ICML 2022开奖!复旦、上交、厦大多篇工作入选杰出论文
新智元
共 7868字,需浏览 16分钟
·
2022-07-25 21:34
新智元报道
新智元报道
【新智元导读】ICML 2022杰出论文揭晓!复旦、上交、厦大等中国高校上榜,吴恩达、Jeff Dean等获时间检验奖荣誉提名。|还在纠结会不会错过元宇宙和web3浪潮?清华大学科学史系副教授胡翌霖,这次给你讲个透!
杰出论文(Outstanding Papers)
Monarch: Expressive Structured Matrices for Efficient and Accurate Training
Solving Stackelberg Prediction Game with Least Squares Loss via Spherically Constrained Least Squares Reformulation
Adversarially Trained Actor Critic for Offline Reinforcement Learning
Stable Conformal Prediction Sets
Do Differentiable Simulators Give Better Policy Gradients?
Learning inverse folding from millions of predicted structures
Learning Mixtures of Linear Dynamical Systems
Causal Conceptions of Fairness and their Consequences
Active fairness auditing
Understanding Dataset Difficulty with V-Usable Information
G-Mixup: Graph Data Augmentation for Graph Classification
The Importance of Non-Markovianity in Maximum State Entropy Exploration
Minimum Cost Intervention Design for Causal Effect Identification
Privacy for Free: How does Dataset Condensation Help Privacy?
Bayesian Model Selection, the Marginal Likelihood, and Generalization
Test of Time Award (时间检验奖)
本届时间检验奖授予了十年前ICML的一篇论文,该论文对机器学习领域产生了实质性影响,包括研究和实践。
评论