Replay Memory as An Empirical MDP: Combining Conservative Estimation with Experience Replay

Hongming Zhang, Chenjun Xiao, Han Wang, Jun Jin, Bo Xu +1 more
2/12/2026

No official code found. Best implementation: initial-h/CEER (4 stars, MIT, last commit 2023-06-22).

Semantic Scholar

Code Implementations(1)

initial-h/CEERCommunity57%
MIT

Replay Memory as An Empirical MDP: Combining Conservative Estimation with Experience Replay. ICLR 2023

40PythonMar 28, 20233 years agoMIT

Similar Papers

MSA Generation with Seqs2Seqs Pretraining: Advancing Protein Structure Predictions

95% match2026

Beyond Textual Constraints: Learning Novel Diffusion Conditions with Fewer Examples

95% match1 repo82026

Decentralized Sporadic Federated Learning: A Unified Algorithmic Framework with Convergence Guarantees

95% match1 repo102026

Contrastive Learning Meets Homophily: Two Birds with One Stone

95% match2026

To Tackle Adversarial Transferability: A Novel Ensemble Training Method with Fourier Transformation

95% match1 repo2026

Visual Transformer with Differentiable Channel Selection: An Information Bottleneck Inspired Approach

95% match2026

Cite this paper

@article{zhang2026replay,
  title  = {Replay Memory as An Empirical MDP: Combining Conservative Estimation with Experience Replay},
  author = {Hongming Zhang and Chenjun Xiao and Han Wang and Jun Jin and Bo Xu and Martin Müller},
  year   = {2026},
  url    = {https://api.semanticscholar.org/CorpusID:3b96b6306ed2fc84cb9cf8076c5084849ae62ae0},
  journal = {ICLR 2023 2023}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/3b96b6306ed2fc84cb9cf8076c5084849ae62ae0.svg)](https://codeofpaper.com/paper/3b96b6306ed2fc84cb9cf8076c5084849ae62ae0)

Discussion