【重磅】ICLR2020 || 106篇深度强化学习顶会论文汇总

关注：决策智能与机器学习，深耕AI脱水干货

来源 | EndtoEnd.ai

作者 | DeepRL

报道 | 深度强化学习实验室

编辑 |  九三山人

【导读】今年的ICLR大会转到了线上举行，DeepMind和哈佛的研究人员投稿了一篇神经网络控制虚拟小白鼠模的论文十分亮眼。此次ICLR大会，华人学者参与论文数占比近60%，Google入选80余篇表现依旧抢眼，而国内的研究团队也不落下风，满分论文频现。本届ICLR 2020共有2594篇投稿，687 篇被接收。其中：48篇 oral 108篇，spotlights 531篇， poster 录取率为 26.5%，相比去年的 31.4% 略有降低。强化学习一直是ICLR投稿的热点，近年来强化学习及深度强化学习不断刷新着人类在游戏、棋牌等领域的最好成绩，关于谷歌研究人员用6小时完成AI芯片设计，也是采用了深度强化学习方法，强化学习的威力不容小觑。本文共列举了106篇深度强化学习领域的论文。

排名	1
平均得分	8
论文地址	https://openreview.net/forum?id=HJgLZR4KvH
标题	Dynamics-aware Unsupervised Skill Discovery
得分	8 8 8
Variance	0
Decision	Accept (Talk)
排名	1
平均得分	8
论文地址	https://openreview.net/forum?id=H1gax6VtDB
标题	Contrastive Learning Of Structured World Models
得分	8 8 8
Variance	0
Decision	Accept (Talk)
排名	1
平均得分	8
论文地址	https://openreview.net/forum?id=r1etN1rtPB
标题	Implementation Matters In Deep Rl: A Case Study On Ppo And Trpo
得分	8 8 8
Variance	0
Decision	Accept (Talk)
排名	1
平均得分	8
论文地址	https://openreview.net/forum?id=HkxlcnVFwB
标题	Gendice: Generalized Offline Estimation Of Stationary Values
得分	8 8 8
Variance	0
Decision	Accept (Talk)
排名	1
平均得分	8
论文地址	https://openreview.net/forum?id=S1g2skStPB
标题	Causal Discovery With Reinforcement Learning
得分	8 8 8
Variance	0
Decision	Accept (Talk)
排名	2
平均得分	7.33
论文地址	https://openreview.net/forum?id=r1genAVKPB
标题	Is A Good Representation Sufficient For Sample Efficient Reinforcement Learning?
得分	8 8 6
Variance	0.89
Decision	Accept (Spotlight)
排名	2
平均得分	7.33
论文地址	https://openreview.net/forum?id=rklHqRVKvH
标题	Harnessing Structures For Value-based Planning And Reinforcement Learning
得分	6 8 8
Variance	0.89
Decision	Accept (Talk)
排名	2
平均得分	7.33
论文地址	https://openreview.net/forum?id=SJgzLkBKPB
标题	Explain Your Move: Understanding Agent Actions Using Focused Feature Saliency
得分	6 8 8
Variance	0.89
Decision	Accept (Poster)
排名	2
平均得分	7.33
论文地址	https://openreview.net/forum?id=SJeD3CEFPH
标题	Meta-q-learning
得分	8 8 6
Variance	0.89
Decision	Accept (Talk)
排名	2
平均得分	7.33
论文地址	https://openreview.net/forum?id=HJl8_eHYvS
标题	Discriminative Particle Filter Reinforcement Learning For Complex Partial Observations
得分	8 6 8
Variance	0.89
Decision	Accept (Poster)
排名	2
平均得分	7.33
论文地址	https://openreview.net/forum?id=rkgbYyHtwB
标题	Disagreement-regularized Imitation Learning
得分	6 8 8
Variance	0.89
Decision	Accept (Spotlight)
排名	2
平均得分	7.33
论文地址	https://openreview.net/forum?id=S1glGANtDr
标题	Doubly Robust Bias Reduction In Infinite Horizon Off-policy Estimation
得分	6 8 8
Variance	0.89
Decision	Accept (Spotlight)
排名	2
平均得分	7.33
论文地址	https://openreview.net/forum?id=rkgvXlrKwH
标题	Seed Rl: Scalable And Efficient Deep-rl With Accelerated Central Inference
得分	8 6 8
Variance	0.89
Decision	Accept (Talk)
排名	2
平均得分	7.33
论文地址	https://openreview.net/forum?id=rJe2syrtvS
标题	The Ingredients Of Real World Robotic Reinforcement Learning
得分	6 8 8
Variance	0.89
Decision	Accept (Spotlight)
排名	2
平均得分	7.33
论文地址	https://openreview.net/forum?id=BJlQtJSKDB
标题	Watch The Unobserved: A Simple Approach To Parallelizing Monte Carlo Tree Search
得分	8 6 8
Variance	0.89
Decision	Accept (Talk)
排名	2
平均得分	7.33
论文地址	https://openreview.net/forum?id=ryeYpJSKwr
标题	Meta-learning Acquisition Functions For Transfer Learning In Bayesian Optimization
得分	8 6 8
Variance	0.89
Decision	Accept (Spotlight)
排名	2
平均得分	7.33
论文地址	https://openreview.net/forum?id=ryxdEkHtPS
标题	A Closer Look At Deep Policy Gradients
得分	8 6 8
Variance	0.89
Decision	Accept (Talk)
排名	2
平均得分	7.33
论文地址	https://openreview.net/forum?id=BJeAHkrYDS
标题	Fast Task Inference With Variational Intrinsic Successor Features
得分	8 6 8
Variance	0.89
Decision	Accept (Talk)
排名	2
平均得分	7.33
论文地址	https://openreview.net/forum?id=rJgJDAVKvB
