Papers › Evolution Strategies as a Scalable Alternative to Reinforcement Learning

Evolution Strategies as a Scalable Alternative to Reinforcement Learning

10 Mar 2017arXiv:1703.03864archive 2025-07-28

Tim Salimans, Jonathan Ho, Xi Chen, Szymon Sidor, Ilya Sutskever

We explore the use of Evolution Strategies (ES), a class of black box optimization algorithms, as an alternative to popular MDP-based RL techniques such as Q-learning and Policy Gradients. Experiments on MuJoCo and Atari show that ES is a viable solution strategy that scales extremely well with the number of CPUs available: By using a novel communication strategy based on common random numbers, our ES implementation only needs to communicate scalars, making it possible to scale to over a thousand parallel workers. This allows us to solve 3D humanoid walking in 10 minutes and obtain competitive results on most Atari games after one hour of training. In addition, we highlight several advantages of ES as a black box optimization technique: it is invariant to action frequency and delayed rewards, tolerant of extremely long horizons, and does not need temporal discounting or value function approximation.

PaperPDFCodeCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

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

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 29 code samples: 28 from 6 repositories linked to this paper, 1 identical to code harvested from another paper's repository. Across all 29: 16 ran (of which 0 constructed an object rather than computing a result; 11 with no instrument failure: 0 honoured, 1 violated, 10 with no contract checked; 5 where Syntology's instrument failed) · 13 unverified. Where Syntology's instrument failed: 4 where our contract draft was wrong; 1 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: official repository: 5 samples from 1 repository, 4 ran (1 with no instrument failure; 3 where Syntology's instrument failed); community (archive-listed): 23 samples from 5 repositories, 11 ran (10 with no instrument failure; 1 where Syntology's instrument failed); 1 identical to code first harvested elsewhere. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

No sample is linked to this paper from 2 of its repositories that Syntology harvested for other papers:

Levels written between 2026-09-22 and 2026-09-26 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

23 repositories listed; official and paper-mentioned ones first.

openai/evolution-strategies-starter officialmentioned on GitHubtfSyntology: 5 samples, 4 ran (3 instrument-failed) · 1 unverified report
AnshMittal1811/AugmentedRandomSearch mentioned on GitHubSyntology: not harvested report
FlixCoder/rust-es-optimizer mentioned on GitHubApache-2.0Syntology: not harvested report
ShangtongZhang/DistributedES mentioned on GitHubpytorchSyntology: 3 samples, 0 ran · 3 unverified report
alisidd/Evolution-Strategies mentioned on GitHubtfSyntology: not harvested report
aspk/space_battle mentioned on GitHubGPL-3.0Syntology: not harvested report
atgambardella/pytorch-es mentioned on GitHubpytorchSyntology: 2 samples, 1 ran (1 instrument-failed) · 1 unverified report
cesch97/NeuroEvolution mentioned on GitHubSyntology: not harvested report
czen88/qtrader mentioned on GitHubtfApache-2.0Syntology: 10 samples, 5 ran (0 instrument-failed) · 5 unverified report
enajx/ES mentioned on GitHubpytorchSyntology: not harvested report
evaboost/evaboost mentioned on GitHubApache-2.0Syntology: 6 samples, 3 ran (0 instrument-failed) · 3 unverified report
fiberleif/evolution-strategies mentioned on GitHubtfSyntology: not harvested report
neilsgp/RL-Algorithms mentioned on GitHubSyntology: not harvested report
neuro-soup/es-torch mentioned on GitHubjaxMITSyntology: not harvested report
nnaisense/pgpelib mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Adam: A Method for Stochastic Optimization (arXiv:1412.6980; that paper's own run record: 27 ran (of which 7 constructed an object rather than computing a result; 23 with no instrument failure: 4 honoured, 2 violated, 17 with no contract checked; 4 where Syntology's instrument failed) · 10 unverified); nothing checks that they implement this paper's method. report
patniemeyer/ga-autoencoder mentioned on GitHubpytorchSyntology: not harvested report
rasmusbergpalm/evostrat mentioned on GitHubpytorchMITSyntology: 2 samples, 2 ran (0 instrument-failed) · 0 unverified report
shakti365/nes mentioned on GitHubSyntology: not harvested report
silvialuu/DRL-2018 mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Proximal Policy Optimization Algorithms (arXiv:1707.06347; that paper's own run record: 99 ran (of which 53 constructed an object rather than computing a result; 71 with no instrument failure: 7 honoured, 2 violated, 62 with no contract checked; 28 where Syntology's instrument failed) · 77 unverified); nothing checks that they implement this paper's method. report
susanwe/ope_worldmodels mentioned on GitHubSyntology: not harvested report
ray-project/ray Syntology: not harvested report
stavrosgreece/MultiAgentLearning pytorchGPL-3.0Syntology: not harvested 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.

