作者：IODA 时间：2020-03-24 点击数：
ICLR是公认的深度学习领域国际顶级会议之一，关注有关深度学习各个方面的前沿研究，在人工智能、统计和数据科学领域以及机器视觉、语音识别、文本理解等重要应用领域中发布了众多极其有影响力的论文。ICLR 2020共有2594篇论文提交，其中48篇被接收为Talk Paper，107篇被接收为Spotlight Paper，532篇作为Poster Paper。
以下是由AI研习社整理的 ICLR 2020 论文实现代码，供大家学习！
60.BinaryDuo: Reducing Gradient Mismatch in Binary Activation Network by Coupling Binary Activations
论文地址：https://openreview.net/forum?id=r1x0lxrFPS
Github地址：https://github.com/Hyungjun-K1m/BinaryDuo
59.BayesOpt Adversarial Attack
论文地址：https://openreview.net/forum?id=Hkem-lrtvH
Github地址：https://github.com/rubinxin/BayesOpt_Attack
58.RaPP: Novelty Detection with Reconstruction along Projection Pathway
论文地址：https://openreview.net/forum?id=HkgeGeBYDB
Github地址：https://drive.google.com/drive/folders/1sknl_i4zmvSsPYZdzYxbg66ZSYDZ_abg?usp=sharing
57.Dynamics-Aware Embeddings
论文地址：https://openreview.net/forum?id=BJgZGeHFPH
Github地址：https://github.com/dyne-submission/dynamics-aware-embeddings
56.AdvectiveNet: An Eulerian-Lagrangian Fluidic Reservoir for Point Cloud Processing
论文地址：https://openreview.net/forum?id=H1eqQeHFDS
Github地址： https://github.com/DIUDIUDIUDIUDIU/AdvectiveNet-An-Eulerian-Lagrangian-Fluidic-Reservoir-for-Point-Cloud-Processing
55.AtomNAS: Fine-Grained End-to-End Neural Architecture Search
论文地址：https://openreview.net/forum?id=BylQSxHFwr
Github地址：https://github.com/meijieru/AtomNAS
54.Deep Audio Priors Emerge From Harmonic Convolutional Networks
论文地址：https://openreview.net/forum?id=rygjHxrYDB
Github地址：http://dap.csail.mit.edu/
53.Expected Information Maximization: Using the I-Projection for Mixture Density Estimation
论文地址：https://openreview.net/forum?id=ByglLlHFDS
Github地址：https://github.com/pbecker93/ExpectedInformationMaximization
52.A Meta-Transfer Objective for Learning to Disentangle Causal Mechanisms
论文地址：https://openreview.net/forum?id=ryxWIgBFPS
Github地址：https://github.com/ec6dde01667145e58de60f864e05a4/CausalOptimizationAnon
51.Latent Normalizing Flows for Many-to-Many Cross-Domain Mappings
论文地址：https://openreview.net/forum?id=SJxE8erKDH
Github地址：https://drive.google.com/file/d/15orWOM0WowN4OevkHstx600TgNtAXGcp/view?usp=sharing
50.Compositional Language Continual Learning
论文地址：https://openreview.net/forum?id=rklnDgHtDS
Github地址：https://github.com/yli1/CLCL
49.Lipschitz constant estimation of Neural Networks via sparse polynomial optimization
论文地址：https://openreview.net/forum?id=rJe4_xSFDB
Github地址：https://drive.google.com/drive/folders/1bkj0H6Thgd9sjRloyq9NBP0uO0v704E9?usp=sharing
48.Unrestricted Adversarial Examples via Semantic Manipulation
论文地址：https://openreview.net/forum?id=Sye_OgHFwH
Github地址：https://www.dropbox.com/s/69zx437t1dgo41b/semantic_attack_code.zip?dl=0
47.DiffTaichi: Differentiable Programming for Physical Simulation
论文地址：https://openreview.net/forum?id=B1eB5xSFvr
Github地址：https://github.com/yuanming-hu/difftaichi
46.Sub-policy Adaptation for Hierarchical Reinforcement Learning
论文地址：https://openreview.net/forum?id=ByeWogStDS
Github地址：https://anonymous.4open.science/r/de105a6d-8f8b-405e-b90a-54ab74adcb17/
45.Co-Attentive Equivariant Neural Networks: Focusing Equivariance On Transformations Co-Occurring in Data
论文地址：https://openreview.net/forum?id=r1g6ogrtDr
Github地址：https://www.dropbox.com/sh/2gghao89strdotw/AAAYJ6XclnfeoS3AfN9Z-n5Wa?dl=0
44.Information Geometry of Orthogonal Initializations and Training
论文地址：https://openreview.net/forum?id=rkg1ngrFPr