标题	Learning To Plan In High Dimensions Via Neural Exploration-exploitation Trees
得分	8 8 6
Variance	0.89
Decision	Accept (Spotlight)
排名	3
平均得分	7
论文地址	https://openreview.net/forum?id=S1lOTC4tDS
标题	Dream To Control: Learning Behaviors By Latent Imagination
得分	8 6 6 8
Variance	1
Decision	Accept (Spotlight)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=SygKyeHKDH
标题	Making Efficient Use Of Demonstrations To Solve Hard Exploration Problems
得分	6 8 6
Variance	0.89
Decision	Accept (Poster)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=SJleNCNtDH
标题	Intrinsic Motivation For Encouraging Synergistic Behavior
得分	6 8 6
Variance	0.89
Decision	Accept (Poster)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=S1xKd24twB
标题	Sqil: Imitation Learning Via Reinforcement Learning With Sparse Rewards
得分	8 6 6
Variance	0.89
Decision	Accept (Poster)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=ryxgJTEYDr
标题	Reinforcement Learning With Competitive Ensembles Of Information-constrained Primitives
得分	8 6 6
Variance	0.89
Decision	Accept (Poster)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=B1gZV1HYvS
标题	Multi-agent Interactions Modeling With Correlated Policies
得分	6 6 8
Variance	0.89
Decision	Accept (Poster)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=BJgy96EYvr
标题	Influence-based Multi-agent Exploration
得分	6 6 8
Variance	0.89
Decision	Accept (Spotlight)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=rylJkpEtwS
标题	Learning The Arrow Of Time For Problems In Reinforcement Learning
得分	6 6 8
Variance	0.89
Decision	Accept (Poster)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=Bkl7bREtDr
标题	Amrl: Aggregated Memory For Reinforcement Learning
得分	6 6 8
Variance	0.89
Decision	Accept (Poster)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=S1xCPJHtDB
标题	Model Based Reinforcement Learning For Atari
得分	6 8 6
Variance	0.89
Decision	Accept (Spotlight)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=r1lL4a4tDB
标题	Variational Recurrent Models For Solving Partially Observable Control Tasks
得分	6 6 8
Variance	0.89
Decision	Accept (Poster)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=HJlxIJBFDr
标题	Sample Efficient Policy Gradient Methods With Recursive Variance Reduction
得分	6 8 6
Variance	0.89
Decision	Accept (Poster)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=H1exf64KwH
标题	Exploring Model-based Planning With Policy Networks
得分	6 8 6
Variance	0.89
Decision	Accept (Poster)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=HygnDhEtvr
标题	Reinforcement Learning Based Graph-to-sequence Model For Natural Question Generation
得分	6 6 8
Variance	0.89
Decision	Accept (Poster)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=rkg-TJBFPB
标题	Ride: Rewarding Impact-driven Exploration For Procedurally-generated Environments
得分	6 6 8
Variance	0.89
Decision	Accept (Poster)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=ryeG924twB
标题	Learning Expensive Coordination: An Event-based Deep Rl Approach
得分	6 8 6
Variance	0.89
Decision	Accept (Poster)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=SJxbHkrKDH
标题	Evolutionary Population Curriculum For Scaling Multi-agent Reinforcement Learning
得分	6 8 6
Variance	0.89
Decision	Accept (Poster)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=S1xitgHtvS
标题	Making Sense Of Reinforcement Learning And Probabilistic Inference
得分	6 6 8
Variance	0.89
Decision	Accept (Spotlight)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=rkxDoJBYPB
标题	Reinforced Genetic Algorithm Learning For Optimizing Computation Graphs
得分	8 6 6
Variance	0.89
Decision	Accept (Poster)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=Sye57xStvB
