London Business School, London, UK
链接：https://arxiv.org/abs/2104.03158

【73】 Dense Dilated UNet: Deep Learning for 3D Photoacoustic Tomography Image Reconstruction
标题：密集扩张UNET：三维光声层析图像重建的深度学习
作者：Steven Guan,Ko-Tsung Hsu,Matthias Eyassu,Parag V. Chitnis
机构*：array of acoustic sensors surrounding the medium [,]. The goal
链接：https://arxiv.org/abs/2104.03130

【74】 Which Neural Network to Choose for Post-Fault Localization, Dynamic State Estimation and Optimal Measurement Placement in Power Systems?
标题：电力系统故障后定位、动态状态估计和最优量测配置应选择哪种神经网络？
作者：Andrei Afonin,Michael Chertkov
备注：12 pages, 8 figures, 10 tables
链接：https://arxiv.org/abs/2104.03115

【75】 Hollow-tree Super: a directional and scalable approach for feature importance in boosted tree models
标题：空心树SUPER：一种增强型树模型中特征重要性的定向可扩展方法
作者：Stephane Doyen,Hugh Taylor,Peter Nicholas,Lewis Crawford,Isabella Young,Michael Sughrue
机构*：Michael E. Sughrue, Omniscient Neurotechnology, Sydney, Australia, Email:, George Street, Level , Sydney, NSW , Australia
备注：28 pages, 1 table, 7 figures, PDF format - Submitted to PLOSONE pending review
链接：https://arxiv.org/abs/2104.03088

【76】 The Properly Use of Google Trends in Forecasting Models
标题：Google Trends在预测模型中的正确使用
作者：Marcelo C. Medeiros,Henrique F. Pires
机构*：Pontificial Catholic University of Rio de janeiro, Pontificial Catholic University of Rio de Janeiro, First version: Fev , This version: Mar
链接：https://arxiv.org/abs/2104.03065

【77】 Utilizing Self-supervised Representations for MOS Prediction
标题：利用自监督表示法进行MOS预测
作者：Wei-Cheng Tseng,Chien-yu Huang,Wei-Tsung Kao,Yist Y. Lin,Hung-yi Lee
机构*：College of Electrical Engineering and Computer Science, National Taiwan University, Taiwan
备注：Submitted to Interspeech 2021
链接：https://arxiv.org/abs/2104.03017

【78】 The AS-NU System for the M2VoC Challenge
标题：面向M2VoC挑战的AS-NU系统
作者：Cheng-Hung Hu,Yi-Chiao Wu,Wen-Chin Huang,Yu-Huai Peng,Yu-Wen Chen,Pin-Jui Ku,Tomoki Toda,Yu Tsao,Hsin-Min Wang
机构*： Academia Sinica, Taiwan Nagoya University, Japan
链接：https://arxiv.org/abs/2104.03009

【79】 Siamese Neural Network with Joint Bayesian Model Structure for Speaker Verification
标题：具有联合贝叶斯模型结构的暹罗神经网络用于说话人确认
作者：Xugang Lu,Peng Shen,Yu Tsao,Hisashi Kawai
机构*：INational Institute of Information and Communication Technology, Japan, Research Center for Information Technology Innovation, Academic Sinica, Taiwan
备注：arXiv admin note: substantial text overlap with arXiv:2101.03329
链接：https://arxiv.org/abs/2104.03004

【80】 CNN Based Segmentation of Infarcted Regions in Acute Cerebral Stroke Patients From Computed Tomography Perfusion Imaging
标题：基于CNN的急性脑卒中患者CT灌注成像梗死区分割
作者：Luca Tomasetti,Kjersti Engan,Mahdieh Khanmohammadi,Kathinka Dæhli Kurz
机构*：Kathinka Daehli Kurz, Stavanger University Hospital, Stavanger medical imaging, laboratory(SMIL), University of Stavanger, Stavanger, Norway, Engineering and Computer Science, BMDLab, Pre-processing steps, Parametric Maps Calculation, Deep Neural, Network, CBF, CBV
链接：https://arxiv.org/abs/2104.03002

【81】 Optimal Algorithms for Differentially Private Stochastic Monotone Variational Inequalities and Saddle-Point Problems
标题：微分私有随机单调变分不等式和鞍点问题的优化算法
作者：Digvijay Boob,Cristóbal Guzmán
机构*：Cristobal Guzman t
链接：https://arxiv.org/abs/2104.02988

【82】 Minimax Kernel Machine Learning for a Class of Doubly Robust Functionals
标题：一类双鲁棒泛函的极小极大核机器学习
作者：AmirEmad Ghassami,Andrew Ying,Ilya Shpitser,Eric Tchetgen Tchetgen
机构*：Johns Hopkins University, The Wharton School, University of Pennsylvania
链接：https://arxiv.org/abs/2104.02929

