cs.AI 方向，今日共计44篇

【1】 CoMoGAN: continuous model-guided image-to-image translation
标题：CoMoGAN：连续模型引导的图像到图像转换
作者：Fabio Pizzati,Pietro Cerri,Raoul de Charette
机构：Inria, Vislab
备注：CVPR 2021 oral
链接：https://arxiv.org/abs/2103.06879

【2】 Understanding the origin of information-seeking exploration in  probabilistic objectives for control
标题：理解概率性控制目标中信息寻求式探索的由来
作者：Beren Millidge,Alexander Tschantz,Anil Seth,Christopher Buckley
机构：Sackler Center for Consciousness Science, University of Edinburgh, University of Sussex, Anil K Seth, Evolutionary and Adaptive Systems Research Group, March
备注：11-03-21 initial upload
链接：https://arxiv.org/abs/2103.06859

【3】 A conditional, a fuzzy and a probabilistic interpretation of  self-organising maps
标题：自组织映射的条件、模糊和概率解释
作者：Laura Giordano,Valentina Gliozzi,Daniele Theseider Dupré
机构：DISIT-Universita del Piemonte Orientale, Alessandria, Italy,  Center for Logic, Language and Cognition, Dipartimento di Informatica, Universita di Torino, Italy
备注：28 pages, 1 figure. arXiv admin note: substantial text overlap with arXiv:2008.13278
链接：https://arxiv.org/abs/2103.06854

【4】 Policy Search with Rare Significant Events: Choosing the Right Partner  to Cooperate with
标题：罕见重大事件的政策搜索：选择合适的合作伙伴
作者：Paul Ecoffet,Nicolas Fontbonne,Jean-Baptiste André,Nicolas Bredeche
机构：Institut des Systemes Intelligents et de robotique, Sorbonne Universite, Paris, France, Jean-Baptiste Andre, Institut Jean Nicod, Departement d'Etudes Cognitives, Ecole Normale Superieure
链接：https://arxiv.org/abs/2103.06846

【5】 COVID-19 Smart Chatbot Prototype for Patient Monitoring
标题：用于患者监护的冠状病毒智能聊天机器人原型
作者：Hannah Lei,Weiqi Lu,Alan Ji,Emmett Bertram,Paul Gao,Xiaoqian Jiang,Arko Barman
机构：Rice University, Houston, Texas, United States;, The University of Texas Health Science Center at Houston
备注：This manuscript is under consideration for the AMIA 2021 Annual Symposium
链接：https://arxiv.org/abs/2103.06816

【6】 Adapting User Interfaces with Model-based Reinforcement Learning
标题：基于模型的强化学习自适应用户界面
作者：Kashyap Todi,Gilles Bailly,Luis A. Leiva,Antti Oulasvirta
机构：Aalto University, Sorbonne Universite, CNRS, ISIR, Finland, France, University of Luxembourg, Current Design, Model-Based Reinforcement Learning, Adapted Design, Final Design, Open, Time, observed, New, State values, adapt, Save As..., a,a, estimated, Export, Move To..., States, s,n, Print, Rename, Duplicate, Exit, t..
备注：13 pages, 10 figures, ACM CHI 2021 Full Paper
链接：https://arxiv.org/abs/2103.06807

【7】 Improved Coherence Index-Based Bound in Compressive Sensing
标题：一种改进的基于相干指数的压缩传感边界
作者：Ljubisa Stankovic,Milos Brajovic,Danilo Mandic,Isidora Stankovic,Milos Dakovic
备注：5 pages, 1 figure
链接：https://arxiv.org/abs/2103.06804

【8】 Intelligent behavior depends on the ecological niche: Scaling up AI to  human-like intelligence in socio-cultural environments
标题：智能行为依赖于生态位：在社会文化环境中将人工智能扩大到类似人类的智能
作者：Manfred Eppe,Pierre-Yves Oudeyer
备注：Keywords: developmental AI, general artificial intelligence,  human-like AI, embodiment, cultural evolution, language, socio-cultural  skills
链接：https://arxiv.org/abs/2103.06769

【9】 ENTRUST: Argument Reframing with Language Models and Entailment
标题：委托：用语言模型和蕴涵重构论证
作者：Tuhin Chakrabarty,Christopher Hidey,Smaranda Muresan
机构：Data Science Institute, Columbia University, Google,
备注：NAACL 2021
链接：https://arxiv.org/abs/2103.06758

【10】 Auto-COP: Adaptation Generation in Context-Oriented Programming using  Reinforcement Learning Options
标题：自动COP：使用强化学习选项在面向上下文的编程中生成适应
作者：Nicolás Cardozo,Ivana Dusparic
备注：Submitted to The Art, Science, and Engineering of Programming Journal. 22 pages
链接：https://arxiv.org/abs/2103.06757