标题	Never Give Up: Learning Directed Exploration Strategies
得分	6 6 8
Variance	0.89
Decision	Accept (Poster)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=HJgC60EtwB
标题	Robust Reinforcement Learning For Continuous Control With Model Misspecification
得分	6 6 8
Variance	0.89
Decision	Accept (Poster)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=S1l8oANFDH
标题	Synthesizing Programmatic Policies That Inductively Generalize
得分	6 8 6
Variance	0.89
Decision	Accept (Poster)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=r1lOgyrKDS
标题	Adaptive Correlated Monte Carlo For Contextual Categorical Sequence Generation
得分	6 6 8
Variance	0.89
Decision	Accept (Poster)
排名	4
平均得分	6.67
论文地址	https://openreview.net/forum?id=S1evHerYPr
标题	Improving Generalization In Meta Reinforcement Learning Using Neural Objectives
得分	6 6 8
Variance	0.89
Decision	Accept (Spotlight)
排名	5
平均得分	6.33
论文地址	https://openreview.net/forum?id=rJeQoCNYDS
标题	Single Episode Transfer For Differing Environmental Dynamics In Reinforcement Learning
得分	3 8 8
Variance	5.56
Decision	Accept (Poster)
排名	5
平均得分	6.33
论文地址	https://openreview.net/forum?id=H1gX8C4YPr
标题	Decentralized Distributed Ppo: Mastering Pointgoal Navigation
得分	3 8 8
Variance	5.56
Decision	Accept (Poster)
排名	6
平均得分	6.25
论文地址	https://openreview.net/forum?id=SJezGp4YPr
标题	Geometric Insights Into The Convergence Of Nonlinear Td Learning
得分	8 3 6 8
Variance	4.19
Decision	Accept (Poster)
排名	6
平均得分	6.25
论文地址	https://openreview.net/forum?id=BJgZGeHFPH
标题	Dynamics-aware Embeddings
得分	3 8 6 8
Variance	4.19
Decision	Accept (Poster)
排名	7
平均得分	6.2
论文地址	https://openreview.net/forum?id=S1ly10EKDS
标题	Reanalysis Of Variance Reduced Temporal Difference Learning
得分	8 8 6 3 6
Variance	3.36
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=BkglSTNFDB
标题	Q-learning With Ucb Exploration Is Sample Efficient For Infinite-horizon Mdp
得分	6 6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=H1e0Wp4KvH
标题	Automated Curriculum Generation Through Setter-solver Interactions
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=r1xGP6VYwH
标题	Optimistic Exploration Even With A Pessimistic Initialisation
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=Syx7A3NFvH
标题	Multi-agent Reinforcement Learning For Networked System Control
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=BJe1334YDH
标题	A Learning-based Iterative Method For Solving Vehicle Routing Problems
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=rkgpv2VFvr
标题	Sharing Knowledge In Multi-task Deep Reinforcement Learning
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=SJgob6NKvH
标题	Rtfm: Generalising To New Environment Dynamics Via Reading
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=HkgsWxrtPB
标题	Meta Reinforcement Learning With Autonomous Inference Of Subtask Dependencies
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=rke3TJrtPS
标题	Projection Based Constrained Policy Optimization
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=B1x6w0EtwH
标题	Graph Constrained Reinforcement Learning For Natural Language Action Spaces
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=SylOlp4FvH
标题	V-mpo: On-policy Maximum A Posteriori Policy Optimization For Discrete And Continuous Control
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=SJexHkSFPS
标题	Thinking While Moving: Deep Reinforcement Learning With Concurrent Control
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=rke7geHtwH
标题	Keep Doing What Worked: Behavior Modelling Priors For Offline Reinforcement Learning
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=Hyg-JC4FDr
标题	Imitation Learning Via Off-policy Distribution Matching
