Unverified50% confidenceFactExact time
IQL（Implicit Q-Learning）由Kostrikov等人于2021年提出，通过期望回归规避对分布外动作的直接查询
1
Sources
50%
Confidence
Long-term
Relevance
7/17/2026
First Seen
Sources
Related Claims
UnverifiedC51（Categorical DQN）由Bellemare、Dabney和Munos于2017年提出，是分布式强化学习的开山之作61% similarUnverifiedAkyürek et al. 2022 argued from the perspective of gradient descent that ICL is essentially implicit meta-learning.60% similarUnverifiedCFR由Zinkevich等人于2007年在NIPS提出，理论基础是在线学习中的遗憾界理论，平均策略在迭代次数趋于无穷时收敛至纳什均衡60% similarUnverifiedIn 2022, Andrew Ng updated and launched a new version of the Machine Learning Specialization on Coursera, incorporating model-tuning methodologies such as bias-variance tradeoff and error analysis.59% similarVerifiedQ-learning is a value-function method in reinforcement learning, proposed by Watkins in 1989, whose core involves maintaining a Q-table and iteratively updating it via the Bellman equation.57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/547321API
curl https://kongchang.com/api/v1/knowledge/claims/547321MCP
get_claim(id=547321)