【11】 Knowledge Graph Question Answering using Graph-Pattern Isomorphism
标题：基于图形-模式同构的知识图问答
作者：Daniel Vollmers,Rricha Jalota,Diego Moussallem,Hardik Topiwala,Axel-Cyrille Ngonga Ngomo,Ricardo Usbeck
机构：Data Science Group, Paderborn University, Germany, Fraunhofer IAIS, Dresden, Germany
链接：https://arxiv.org/abs/2103.06752

【12】 Intraclass clustering: an implicit learning ability that regularizes  DNNs
标题：类内聚类：规则化DNN的内隐学习能力
作者：Carbonnelle Simon,Christophe De Vleeschouwer
机构：FNRS research fellows, ICTEAM, Universite catholique de louvain, Louvain-La-Neuve, Belgium
备注：Published as a conference paper at ICLR 2021
链接：https://arxiv.org/abs/2103.06733

【13】 Beta-CROWN: Efficient Bound Propagation with Per-neuron Split  Constraints for Complete and Incomplete Neural Network Verification
标题：Beta-Crown：完全和不完全神经网络验证的每神经元分裂约束下的有效有界传播
作者：Shiqi Wang,Huan Zhang,Kaidi Xu,Xue Lin,Suman Jana,Cho-Jui Hsieh,J. Zico Kolter
机构：Columbia University ,UCLA ,Northeastern Univerisity ,CMU,  Equal Contribution.
备注：Shiqi Wang, Huan Zhang and Kaidi Xu contributed equally
链接：https://arxiv.org/abs/2103.06624

【14】 A Quadratic Actor Network for Model-Free Reinforcement Learning
标题：一种无模型强化学习的二次型执行者网络
作者：Matthias Weissenbacher,Yoshinobu Kawahara
备注：8 pages, 15 figures
链接：https://arxiv.org/abs/2103.06617

【15】 Symbolic Reinforcement Learning for Safe RAN Control
标题：用于安全跑道控制的符号强化学习
作者：Alexandros Nikou,Anusha Mujumdar,Marin Orlic,Aneta Vulgarakis Feljan
备注：The paper has been accepted to be presented in 20th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2021), May 3-7, London, UK (demo track)
链接：https://arxiv.org/abs/2103.06602

【16】 Integrated Age Estimation Mechanism
标题：综合年龄估计机制
作者：Fan Li,Yongming Li,Pin Wang,Jie Xiao,Fang Yan,Xinke Li
机构：China;, . Collaborative Innovation Center for Brain Science, Chongqing University, Chongqing , China)
链接：https://arxiv.org/abs/2103.06546

【17】 Causal Learner: A Toolbox for Causal Structure and Markov Blanket  Learning
标题：因果学习器：因果结构和马尔可夫毯子学习工具箱
作者：Zhaolong Ling,Kui Yu,Yiwen Zhang,Lin Liu,Jiuyong Li
机构：Anhui University, Hefei, Anhui, China, YUKUIQHFUT.EDU.CN, Hefei University of Technology, Hefei, Anhui, China, ZHANGYIWENQAHU.EDU.CN, LIN. LIUQUNISA.EDU.AU, JIUYONG. LIQUNISA.EDU.AU, UniSA STEM, University of South Australia, Adelaide, SA, australia, Editor:
链接：https://arxiv.org/abs/2103.06544

【18】 Affect2MM: Affective Analysis of Multimedia Content Using Emotion  Causality
标题：Affect2MM：基于情感因果关系的多媒体内容情感分析
作者：Trisha Mittal,Puneet Mathur,Aniket Bera,Dinesh Manocha
机构：University of Maryland, College Park, USA
备注：Accepted to CVPR 2021
链接：https://arxiv.org/abs/2103.06541

【19】 Memristive Stochastic Computing for Deep Learning Parameter Optimization
标题：记忆随机计算在深度学习参数优化中的应用
作者：Corey Lammie,Jason K. Eshraghian,Wei D. Lu,Mostafa Rahimi Azghadi
备注：Accepted by IEEE Transactions on Circuits and Systems Part II:  Express Briefs
链接：https://arxiv.org/abs/2103.06506

【20】 Adversarial Laser Beam: Effective Physical-World Attack to DNNs in a  Blink
标题：敌意激光束：对DNNs的有效物理世界攻击
作者：Ranjie Duan,Xiaofeng Mao,A. K. Qin,Yun Yang,Yuefeng Chen,Shaokai Ye,Yuan He
机构：Swinburne University of Technology ,Alibaba Group, EPFL
备注：Accepted to CVPR2021
链接：https://arxiv.org/abs/2103.06504

