Papers › Munchausen Reinforcement Learning

Munchausen Reinforcement Learning

28 Jul 2020NeurIPS 2020 12arXiv:2007.14430archive 2025-07-28

Nino Vieillard, Olivier Pietquin, Matthieu Geist

Bootstrapping is a core mechanism in Reinforcement Learning (RL). Most algorithms, based on temporal differences, replace the true value of a transiting state by their current estimate of this value. Yet, another estimate could be leveraged to bootstrap RL: the current policy. Our core contribution stands in a very simple idea: adding the scaled log-policy to the immediate reward. We show that slightly modifying Deep Q-Network (DQN) in that way provides an agent that is competitive with distributional methods on Atari games, without making use of distributional RL, n-step returns or prioritized replay. To demonstrate the versatility of this idea, we also use it together with an Implicit Quantile Network (IQN). The resulting agent outperforms Rainbow on Atari, installing a new State of the Art with very little modifications to the original algorithm. To add to this empirical study, we provide strong theoretical insights on what happens under the hood -- implicit Kullback-Leibler regularization and increase of the action-gap.

PaperPDFConference PDFCodeCode Syntology ran

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, Syntology's MCP tool lists every function and class Syntology harvested from this paper and whether it ran (how to connect): get_harvested_code_for_paper(arxiv_id="2007.14430")

Code

Syntology Ran 12 of 22 code samples harvested from 2 repositories linked to this paper; 10 have no recorded run. Of those that ran: 1 ran · violated contract; 11 ran with no contract checked.

By repository: community (archive-listed): 22 samples from 2 repositories, 12 ran. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

BY571/Munchausen-RL mentioned on GitHubpytorch report
floringogianu/atari-agents mentioned on GitHubpytorch report
opendilab/DI-engine mentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

22 samples harvested; 12 ran; 0 honoured the contract we drafted; 10 have no recorded run. Read from Syntology's graph 2026-09-24; that is when this build read the record, not when the samples ran.

1ran · violated contract
11ran
10unverified

Licence: 0 of the 22 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 2 repositories linked to this paper, official or community; each sample names its own and says which. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself; “our draft was wrong” and “fixture could not drive it” are failures of Syntology's instrument, not of the code.

Each sample ends with its code_sha256, Syntology's identity for that exact code. An agent fetches the stored sample with Syntology's MCP tool get_code(code_sha256="…") (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples from a repository marked official are listed first. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of four reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; or the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar). Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

CateoricalPolicy lingweizhu/Pytorch-MunchausenActorCritic/Munchausen_actorcritic_discrete/agent/munchausen_ac.py community (archive-listed) ran · metamorphic tier: deterministic MIT (permissive) · 5755feadde38c8e8 · report
DQNBase lingweizhu/Pytorch-MunchausenActorCritic/Munchausen_actorcritic_discrete/agent/munchausen_ac.py community (archive-listed) ran · metamorphic tier: deterministic MIT (permissive) · 81e3bb6b8f3a8963 · report
LazyMemory lingweizhu/Pytorch-MunchausenActorCritic/Munchausen_actorcritic_discrete/agent/munchausen_ac.py community (archive-listed) ran MIT (permissive) · 3dafc0e0e5fc8869 · report
LazyMultiStepMemory lingweizhu/Pytorch-MunchausenActorCritic/Munchausen_actorcritic_discrete/agent/munchausen_ac.py community (archive-listed) ran MIT (permissive) · 95cfd13ecea4b3d9 · report
MPOParams deepmind/acme/acme/jax/losses/mpo.py community (archive-listed) ran · metamorphic tier: deterministic Apache-2.0 (permissive) · b0bd64a5291ade62 · report
MPOStats deepmind/acme/acme/jax/losses/mpo.py community (archive-listed) ran Apache-2.0 (permissive) · 3c56fd89d0f7a624 · report
MultiStepBuff lingweizhu/Pytorch-MunchausenActorCritic/Munchausen_actorcritic_discrete/agent/munchausen_ac.py community (archive-listed) ran MIT (permissive) · a28f79e622630932 · report
MunchausenACAgent lingweizhu/Pytorch-MunchausenActorCritic/Munchausen_actorcritic_discrete/agent/munchausen_ac.py community (archive-listed) ran MIT (permissive) · acef77cb6c31e18f · report
QNetwork lingweizhu/Pytorch-MunchausenActorCritic/Munchausen_actorcritic_discrete/agent/munchausen_ac.py community (archive-listed) ran · metamorphic tier: deterministic MIT (permissive) · c0deed6c236f0456 · report
RunningMeanStats lingweizhu/Pytorch-MunchausenActorCritic/Munchausen_actorcritic_discrete/agent/munchausen_ac.py community (archive-listed) ran MIT (permissive) · 4307d53aa4781ea6 · report
TwinnedQNetwork lingweizhu/Pytorch-MunchausenActorCritic/Munchausen_actorcritic_discrete/agent/munchausen_ac.py community (archive-listed) ran · metamorphic tier: deterministic MIT (permissive) · e4242dcaea044cb7 · report
compute_nonparametric_kl_from_normalized_weights deepmind/acme/acme/jax/losses/mpo.py community (archive-listed) ran · violated contract fingerprinted Apache-2.0 (permissive) · 26aa15e8de106e8a · report
BaseAgent lingweizhu/Pytorch-MunchausenActorCritic/Munchausen_actorcritic_discrete/agent/munchausen_ac.py community (archive-listed) unverified MIT (permissive) · 636a9ac9309ad92c · report
BaseNetwork lingweizhu/Pytorch-MunchausenActorCritic/Munchausen_actorcritic_discrete/agent/munchausen_ac.py community (archive-listed) unverified MIT (permissive) · c8be42d3d03f50ae · report
LazyPrioritizedMultiStepMemory lingweizhu/Pytorch-MunchausenActorCritic/Munchausen_actorcritic_discrete/agent/munchausen_ac.py community (archive-listed) unverified MIT (permissive) · cd0a7dde773aaf77 · report
MPO deepmind/acme/acme/jax/losses/mpo.py community (archive-listed) unverified Apache-2.0 (permissive) · 42136719f9861313 · report
compute_cross_entropy_loss deepmind/acme/acme/jax/losses/mpo.py community (archive-listed) unverified Apache-2.0 (permissive) · 6e0977288094358f · report
compute_parametric_kl_penalty_and_dual_loss deepmind/acme/acme/jax/losses/mpo.py community (archive-listed) unverified Apache-2.0 (permissive) · fb4ceee742119787 · report
compute_weights_and_temperature_loss deepmind/acme/acme/jax/losses/mpo.py community (archive-listed) unverified Apache-2.0 (permissive) · 7fa0e09c9f031f21 · report
disable_gradients lingweizhu/Pytorch-MunchausenActorCritic/Munchausen_actorcritic_discrete/agent/munchausen_ac.py community (archive-listed) unverified MIT (permissive) · 56dd8863f4522252 · report
initialize_weights_he lingweizhu/Pytorch-MunchausenActorCritic/Munchausen_actorcritic_discrete/agent/munchausen_ac.py community (archive-listed) unverified MIT (permissive) · 2b7b028ed940553c · report
update_params lingweizhu/Pytorch-MunchausenActorCritic/Munchausen_actorcritic_discrete/agent/munchausen_ac.py community (archive-listed) unverified MIT (permissive) · 33523284389291bf · report

Tasks

Atari GamesReinforcement LearningReinforcement Learning (RL)reinforcement-learning

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Atari Games Atari-57 M-IQN Mean Human Normalized Score 504% #10 of 11 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

N-step Returns

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections