Pith. sign in

REVIEW 6 cited by

Mirror Descent Policy Optimization

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2005.09814 v5 pith:FIN5YJAD submitted 2020-05-20 cs.LG cs.AIstat.ML

classification cs.LGcs.AIstat.ML
keywords mdpoalgorithmstrust-regiondescentmirrorobjectiveoptimizationpolicy
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Mirror descent (MD), a well-known first-order method in constrained convex optimization, has recently been shown as an important tool to analyze trust-region algorithms in reinforcement learning (RL). However, there remains a considerable gap between such theoretically analyzed algorithms and the ones used in practice. Inspired by this, we propose an efficient RL algorithm, called {\em mirror descent policy optimization} (MDPO). MDPO iteratively updates the policy by {\em approximately} solving a trust-region problem, whose objective function consists of two terms: a linearization of the standard RL objective and a proximity term that restricts two consecutive policies to be close to each other. Each update performs this approximation by taking multiple gradient steps on this objective function. We derive {\em on-policy} and {\em off-policy} variants of MDPO, while emphasizing important design choices motivated by the existing theory of MD in RL. We highlight the connections between on-policy MDPO and two popular trust-region RL algorithms: TRPO and PPO, and show that explicitly enforcing the trust-region constraint is in fact {\em not} a necessity for high performance gains in TRPO. We then show how the popular soft actor-critic (SAC) algorithm can be derived by slight modifications of off-policy MDPO. Overall, MDPO is derived from the MD principles, offers a unified approach to viewing a number of popular RL algorithms, and performs better than or on-par with TRPO, PPO, and SAC in a number of continuous control tasks. Code is available at \url{https://github.com/manantomar/Mirror-Descent-Policy-Optimization}.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 6 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. OpenAlex reports about 24 citations worldwide. Full citation record

  1. On the Policy Convergence of Policy Mirror Descent Methods

    math.OC 2026-07 accept novelty 7.0 of 10

    Unregularized PMD with any constant step size converges to a limiting optimal policy for general decomposable Legendre mirror maps, with behavior governed by differentiability of ψ at 0 and 1.

  2. StaQ it! Growing neural networks for Policy Mirror Descent

    cs.LG 2025-06 conditional novelty 7.0 of 10

    StaQ, a finite-memory Policy Mirror Descent algorithm, converges to the optimal entropy-regularized policy with a sufficiently large window of past Q-functions and performs competitively with baselines.

  3. On the Effect of Regularization in Policy Mirror Descent

    cs.LG 2025-07 conditional novelty 6.0 of 10

    A large empirical sweep shows that in Policy Mirror Descent, MDP and Drift regularizers are partly substitutable, yet their precise combination determines temperature robustness.

  4. Convergence and Sample Complexity of First-Order Methods for Agnostic Reinforcement Learning

    cs.LG 2025-07 reject novelty 6.0 of 10

    Under variational gradient dominance, the paper derives state-space-independent sample complexity bounds for SDPO, CPI, DA-CPI, and PMD in agnostic policy learning.

  5. Efficient Online Reinforcement Learning for Diffusion Policy

    cs.LG 2025-02 conditional novelty 6.0 of 10

    Diffusion policies can be trained online with reweighted score matching using only Q-functions, and the resulting DPMD and SDAC algorithms beat SAC and prior diffusion-policy RL on most MuJoCo tasks.

  6. A Single Diffusion-Policy Controller for Multi-Task Block Pushing with Zero-Shot Sim-to-Real Transfer

    cs.RO 2026-07 conditional novelty 5.0 of 10

    One diffusion policy trained via energy-guided RL solves multi-shape block pushing without demos and transfers zero-shot to real robots under varied conditions.

Pith tools