【21】 3D Human Pose, Shape and Texture from Low-Resolution Images and Videos
标题：来自低分辨率图像和视频的三维人体姿势、形状和纹理
作者：Xiangyu Xu,Hao Chen,Francesc Moreno-Noguer,Laszlo A. Jeni,Fernando De la Torre
备注：arXiv admin note: substantial text overlap with arXiv:2007.13666
链接：https://arxiv.org/abs/2103.06498

【22】 Active$^2$ Learning: Actively reducing redundancies in Active Learning  methods for Sequence Tagging and Machine Translation
标题：主动$^2$学习：主动减少序列标注和机器翻译主动学习方法中的冗余
作者：Rishi Hazra,Parag Dutta,Shubham Gupta,Mohammed Abdul Qaathir,Ambedkar Dukkipati
机构：Indian Institute of Science, Bangalore, KA, IN-
备注：Accepted in NAACL-HLT-2021
链接：https://arxiv.org/abs/2103.06490

【23】 DAFAR: Detecting Adversaries by Feedback-Autoencoder Reconstruction
标题：Dafar：利用反馈-自动编码器重构检测攻击者
作者：Haowen Liu,Ping Yi,Hsiao-Ying Lin,Jie Shi
机构：Shanghai Jiao Tong University, China, Huawei International, Singapore
链接：https://arxiv.org/abs/2103.06487

【24】 Metapaths guided Neighbors aggregated Network for?Heterogeneous Graph  Reasoning
标题：基于元路径引导的邻居聚合网络异构图推理
作者：Bang Lin,Xiuchong Wang,Yu Dong,Chengfu Huo,Weijun Ren,Chuanyu Xu
机构：Alibaba Group
链接：https://arxiv.org/abs/2103.06474

【25】 Multi-Task Federated Reinforcement Learning with Adversaries
标题：具有对手的多任务联合强化学习
作者：Aqeel Anwar,Arijit Raychowdhury
机构：Georgia Institute of Technology, Atlanta, GA, USA
备注：14 pages, 19 figures, 4 tables, journal
链接：https://arxiv.org/abs/2103.06473

【26】 Generalizable Episodic Memory for Deep Reinforcement Learning
标题：深度强化学习的泛化情景记忆
作者：Hao Hu,Jianing Ye,Zhizhou Ren,Guangxiang Zhu,Chongjie Zhang
链接：https://arxiv.org/abs/2103.06469

【27】 Emerging Paradigms of Neural Network Pruning
标题：新兴的神经网络修剪范式
作者：Huan Wang,Can Qin,Yulun Zhang,Yun Fu
机构：Northeastern University, Boston, MA, USA
链接：https://arxiv.org/abs/2103.06460

【28】 Towards Multi-Sense Cross-Lingual Alignment of Contextual Embeddings
标题：论语境嵌入的多义跨语言对齐
作者：Linlin Liu,Thien Hai Nguyen,Shafiq Joty,Lidong Bing,Luo Si
机构：Nanyang Technological University, Singapore, DAMO Academy, Alibaba Group, Salesforce Research Asia, Singapore
链接：https://arxiv.org/abs/2103.06459

【29】 Full Page Handwriting Recognition via Image to Sequence Extraction
标题：基于图像到序列提取的整页手写识别
作者：Sumeet S. Singh,Sergey Karayev
机构：Turnitin LLC
链接：https://arxiv.org/abs/2103.06450

【30】 Where is your place, Visual Place Recognition?
标题：你的位置在哪里，视觉位置识别？
作者：Sourav Garg,Tobias Fischer,Michael Milford
机构：QUT Centre for Robotics, Queensland University of Technology
备注：Under Review
链接：https://arxiv.org/abs/2103.06443

【31】 Hierarchical Bayesian Model for the Transfer of Knowledge on Spatial  Concepts based on Multimodal Information
标题：基于多模态信息的空间概念知识传递的分层贝叶斯模型
作者：Yoshinobu Hagiwara,Keishiro Taguchi,Satoshi Ishibushi,Akira Taniguchi,Tadahiro Taniguchi
机构：Taniguchia, a Ritsumeikan University, -,-, Noji Higashi, Kusatsu, Shiga ,-, Japan;, b Panasonic Corporation, Osaka, Japan, (v,., released December ,)
备注：17 pages, 12 figures, 6 tables
链接：https://arxiv.org/abs/2103.06442

【32】 Topical Language Generation using Transformers
标题：使用Transformer生成主题语言
作者：Rohola Zandie,Mohammad H. Mahoor
备注：Accepted in the Journal of Natural Language Engineering
链接：https://arxiv.org/abs/2103.06434

