编译|极市平台
转载自|极市平台
1. star:9819|Weakly Supervised Disentanglement with Guarantees(弱监督学习)
2. star:9819|Measuring Compositional Generalization: A Comprehensive Method on Realistic Data
3. star:9819|Meta-Learning without Memorization(元学习/小样本图像分类)
4. star:4977|U-GAT-IT: Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation(图像翻译/无监督)
5. star:2106|On the Variance of the Adaptive Learning Rate and Beyond
6. star:1469|DiffTaichi: Differentiable Programming for Physical Simulation
7. star:1018|Generative Models for Effective ML on Private, Decentralized Datasets
8. star:963|Behaviour Suite for Reinforcement Learning(强化学习)
9. star:534|Contrastive Representation Distillation(知识蒸馏)
10. star:516|On the Relationship between Self-Attention and Convolutional Layers(注意力机制)
11. star:469|AugMix: A Simple Data Processing Method to Improve Robustness and Uncertainty
12. star:443|NAS-Bench-201: Extending the Scope of Reproducible Neural Architecture Search(神经网络架构搜索)
13. star:393|Once for All: Train One Network and Specialize it for Efficient Deployment(神经网络训练)
14. star:246|BatchEnsemble: An Alternative Approach to Efficient Ensemble and Lifelong Learning(神经网络训练)
15. star:243|FasterSeg: Searching for Faster Real-time Semantic Segmentation(语义分割)
16. star:213|Contrastive Learning of Structured World Models
17. star:191|Real or Not Real, that is the Question(GAN)
18. star:186|Pseudo-LiDAR++: Accurate Depth for 3D Object Detection in Autonomous Driving(3D目标检测)
19. star:182|Learning to Explore using Active Neural SLAM(三维SLAM)
20. star:175|Mutual Mean-Teaching: Pseudo Label Refinery for Unsupervised Domain Adaptation on Person Re-identification(行人重识别/无监督)
21. star:132|AtomNAS: Fine-Grained End-to-End Neural Architecture Search(神经网络架构搜索)
22. star:128|Strategies for Pre-training Graph Neural Networks(神经网络训练)
23. star117|Towards Stabilizing Batch Statistics in Backward Propagation of Batch Normalization(归一化)
24. star:107|DropEdge: Towards Deep Graph Convolutional Networks on Node Classification
25. star:107|Neural Arithmetic Units
26. star:106|Semantically-Guided Representation Learning for Self-Supervised Monocular Depth(单目深度估计)
27. star:100|Composition-based Multi-Relational Graph Convolutional Networks
28. star:93|Deformable Kernels: Adapting Effective Receptive Fields for Object Deformation(图像分割/目标检测)
29. star:80|NAS evaluation is frustratingly hard(神经网络架构搜索)
30. star:74|Understanding and Robustifying Differentiable Architecture Search(图像分类)
31. star:72|Fast Neural Network Adaptation via Parameter Remapping and Architecture Search(图像分类/目标检测/语义分割)
32. star:72|Capsules with Inverted Dot-Product Attention Routing(图像分类)
33. star:53|Deep Semi-Supervised Anomaly Detection(异常检测)
34. star:51|Network Deconvolution(图像分类)
35. star:49|Pitfalls of In-Domain Uncertainty Estimation and Ensembling in Deep Learning(图像分类)
36. star:36|A Neural Dirichlet Process Mixture Model for Task-Free Continual Learning(图像分类)
37. star:33|Empirical Bayes Transductive Meta-Learning with Synthetic Gradients(小样本图像分类/元学习)
38. star:32|Query2box: Reasoning over Knowledge Graphs in Vector Space using Box Embeddings(知识图谱)
39. star:27|Kaleidoscope: An Efficient, Learnable Representation For All Structured Linear Maps(图像分类)
40. star:22|Fooling Detection Alone is Not Enough: Adversarial Attack against Multiple Object Tracking(多目标跟踪)