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=ByxdUySKvS
标题	Adversarial Autoaugment
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=B1gqipNYwH
标题	Option Discovery Using Deep Skill Chaining
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=HJgLLyrYwB
标题	State-only Imitation With Transition Dynamics Mismatch
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=HyxnMyBKwB
标题	The Gambler’s Problem And Beyond
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=B1e-kxSKDH
标题	Structured Object-aware Physics Prediction For Video Modeling And Planning
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=H1lmhaVtvr
标题	Dynamical Distance Learning For Semi-supervised And Unsupervised Skill Discovery
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=SkeIyaVtwB
标题	Exploration In Reinforcement Learning With Deep Covering Options
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=S1lEX04tPr
标题	Cm3: Cooperative Multi-goal Multi-stage Multi-agent Reinforcement Learning
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=ryxB2lBtvH
标题	Learning To Coordinate Manipulation Skills Via Skill Behavior Diversification
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=H1ezFREtwH
标题	Composing Task-agnostic Policies With Deep Reinforcement Learning
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=B1gskyStwr
标题	Frequency-based Search-control In Dyna
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=S1ltg1rFDS
标题	Black-box Off-policy Estimation For Infinite-horizon Reinforcement Learning
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=ryg48p4tPH
标题	Action Semantics Network: Considering The Effects Of Actions In Multiagent Systems
得分	6 6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=BkxXe0Etwr
标题	Caql: Continuous Action Q-learning
得分	6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=SkgC6TNFvr
标题	Reinforced Active Learning For Image Segmentation
得分	6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=Hye1kTVFDS
标题	The Variational Bandwidth Bottleneck: Stochastic Evaluation On An Information Budget
得分	6 6
Variance	0
Decision	Accept (Poster)
排名	8
平均得分	6
论文地址	https://openreview.net/forum?id=H1gzR2VKDH
标题	Hierarchical Foresight: Self-supervised Learning Of Long-horizon Tasks Via Visual Subgoal Generation
得分	6 6
Variance	0
Decision	Accept (Poster)
排名	9
平均得分	5.75
论文地址	https://openreview.net/forum?id=BJliakStvH
标题	Maximum Likelihood Constraint Inference For Inverse Reinforcement Learning
得分	8 6 3 6
Variance	3.19
Decision	Accept (Spotlight)
排名	9
平均得分	5.75
论文地址	https://openreview.net/forum?id=rygfnn4twS
标题	Autoq: Automated Kernel-wise Neural Network Quantization
得分	6 6 8 3
Variance	3.19
Decision	Accept (Poster)
排名	9
平均得分	5.75
论文地址	https://openreview.net/forum?id=Hkl9JlBYvr
标题	Varibad: A Very Good Method For Bayes-adaptive Deep Rl Via Meta-learning
得分	8 6 8 1
Variance	8.19
Decision	Accept (Poster)
排名	10
平均得分	5.67
论文地址	https://openreview.net/forum?id=SJg5J6NtDr
标题	Watch, Try, Learn: Meta-learning From Demonstrations And Rewards
得分	8 3 6
Variance	4.22
Decision	Accept (Poster)
排名	10
平均得分	5.67
论文地址	https://openreview.net/forum?id=rJeINp4KwH
标题	Population-guided Parallel Policy Search For Reinforcement Learning
得分	6 8 3
Variance	4.22
Decision	Accept (Poster)
排名	10
平均得分	5.67
论文地址	https://openreview.net/forum?id=HJgcvJBFvB
标题	A Simple Randomization Technique For Generalization In Deep Reinforcement Learning
得分	8 3 6
Variance	4.22
Decision	Accept (Poster)
排名	10
平均得分	5.67
论文地址	https://openreview.net/forum?id=H1eCw3EKvH
标题	On The Weaknesses Of Reinforcement Learning For Neural Machine Translation
得分	8 6 3
Variance	4.22
Decision	Accept (Poster)
排名	10
平均得分	5.67
论文地址	https://openreview.net/forum?id=rylrdxHFDr