【33】 XDO: A Double Oracle Algorithm for Extensive-Form Games
标题：XDO：一种求解扩展形式博弈的双Oracle算法
作者：Stephen McAleer,John Lanier,Pierre Baldi,Roy Fox
链接：https://arxiv.org/abs/2103.06426

【34】 3D Head-Position Prediction in First-Person View by Considering Head  Pose for Human-Robot Eye Contact
标题：考虑头部姿态的人-机器人眼神接触第一人称视角三维头部位置预测
作者：Yuki Tamaru,Yasunori Ozaki,Yuki Okafuji,Jun Baba,Junya Nakanishi,Yuichiro Yoshikawa
备注：Submit to IROS 2021
链接：https://arxiv.org/abs/2103.06417

【35】 A Vision Based Deep Reinforcement Learning Algorithm for UAV Obstacle  Avoidance
标题：一种基于视觉的无人机避障深度强化学习算法
作者：Jeremy Roghair,Kyungtae Ko,Amir Ehsan Niaraki Asli,Ali Jannesari
机构：Iowa State University, IA, USA
备注：12 pages, 6 figures
链接：https://arxiv.org/abs/2103.06403

【36】 Improving Context-Based Meta-Reinforcement Learning with Self-Supervised  Trajectory Contrastive Learning
标题：用自监督轨迹对比学习改进基于上下文的元强化学习
作者：Bernie Wang,Simon Xu,Kurt Keutzer,Yang Gao,Bichen Wu
链接：https://arxiv.org/abs/2103.06386

【37】 HOT-VAE: Learning High-Order Label Correlation for Multi-Label  Classification via Attention-Based Variational Autoencoders
标题：HOT-VAE：基于注意力的变分自动编码器学习高阶标签相关性用于多标签分类
作者：Wenting Zhao,Shufeng Kong,Junwen Bai,Daniel Fink,Carla Gomes
机构：Cornell University, USA,  Cornell Lab of Ornithology, Ithaca, NY, USA
备注：accepted at AAAI'21 AISI
链接：https://arxiv.org/abs/2103.06375

【38】 Hard Attention Control By Mutual Information Maximization
标题：基于互信息最大化的硬注意力控制
作者：Himanshu Sahni,Charles Isbell
机构：Georgia Institute of Technology, USA
链接：https://arxiv.org/abs/2103.06371

【39】 Causal-aware Safe Policy Improvement for Task-oriented dialogue
标题：面向任务型对话的因果感知安全政策改进
作者：Govardana Sachithanandam Ramachandran,Kazuma Hashimoto,Caiming Xiong
机构：Salesforce Research
链接：https://arxiv.org/abs/2103.06370

【40】 Majority Voting with Bidirectional Pre-translation Improves Bitext  Retrieval
标题：采用双向预翻译的多数投票改进了比特文检索
作者：Alex Jones,Derry Tanti Wijaya
机构：Dartmouth College, Boston University
链接：https://arxiv.org/abs/2103.06369

【41】 User-centered Evaluation of Popularity Bias in Recommender Systems
标题：推荐系统中以用户为中心的人气偏向评价
作者：Himan Abdollahpouri,Masoud Mansoury,Robin Burke,Bamshad Mobasher,Edward Malthouse
机构：Northwestern University, Eindhoven University of Technology University of Colorado Boulder, Evanston, USA, Eindhoven, Netherlands, Boulder,USA, DePaul University, Chicago, USA
备注：Proceedings of the 29th ACM Conference on User Modeling, Adaptation and Personalization (UMAP '21), June 21--25, 2021, Utrecht, Netherlands. arXiv admin note: text overlap with arXiv:2007.12230
链接：https://arxiv.org/abs/2103.06364

【42】 Continual Semantic Segmentation via Repulsion-Attraction of Sparse and  Disentangled Latent Representations
标题：基于排斥-吸引的稀疏和解缠潜在表征的连续语义分割
作者：Umberto Michieli,Pietro Zanuttigh
机构：University of Padova
备注：CVPR 2021. 22 pages, 10 figures, 11 tables
链接：https://arxiv.org/abs/2103.06342

【43】 The AI Index 2021 Annual Report
标题：AI指数2021年年度报告
作者：Daniel Zhang,Saurabh Mishra,Erik Brynjolfsson,John Etchemendy,Deep Ganguli,Barbara Grosz,Terah Lyons,James Manyika,Juan Carlos Niebles,Michael Sellitto,Yoav Shoham,Jack Clark,Raymond Perrault
链接：https://arxiv.org/abs/2103.06312

【44】 What is Multimodality?
标题：什么是多模态？
作者：Letitia Parcalabescu,Nils Trost,Anette Frank
机构：Computational Linguistics, Center for Molecular, Biology (ZMBH), Heidelberg University
备注：9 pages, 3 figures
链接：https://arxiv.org/abs/2103.06304