A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.

Code Syntology ran Syntology

29 samples harvested: 16 ran (of which 0 constructed an object rather than computing a result; 11 with no instrument failure: 0 honoured, 1 violated, 10 with no contract checked; 5 where Syntology's instrument failed) · 13 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

1ran · violated contract
4ran · our draft was wrong
1ran; the contract check could not run
10ran
13unverified

Licence: 2 of the 29 samples are pointer only, meaning Syntology does not serve that copy's text. For 1 of them, this paper's copy is under a permissive licence and cleared, but the stored copy of the same code that Syntology serves is not; the licence label's tooltip gives Syntology's recorded reason. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 6 repositories linked to this paper, official or community; the list below is grouped by repository, each group headed by its name and what it is to the paper. 1 sample is listed last, in a group without a repository. “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. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (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 are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. 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 six 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; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. 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.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

openai/evolution-strategies-starter official repository 5 samples, 4 ran (3 instrument-failed) · 1 unverified
compute_centered_ranks in 6 papers es_distributed/es.py ran · violated contract fingerprinted MIT (permissive) · c8c2960f6687b93c · report fixture: signature x:ndarray/1/float · returned ndarray · level recorded 2026-09-22
compute_ranks in 8 papers es_distributed/es.py ran; the contract check could not run (RAISES) · metamorphic tier: invariant fingerprinted MIT (permissive) · e2d29b359d6b9487 · report fixture: signature x:ndarray/1/float · returned ndarray · level recorded 2026-09-22
make_download_and_run_script scripts/launch.py ran · our draft was wrong fingerprinted MIT (permissive) · a4a771117d6a020c · report fixture: signature code_url:str/0/str, cmd:str/0/str · returned str · level recorded 2026-09-22
make_master_script scripts/launch.py ran · our draft was wrong fingerprinted MIT (permissive) · 79c7eb37f654ff5e · report fixture: signature code_url:str/0/str, exp_str:str/0/str · returned str · level recorded 2026-09-22
make_session in 5 papers es_distributed/es.py unverified MIT (permissive) · 703807555aa85c6d · report
ShangtongZhang/DistributedES community (archive-listed) 3 samples, 0 ran · 3 unverified
test in 377 papers cma_es.py unverified Apache-2.0 (permissive) · 0f3373194a3a7f49 · report
test in 377 papers natural_es.py unverified Apache-2.0 (permissive) · d216d50a731d0266 · report
train in 522 papers natural_es.py unverified Apache-2.0 (permissive) · c481166be3fcc928 · report
atgambardella/pytorch-es community (archive-listed) 2 samples, 1 ran (1 instrument-failed) · 1 unverified
selu model.py ran · our draft was wrong fingerprinted MIT (permissive) · 372b61b06977ac30 · report fixture: signature x:Tensor/2/float · returned Tensor · level recorded 2026-09-22
normalized_columns_initializer in 8 papers model.py unverified MIT (permissive) · 2b7f271057272097 · report
czen88/qtrader community (archive-listed) 10 samples, 5 ran (0 instrument-failed) · 5 unverified