标题	State Alignment-based Imitation Learning
得分	6 8 3
Variance	4.22
Decision	Accept (Poster)
排名	10
平均得分	5.67
论文地址	https://openreview.net/forum?id=rylvYaNYDH
标题	Finding And Visualizing Weaknesses Of Deep Reinforcement Learning Agents
得分	8 6 3
Variance	4.22
Decision	Accept (Poster)
排名	10
平均得分	5.67
论文地址	https://openreview.net/forum?id=Skln2A4YDB
标题	Model-augmented Actor-critic: Backpropagating Through Paths
得分	3 6 8
Variance	4.22
Decision	Accept (Poster)
排名	10
平均得分	5.67
论文地址	https://openreview.net/forum?id=rygf-kSYwH
标题	Behaviour Suite For Reinforcement Learning
得分	8 3 6
Variance	4.22
Decision	Accept (Spotlight)
排名	10
平均得分	5.67
论文地址	https://openreview.net/forum?id=BJluxREKDB
标题	Learning Heuristics For Quantified Boolean Formulas Through Reinforcement Learning
得分	6 8 3
Variance	4.22
Decision	Accept (Poster)
排名	10
平均得分	5.67
论文地址	https://openreview.net/forum?id=Bkg0u3Etwr
标题	Maxmin Q-learning: Controlling The Estimation Bias Of Q-learning
得分	8 6 3
Variance	4.22
Decision	Accept (Poster)
排名	10
平均得分	5.67
论文地址	https://openreview.net/forum?id=ryx6WgStPB
标题	Hypermodels For Exploration
得分	8 3 6
Variance	4.22
Decision	Accept (Poster)
排名	11
平均得分	5.5
论文地址	https://openreview.net/forum?id=ByeWogStDS
标题	Sub-policy Adaptation For Hierarchical Reinforcement Learning
得分	3 8
Variance	6.25
Decision	Accept (Poster)
排名	11
平均得分	5.5
论文地址	https://openreview.net/forum?id=r1xPh2VtPB
标题	Svqn: Sequential Variational Soft Q-learning Networks
得分	3 8
Variance	6.25
Decision	Accept (Poster)
排名	12
平均得分	5.25
论文地址	https://openreview.net/forum?id=BJeGlJStPr
标题	Impact: Importance Weighted Asynchronous Architectures With Clipped Target Networks
得分	6 3 6 6
Variance	1.69
Decision	Accept (Poster)
排名	13
平均得分	5
论文地址	https://openreview.net/forum?id=rJld3hEYvS
标题	排名ing Policy Gradient
得分	6 3 6
Variance	2
Decision	Accept (Poster)
排名	13
平均得分	5
论文地址	https://openreview.net/forum?id=HklxbgBKvr
标题	Model-based Reinforcement Learning For Biological Sequence Design
得分	6 3 6
Variance	2
Decision	Accept (Poster)
排名	13
平均得分	5
论文地址	https://openreview.net/forum?id=HJx-3grYDB
标题	Learning Nearly Decomposable Value Functions Via Communication Minimization
得分	6 6 3
Variance	2
Decision	Accept (Poster)
排名	13
平均得分	5
论文地址	https://openreview.net/forum?id=Byx4NkrtDS
标题	Implementing Inductive Bias For Different Navigation Tasks Through Diverse Rnn Attrractors
得分	3 6 6
Variance	2
Decision	Accept (Poster)
排名	13
平均得分	5
论文地址	https://openreview.net/forum?id=SylL0krYPS
标题	Toward Evaluating Robustness Of Deep Reinforcement Learning With Continuous Control
得分	6 3 6
Variance	2
Decision	Accept (Poster)
排名	13
平均得分	5
论文地址	https://openreview.net/forum?id=rJxX8T4Kvr
标题	Learning Efficient Parameter Server Synchronization Policies For Distributed Sgd
得分	6 3 6
Variance	2
Decision	Accept (Poster)
排名	13
平均得分	5
论文地址	https://openreview.net/forum?id=HkxjqxBYDB
标题	Episodic Reinforcement Learning With Associative Memory
得分	6 3 6
Variance	2
Decision	Accept (Poster)
排名	14
平均得分	4.67
论文地址	https://openreview.net/forum?id=rkecJ6VFvr
标题	Logic And The 2-simplicial Transformer
得分	8 3 3
Variance	5.56
Decision	Accept (Poster)
排名	15
平均得分	4
论文地址	https://openreview.net/forum?id=rkl3m1BFDB
标题	Exploratory Not Explanatory: Counterfactual Analysis Of Saliency Maps For Deep Rl
得分	1 3 8
Variance	8.67
Decision	Accept (Poster)
排名	15
平均得分	4
论文地址	https://openreview.net/forum?id=S1xnXRVFwH
标题	Playing The Lottery With Rewards And Multiple Languages: Lottery Tickets In Rl And Nlp
得分	3 3 6
Variance	2
Decision	Accept (Poster)

 

本文同步推送至

知乎：https://zhuanlan.zhihu.com/c_1196078521515343872

Github: https://github.com/NeuronDance/DeepRL/tree/master/DRL-ConferencePaper/ICLR/2020

交流合作

请加微信号：yan_kylin_phenix，注明姓名+单位+从业方向+地点，非诚勿扰。