Reproduction: OpPMD Discounted MDPs (ICML 2026 #21515) - a debajyotidasgupta Collection

debajyotidasgupta 's Collections

Repro: CoT Controllability (ICML 2026 #18936)

Reproduction: OpPMD Discounted MDPs (ICML 2026 #21515)

MindFlow reproduction (ICML 2026 #1894)

SpecPrune-VLA reproduction

OrderPlace ICML-2026 Reproduction

Repro: Learning for Branch-and-Cut (ICML26)

Reproduction: FOCUS DLLMs Compute Bound (ICML 2026)

AsyncSpade ICML-2026 Reproduction

ICML 2026 Repro: The Label Horizon Paradox

EU-Halt — Frontier Panel (When Uncertainty Lies)

EU-Halt — Evaluation Datasets

EU-Halt — Production Heads

Reproduction: OpPMD Discounted MDPs (ICML 2026 #21515)

updated 28 days ago

Independent reproduction of ICML 2026 #21515 (arXiv 2502.15345): OpPMD, Theorems 1-2, and the Dist regime on the paper's tabular MDP.