参与:杜伟、楚航、罗若天
本周的重要研究包括王者荣耀 AI 绝悟完全体以及全新的目标检测范式 Sparse R-CNN。
目录:
-
Towards Playing Full MOBA Games with Deep Reinforcement Learning
-
Long Range Arena : A Benchmark for Efficient Transformers
-
Sparse R-CNN: End-to-End Object Detection with Learnable Proposals
-
The Mathematical Foundations of Manifold Learning
-
Learning to Reconstruct and Segment 3D Objects
-
U^2-Net: Going Deeper with Nested U-Structure for Salient Object Detection
-
Stylized Neural Painting
-
ArXiv Weekly Radiostation:NLP、CV、ML 更多精选论文(附音频)
-
作者:Deheng Ye、Guibin Chen、Wen Zhang 等
-
论文链接:https://arxiv.org/abs/2011.12692
-
作者:Yi Tay、Mostafa Dehghani、Samira Abnar 等
-
论文链接:https://arxiv.org/pdf/2011.04006.pdf
-
作者:Peize Sun、Rufeng Zhang、Yi Jiang、Tao Kong, 等
-
论文链接:https://arxiv.org/abs/2011.12450
-
作者:Luke Melas-Kyriazi
-
论文链接:https://arxiv.org/pdf/2011.01307.pdf
-
作者:Bo Yang
-
论文链接:https://arxiv.org/pdf/2010.09582.pdf
-
作者:Xuebin Qin、Zichen Zhang、Chenyang Huang 等
-
论文链接:https://arxiv.org/pdf/2005.09007.pdf
-
作者:Zhengxia Zou、Tianyang Shi、Shuang Qiu 等
-
论文链接:https://arxiv.org/pdf/2011.08114.pdf
1. Tight Integrated End-to-End Training for Cascaded Speech Translation. (from Hermann Ney)2. Two-Way Neural Machine Translation: A Proof of Concept for Bidirectional Translation Modeling using a Two-Dimensional Grid. (from Hermann Ney)3. GLGE: A New General Language Generation Evaluation Benchmark. (from Ruofei Zhang)4. ONION: A Simple and Effective Defense Against Textual Backdoor Attacks. (from Maosong Sun)5. Cross-Document Event Coreference Resolution Beyond Corpus-Tailored Systems. (from Iryna Gurevych)6. Acoustic span embeddings for multilingual query-by-example search. (from Karen Livescu)7. XTQA: Span-Level Explanations of the Textbook Question Answering. (from Jun Liu)8. Enhancing deep neural networks with morphological information. (from Marko Robnik-Šikonja)9. A Panoramic Survey of Natural Language Processing in the Arab World. (from Kareem Darwish)10. 1st AfricaNLP Workshop Proceedings, 2020. (from Vukosi Marivate)本周 10 篇 CV 精选论文是: 1. Temporal Action Detection with Multi-level Supervision. (from Kate Saenko, Trevor Darrell)2. Exploring Simple Siamese Representation Learning. (from Kaiming He)3. Emotional Semantics-Preserved and Feature-Aligned CycleGAN for Visual Emotion Adaptation. (from Alberto L. Sangiovanni-Vincentelli, Kurt Keutzer)4. Learning to Sample the Most Useful Training Patches from Images. (from Liang Chen, Philip Torr)5. SLADE: A Self-Training Framework For Distance Metric Learning. (from Larry Davis, C.-C. Jay Kuo)6. Building 3D Morphable Models from a Single Scan. (from Joshua Tenenbaum)7. Attention Aware Cost Volume Pyramid Based Multi-view Stereo Network for 3D Reconstruction. (from Bing Liu)8. Unsupervised Discovery of DisentangledManifolds in GANs. (from Ming-Hsuan Yang)9. SegBlocks: Block-Based Dynamic Resolution Networks for Real-Time Segmentation. (from Tinne Tuytelaars)10. MicroNet: Towards Image Recognition with Extremely Low FLOPs. (from Zicheng Liu, Lei Zhang, Nuno Vasconcelos)本周 10 篇 ML 精选论文是:1. Energy-Based Models for Continual Learning. (from Antonio Torralba)2. TLeague: A Framework for Competitive Self-Play based Distributed Multi-Agent Reinforcement Learning. (from Zhengyou Zhang)3. Unsupervised learning of disentangled representations in deep restricted kernel machines with orthogonality constraints. (from Johan A. K. Suykens)4. Equivariant Conditional Neural Processes. (from Yee Whye Teh)5. MetaGater: Fast Learning of Conditional Channel Gated Networks via Federated Meta-Learning. (from Junshan Zhang)6. Ensemble- and Distance-Based Feature Ranking for Unsupervised Learning. (from Sašo Džeroski)7. Convergence Analysis of Homotopy-SGD for non-convex optimization. (from Moritz Diehl, Frank Hutter)8. Explainable Multivariate Time Series Classification: A Deep Neural Network Which Learns To Attend To Important Variables As Well As Informative Time Intervals. (from Vasant Honavar)9. No Subclass Left Behind: Fine-Grained Robustness in Coarse-Grained Classification Problems. (from Christopher Ré)10. Cyclic Label Propagation for Graph Semi-supervised Learning. (from Jiajun Bu)