Github地址：https://github.com/mandt-lab/adversarial-negative-sampling
43.Extreme Classification via Adversarial Softmax Approximation
论文地址：https://openreview.net/forum?id=rJxe3xSYDS
42.NAS-Bench-1Shot1: Benchmarking and Dissecting One-shot Neural Architecture Search
论文地址：https://openreview.net/forum?id=SJx9ngStPH
Github地址：https://github.com/automl/nasbench-1shot1
41.The Shape of Data: Intrinsic Distance for Data Distributions
论文地址：https://openreview.net/forum?id=HyebplHYwB
Github地址：https://github.com/xgfs/imd
40.Enabling Deep Spiking Neural Networks with Hybrid Conversion and Spike Timing Dependent Backpropagation
论文地址：https://openreview.net/forum?id=B1xSperKvH
Github地址：https://github.com/nitin-rathi/hybrid-snn-conversion
39.Massively Multilingual Sparse Word Representations
论文地址：https://openreview.net/forum?id=HyeYTgrFPB
Github地址：https://github.com/begab/mamus
38.Learning The Difference That Makes A Difference With Counterfactually-Augmented Data
论文地址：https://openreview.net/forum?id=Sklgs0NFvr
Github地址：https://github.com/dkaushik96/counterfactually-augmented-data
37.Fractional Graph Convolutional Networks (FGCN) for Semi-Supervised Learning
论文地址：https://openreview.net/forum
Github地址：https://github.com/yuzhouchen92/
36.Unsupervised Learning of Graph Hierarchical Abstractions with Differentiable Coarsening and Optimal Transport
论文地址：https://openreview.net/forum?id=Bkf4XgrKvS
Github地址：https://github.com/matenure/OTCoarsening
35.Deep Semi-Supervised Anomaly Detection
论文地址：https://arxiv.org/abs/1801.03149
Github地址：https://github.com/lukasruff/Deep-SAD-PyTorch
34.Model Based Reinforcement Learning for Atari
论文地址：https://openreview.net/forum?id=S1xCPJHtDB
Github地址：http://bit.ly/2wjgn1a
33.And the Bit Goes Down: Revisiting the Quantization of Neural Networks
论文地址：https://openreview.net/forum?id=rJehVyrKwH
Github地址：https://drive.google.com/file/d/12QK7onizf2ArpEBK706ly8bNfiM9cPzp/view?usp=sharing
32.Behaviour Suite for Reinforcement Learning
论文地址：https://openreview.net/forum?id=rygf-kSYwH
Github地址： https://github.com/deepmind/bsuite
31.NAS-Bench-201: Extending the Scope of Reproducible Neural Architecture Search
论文地址：https://openreview.net/forum?id=HJxyZkBKDr
Github地址： https://github.com/D-X-Y/NAS-Bench-201
30.Network Deconvolution
论文地址：https://openreview.net/forum?id=rkeu30EtvS
Github地址： https://github.com/yechengxi/deconvolution
29.Simplified Action Decoder for Deep Multi-Agent Reinforcement Learning
论文地址：https://openreview.net/forum?id=B1xm3RVtwB
Github地址： https://bit.ly/2mBJLyk
28.Learning to Plan in High Dimensions via Neural Exploration-Exploitation Trees
论文地址：https://openreview.net/forum?id=rJgJDAVKvB
Github地址： https://github.com/NeurEXT/NEXT-learning-to-plan/blob/master/main.ipynb
27.Disentangling neural mechanisms for perceptual grouping
论文地址：https://openreview.net/forum?id=HJxrVA4FDS
Github地址：https://bit.ly/2wdQYGd
26. The Logical Expressiveness of Graph Neural Networks
论文地址：https://openreview.net/forum?id=r1lZ7AEKvB
Github地址：https://anonymous.4open.science/r/787222e2-ad5e-4810-a788-e80f0fe7eff0/
25. Influence-Based Multi-Agent Exploration
论文地址：https://openreview.net/forum?id=BJgy96EYvr
Github地址：https://github.com/TonghanWang/EITI-EDTI
24. Ridge Regression: Structure, Cross-Validation, and Sketching
论文地址：https://openreview.net/forum?id=HklRwaEKwB
Github地址：https://github.com/liusf15/RidgeRegression