beta stats.py ran fingerprinted Apache-2.0 (permissive) · 4fb9910c23879e26 · report fixture: signature returns:ndarray/1/float, market:ndarray/1/float · returned float64 · level recorded 2026-09-26
create_lagged_series research/forecast.py ran Apache-2.0 (permissive) · b3546aca07016adb · report fixture: signature ts:DataFrame/0/?, lags:int/0/int · returned DataFrame · level recorded 2026-09-26
get_stats_mon backtest.py ran Apache-2.0 (permissive) · f31e6801b6f2ef85 · report fixture: signature ds:DataFrame/0/? · returned DataFrame · level recorded 2026-09-26
truncate in 26 papers params.py ran fingerprinted Apache-2.0 (permissive) · 9ceef2b01661079b · report fixture: signature n:float/0/float, digits:int/0/int · returned float · level recorded 2026-09-26
vol stats.py ran fingerprinted Apache-2.0 (permissive) · f6c2c3eb9800bb80 · report fixture: signature returns:ndarray/1/float · returned float64 · level recorded 2026-09-26
create_model in 163 papers research/neuralnet.py unverified Apache-2.0 (permissive) · f4c39c5bdd65e1e8 · report
gen_trades backtest.py unverified Apache-2.0 (permissive) · b23b7a9cdd5a560e · report
get_stats in 37 papers backtest.py unverified Apache-2.0 (permissive) · 9140472b1b7f5e55 · report
load_data_kr datalib.py unverified Apache-2.0 (permissive) · 063c73c605f93af1 · report
lpm stats.py unverified Apache-2.0 (permissive) · 63d526aab1237921 · report
evaboost/evaboost community (archive-listed) 6 samples, 3 ran (0 instrument-failed) · 3 unverified
IsErrorSuppressedByNolint in 5 papers 3rdparty/styleguide/cpplint/cpplint.py ran Apache-2.0 (permissive) · f411b1d1a4650685 · report fixture: signature category:str/0/str, linenum:int/0/int · returned bool · level recorded 2026-09-24
IsHeaderExtension 3rdparty/styleguide/cpplint/cpplint.py ran Apache-2.0 (permissive) · 33e5324385b943de · report fixture: signature file_extension:str/0/str · returned bool · level recorded 2026-09-26
register_check 3rdparty/styleguide/pycodestyle/pycodestyle.py ran Apache-2.0 (permissive) · 1e6800b0d86136c6 · report fixture: signature check:callable/0/?, codes:list/1/str · returned function · level recorded 2026-09-26
Match 3rdparty/styleguide/cpplint/cpplint.py unverified Apache-2.0 (permissive) here; the stored copy is pointer only · cc4a507875b5c7a1 · report
tabs_obsolete 3rdparty/styleguide/pycodestyle/pycodestyle.py unverified Apache-2.0 (permissive) · 8b07c60c858a0a5f · report
tabs_or_spaces 3rdparty/styleguide/pycodestyle/pycodestyle.py unverified Apache-2.0 (permissive) · 52c0ec49a4cf9407 · report
rasmusbergpalm/evostrat community (archive-listed) 2 samples, 2 ran (0 instrument-failed) · 0 unverified
compute_centered_ranks in 6 papers evostrat/util.py ran fingerprinted MIT (permissive) · 434d7be1e2631103 · report fixture: signature x:list/1/float · returned list · level recorded 2026-09-26
normalize in 638 papers
census: 70 implementations from 82 papers, 51 distinct outputs across 11 buckets, one shared input per bucket
evostrat/util.py ran fingerprinted MIT (permissive) · fe8d306c84b63ec5 · report fixture: signature x:list/1/float · returned list · level recorded 2026-09-26
Repository not located 1 sample, 1 ran (1 instrument-failed) · 0 unverified
1 sample is identical to code Syntology harvested from another paper's repository; this paper's own copy was not located, so no repository is named and no licence is recorded.
mkdir in 12 papers repository not located ran · our draft was wrong licence of this copy not recorded · 6b3bc12652205efd · report fixture: signature base:str/0/str, name:str/0/str · returned str · level recorded 2026-09-22

