摘要:确定生物分子的 3D 形状是现代生物学和医学发现中最困难的问题之一。许多公司和研究机构花费数百万美元来确定分子结构,却也常常无果。来自斯坦福大学的研究团队利用机器学习的方法解决了这个难题。在计算机科学系副教授 Ron Dror 的指导下,斯坦福大学博士生 Stephan Eismann 和 Raphael Townshend 巧妙地使用机器学习技术开发了一种通过计算预测生物分子准确结构的方法。并且即使仅从少数已知结构中学习,他们的方法也能成功,使其适用于结构最难通过实验确定的分子类型. 8 月 27 日,该团队与斯坦福大学生物化学系副教授 Rhiju Das 合作的研究论文在《Science》上发表并登上封面。
1. Semantic-Preserving Adversarial Text Attacks. (from Dacheng Tao, Wei Liu)2. Open Relation Modeling: Learning to Define Relations between Entities. (from Kevin Chen-Chuan Chang, Wen-mei Hwu)3. Reducing Exposure Bias in Training Recurrent Neural Network Transducers. (from Brian Kingsbury)4. A Framework for Neural Topic Modeling of Text Corpora. (from Majid Sarrafzadeh)5. One Chatbot Per Person: Creating Personalized Chatbots based on Implicit User Profiles. (from Ji-Rong Wen)6. A Hierarchical Entity Graph Convolutional Network for Relation Extraction across Documents. (from Hwee Tou Ng)7. Exploring the Promises of Transformer-Based LMs for the Representation of Normative Claims in the Legal Domain. (from Siegfried Handschuh)8. Prompt-Learning for Fine-Grained Entity Typing. (from Juanzi Li)9. LayoutReader: Pre-training of Text and Layout for Reading Order Detection. (from Furu Wei)10. Just Say No: Analyzing the Stance of Neural Dialogue Generation in Offensive Contexts. (from Alan Ritter) 10 CV Papers.mp3音频:进度条00:00/21:22
本周 10 篇 CV 精选论文是:
1. Airbert: In-domain Pretraining for Vision-and-Language Navigation. (from Ivan Laptev, Cordelia Schmid)2. Region-level Active Learning for Cluttered Scenes. (from Trevor Darrell)3. A Synthesis-Based Approach for Thermal-to-Visible Face Verification. (from Rama Chellappa)4. Discriminative Region-based Multi-Label Zero-Shot Learning. (from Ling Shao, Mubarak Shah)5. Learning Motion Priors for 4D Human Body Capture in 3D Scenes. (from Yan Zhang, Marc Pollefeys)6. Glimpse-Attend-and-Explore: Self-Attention for Active Visual Exploration. (from Tinne Tuytelaars)7. BlockCopy: High-Resolution Video Processing with Block-Sparse Feature Propagation and Online Policies. (from Tinne Tuytelaars)8. Layer-wise Customized Weak Segmentation Block and AIoU Loss for Accurate Object Detection. (from Lei Zhang)9. Reconcile Prediction Consistency for Balanced Object Detection. (from Lei Zhang)10. Recall@k Surrogate Loss with Large Batches and Similarity Mixup. (from Jiri Matas) 10 ML Papers.mp3音频:进度条00:00/21:55
本周 10 篇 ML 精选论文是:
1. DSKReG: Differentiable Sampling on Knowledge Graph for Recommendation with Relational GNN. (from Philip S. Yu)2. Explaining Bayesian Neural Networks. (from Klaus-Robert Müller)3. SreaMRAK a Streaming Multi-Resolution Adaptive Kernel Algorithm. (from Yoav Freund)4. Latent Space Energy-Based Model of Symbol-Vector Coupling for Text Generation and Classification. (from Bo Pang)5. sigmoidF1: A Smooth F1 Score Surrogate Loss for Multilabel Classification. (from Maarten de Rijke)6. Adaptive Control of Differentially Private Linear Quadratic Systems. (from Ness Shroff)7. PASTO: Strategic Parameter Optimization in Recommendation Systems -- Probabilistic is Better than Deterministic. (from Yang Liu)8. Plug and Play, Model-Based Reinforcement Learning. (from Svetha Venkatesh)9. Integer-arithmetic-only Certified Robustness for Quantized Neural Networks. (from Cyrus Shahabi)10. SemiFed: Semi-supervised Federated Learning with Consistency and Pseudo-Labeling. (from Cyrus Shahabi)