【83】 An Iteratively Reweighted Method for Sparse Optimization on Nonconvex $\ell_{p}$ Ball
作者：Hao Wang,Xiangyu Yang,Wei Jiang
机构*：School of Information Science and Technology, ShanghaiTech, University, Shanghai , China
备注：This work has been submitted and may be published. Copyright may be transferred without notice, after which this version may no longer be accessible
链接：https://arxiv.org/abs/2104.02912

【84】 PyNET-CA: Enhanced PyNET with Channel Attention for End-to-End Mobile Image Signal Processing
标题：PyNET-CA：用于端到端移动图像信号处理的具有信道关注的增强型PyNET
作者：Byung-Hoon Kim,Joonyoung Song,Jong Chul Ye,JaeHyun Baek
机构*： Korea Advanced Institute of Science and Technology, Daejeon, South Korea, Amazon Web Services, Seoul, South Korea
备注：ECCV 2020 AIM workshop accepted version
链接：https://arxiv.org/abs/2104.02895

【85】 Online Feature Screening for Data Streams with Concept Drift
标题：具有概念漂移的数据流在线特征筛选
作者：Mingyuan Wang,Adrian Barbu
备注：8 figures, 30 pages
链接：https://arxiv.org/abs/2104.02883

【86】 Adapting Speaker Embeddings for Speaker Diarisation
标题：用于说话人区分的自适应说话人嵌入
作者：Youngki Kwon,Jee-weon Jung,Hee-Soo Heo,You Jin Kim,Bong-Jin Lee,Joon Son Chung
机构*：Naver Corporation, South Korea
备注：5 pages, 2 figures, 3 tables, submitted to Interspeech as a conference paper
链接：https://arxiv.org/abs/2104.02879

【87】 Three-class Overlapped Speech Detection using a Convolutional Recurrent Neural Network
标题：基于卷积递归神经网络的三类重叠语音检测
作者：Jee-weon Jung,Hee-Soo Heo,Youngki Kwon,Joon Son Chung,Bong-Jin Lee
机构*：Naver Corporation, South Korea
备注：5 pages, 2 figures, 4 tables, submitted to Interspeech as a conference paper
链接：https://arxiv.org/abs/2104.02878

【88】 Harmless label noise and informative soft-labels in supervised classification
标题：监督分类中的无害标签噪声和信息软标签
作者：Daniel Ahfock,Geoffrey J. McLachlan
链接：https://arxiv.org/abs/2104.02872

【89】 Information Bottleneck Attribution for Visual Explanations of Diagnosis and Prognosis
标题：诊断和预后可视化解释的信息瓶颈归因
作者：Ugur Demir,Ismail Irmakci,Elif Keles,Ahmet Topcu,Ziyue Xu,Concetto Spampinato,Sachin Jambawalikar,Evrim Turkbey,Baris Turkbey,Ulas Bagci
机构*：Northwestern University, Chicago, IL, ECE, Ege University, Izmir, Turkey, Tokat State Hospital, Tokat, Turkey, NVIDIA, Bethesda, MD, University of Catania, Catania, Italy, Columbia University Medical Center, New York, NY
链接：https://arxiv.org/abs/2104.02869

【90】 Capturing Multi-Resolution Context by Dilated Self-Attention
标题：利用扩张型自我注意捕获多分辨率语境
作者：Niko Moritz,Takaaki Hori,Jonathan Le Roux
机构*：Mitsubishi Electric Research Laboratories (MERL), Cambridge, MA, USA
备注：In Proc. ICASSP 2021
链接：https://arxiv.org/abs/2104.02858

【91】 Learning to Rank Microphones for Distant Speech Recognition
标题：学习对麦克风进行排序以进行远程语音识别
作者：Samuele Cornell,Alessio Brutti,Marco Matassoni,Stefano Squartini
链接：https://arxiv.org/abs/2104.02819

【92】 Sparse Partial Least Squares for Coarse Noisy Graph Alignment
标题：基于稀疏偏最小二乘的粗糙噪声图对齐
作者：Michael Weylandt,George Michailidis,T. Mitchell Roddenberry
机构*：University of Florida Informatics Institute, Rice University, Gainesville, FL , Houston, TX
链接：https://arxiv.org/abs/2104.02810

【93】 Robust Semantic Interpretability: Revisiting Concept Activation Vectors
标题：健壮的语义可解释性：重访概念激活向量
作者：Jacob Pfau,Albert T. Young,Jerome Wei,Maria L. Wei,Michael J. Keiser
机构*：UCSF and Dermatology Service, SFVAMC, Institute for Neurodegenerative Diseases, and Bakar Computational Health Sciences Institute, UCSF, UC Berkeley
备注：ICML WHI 2020
链接：https://arxiv.org/abs/2104.02768

【94】 Exploring Targeted Universal Adversarial Perturbations to End-to-end ASR Models
标题：探索端到端ASR模型的目标通用对抗性扰动
作者：Zhiyun Lu,Wei Han,Yu Zhang,Liangliang Cao
机构*：Google Inc., USA
备注：Submitted to INTERSPEECH 2021
链接：https://arxiv.org/abs/2104.02757

机器翻译，仅供参考；“机构”仅供参考