23. Encoding word order in complex embeddings
论文地址：https://openreview.net/forum?id=Hke-WTVtwr
Github地址：https://github.com/iclr-complex-order/complex-order
22. Enhancing Adversarial Defense by k-Winners-Take-All
论文地址：https://openreview.net/forum?id=Skgvy64tvr
Github地址：https://github.com/a554b554/kWTA-Activation
21. PC-DARTS: Partial Channel Connections for Memory-Efficient Architecture Search
论文地址：https://openreview.net/forum?id=BJlS634tPr
Github地址：https://www.dropbox.com/sh/on9lg3rpx1r6dkf/AABG5mt0sMHjnEJyoRnLEYW4a?dl=0
20. Intensity-Free Learning of Temporal Point Processes
论文地址：https://openreview.net/forum?id=HygOjhEYDH
Github地址：https://github.com/shchur/ifl-tpp
19. Estimating Gradients for Discrete Random Variables by Sampling without Replacement
论文地址：https://openreview.net/forum?id=rklEj2EFvB
Github地址：https://github.com/wouterkool/estimating-gradients-without-replacement
18. Graph Neural Networks Exponentially Lose Expressive Power for Node Classification
论文地址：https://openreview.net/forum?id=S1ldO2EFPr
Github地址： https://github.com/delta2323/gnn-asymptotics
17. BackPACK: Packing more into Backprop
论文地址：https://arxiv.org/abs/1912.10985
Github地址：https://github.com/toiaydcdyywlhzvlob/backpack
16. AtomNAS: Fine-Grained End-to-End Neural Architecture Search
论文地址：https://arxiv.org/abs/1912.12522
15.FasterSeg: Searching for Faster Real-time Semantic Segmentation
论文地址：https://paperswithcode.com/paper/fasterseg-searching-for-faster-real-time
Github地址：https://github.com/TAMU-VITA/FasterSeg
14. NAS evaluation is frustratingly hard
Github地址：https://github.com/antoyang/NAS-Benchmark
13. Geom-GCN: Geometric Graph Convolutional Networks
论文地址：https://openreview.net/forum?id=S1e2agrFvS
Github地址：https://github.com/graphdml-uiuc-jlu/geom-gcn
12. Deep Graph Matching Consensus
论文地址：https://arxiv.org/abs/2001.02525?context=cs
Github地址：https://github.com/JaminFong/FNA
11. Convolutional Convolutional Neural Processes
论文地址：https://arxiv.org/abs/1910.13556v1
Github地址：https://github.com/cambridge-mlg/convcnp
10. Neural Machine Translation with Universal Visual Representation
论文地址：暂无
Github地址：https://github.com/cooelf/UVR-NMT
9. A Neural Dirichlet Process Mixture Model for Task-Free Continual Learning
论文地址：https://arxiv.org/abs/2001.0689
Github地址：https://github.com/soochan-lee/CN-DPM
8. Gradient-Based Neural DAG Learning
论文地址：https://arxiv.org/abs/1906.02226
Github地址：https://github.com/kurowasan/GraN-DAG
7. Scale-Equivariant Steerable Networks
论文地址：https://arxiv.org/abs/1910.11093v1
Github地址：https://github.com/ISosnovik/sesn
6.Composition-Based Multi-Relational Graph Convolutional Networks
论文地址：https://arxiv.org/abs/1802.04944v1
Github地址：https://github.com/malllabiisc/CompGCN
5.Automatically Discovering and Learning New Visual Categories with Ranking Statistics
论文地址：
Github地址：https://github.com/k-han/AutoNovel
4.Pitfalls of In-Domain Uncertainty Estimation and Ensembling in Deep Learning
Github地址：https://github.com/bayesgroup/pytorch-ensembles
3. Strategies for Pre-training Graph Neural Networks
Github地址：https://github.com/snap-stanford/pretrain-gnns
2. Learning to Retrieve Reasoning Paths over Wikipedia Graph for Question Answering
Github地址：https://github.com/AkariAsai/learning_to_retrieve_reasoning_paths
1. FreeLB: Enhanced Adversarial Training for Language Understanding
论文地址：https://arxiv.org/abs/1909.11764v2
Github地址：https://github.com/zhuchen03/FreeLB