Tasks

Atari GamesMuJoCoQ-LearningReinforcement LearningReinforcement Learning (RL)reinforcement-learning

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Atari Games Atari 2600 Alien ES FF (1 hour) noop Score 994.0 #39 of 49 Archive leaderboard report
Atari Games Atari 2600 Amidar ES FF (1 hour) noop Score 112.0 #46 of 48 Archive leaderboard report
Atari Games Atari 2600 Assault ES FF (1 hour) noop Score 1673.9 #38 of 45 Archive leaderboard report
Atari Games Atari 2600 Asterix ES FF (1 hour) noop Score 1440 #45 of 49 Archive leaderboard report
Atari Games Atari 2600 Asteroids ES FF (1 hour) noop Score 1562.0 #32 of 43 Archive leaderboard report
Atari Games Atari 2600 Atlantis ES FF (1 hour) noop Score 1267410.0 #9 of 42 Archive leaderboard report
Atari Games Atari 2600 Bank Heist ES FF (1 hour) noop Score 225.0 #40 of 45 Archive leaderboard report
Atari Games Atari 2600 Battle Zone ES FF (1 hour) noop Score 16600.0 #39 of 47 Archive leaderboard report
Atari Games Atari 2600 Beam Rider ES FF (1 hour) noop Score 744.0 #47 of 49 Archive leaderboard report
Atari Games Atari 2600 Berzerk ES FF (1 hour) noop Score 686.0 #35 of 39 Archive leaderboard report
Atari Games Atari 2600 Bowling ES FF (1 hour) noop Score 30 #43 of 44 Archive leaderboard report
Atari Games Atari 2600 Boxing ES FF (1 hour) noop Score 49.8 #39 of 45 Archive leaderboard report
Atari Games Atari 2600 Breakout ES FF (1 hour) noop Score 9.5 #55 of 58 Archive leaderboard report
Atari Games Atari 2600 Centipede ES FF (1 hour) noop Score 7783.9 #22 of 45 Archive leaderboard report
Atari Games Atari 2600 Chopper Command ES FF (1 hour) noop Score 3710.0 #37 of 45 Archive leaderboard report
Atari Games Atari 2600 Crazy Climber ES FF (1 hour) noop Score 26430.0 #45 of 49 Archive leaderboard report
Atari Games Atari 2600 Demon Attack ES FF (1 hour) noop Score 1166.5 #42 of 46 Archive leaderboard report
Atari Games Atari 2600 Double Dunk ES FF (1 hour) noop Score 0.2 #21 of 43 Archive leaderboard report
Atari Games Atari 2600 Enduro ES FF (1 hour) noop Score 95.0 #41 of 48 Archive leaderboard report
Atari Games Atari 2600 Fishing Derby ES FF (1 hour) noop Score -49.0 #41 of 44 Archive leaderboard report
Atari Games Atari 2600 Freeway ES FF (1 hour) noop Score 31.0 #30 of 59 Archive leaderboard report
Atari Games Atari 2600 Frostbite ES FF (1 hour) noop Score 370.0 #41 of 53 Archive leaderboard report
Atari Games Atari 2600 Gopher ES FF (1 hour) noop Score 582.0 #43 of 43 Archive leaderboard report
Atari Games Atari 2600 Gravitar ES FF (1 hour) noop Score 805.0 #25 of 53 Archive leaderboard report
Atari Games Atari 2600 Ice Hockey ES FF (1 hour) noop Score -4.1 #34 of 43 Archive leaderboard report
Atari Games Atari 2600 Kangaroo ES FF (1 hour) noop Score 11200.0 #23 of 47 Archive leaderboard report
Atari Games Atari 2600 Krull ES FF (1 hour) noop Score 8647.2 #25 of 45 Archive leaderboard report
Atari Games Atari 2600 Name This Game ES FF (1 hour) noop Score 4503.0 #39 of 43 Archive leaderboard report
Atari Games Atari 2600 Pong ES FF (1 hour) noop Score 21.0 #2 of 52 Archive leaderboard report
Atari Games Atari 2600 Private Eye ES FF (1 hour) noop Score 100.0 #43 of 52 Archive leaderboard report
Atari Games Atari 2600 Q*Bert ES FF (1 hour) noop Score 147.5 #55 of 57 Archive leaderboard report
Atari Games Atari 2600 River Raid ES FF (1 hour) noop Score 5009.0 #35 of 41 Archive leaderboard report
Atari Games Atari 2600 Road Runner ES FF (1 hour) noop Score 16590.0 #38 of 44 Archive leaderboard report
Atari Games Atari 2600 Robotank ES FF (1 hour) noop Score 11.9 #38 of 42 Archive leaderboard report
Atari Games Atari 2600 Seaquest ES FF (1 hour) noop Score 1390.0 #45 of 57 Archive leaderboard report
Atari Games Atari 2600 Space Invaders ES FF (1 hour) noop Score 678.5 #48 of 55 Archive leaderboard report
Atari Games Atari 2600 Star Gunner ES FF (1 hour) noop Score 1470.0 #39 of 43 Archive leaderboard report
Atari Games Atari 2600 Tennis ES FF (1 hour) noop Score -4.5 #38 of 43 Archive leaderboard report
Atari Games Atari 2600 Time Pilot ES FF (1 hour) noop Score 4970.0 #36 of 44 Archive leaderboard report
Atari Games Atari 2600 Tutankham ES FF (1 hour) noop Score 130.3 #32 of 44 Archive leaderboard report
Atari Games Atari 2600 Up and Down ES FF (1 hour) noop Score 67974.0 #19 of 44 Archive leaderboard report
Atari Games Atari 2600 Venture ES FF (1 hour) noop Score 760.0 #20 of 55 Archive leaderboard report
Atari Games Atari 2600 Video Pinball ES FF (1 hour) noop Score 22834.8 #40 of 42 Archive leaderboard report
Atari Games Atari 2600 Wizard of Wor ES FF (1 hour) noop Score 3480.0 #35 of 41 Archive leaderboard report
Atari Games Atari 2600 Zaxxon ES FF (1 hour) noop Score 6380.0 #33 of 41 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

Q-Learning

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