Browse State-of-the-Art › Continuous Control

Continuous Control

494 papers with code · 73 benchmarks · 10 datasets archive 2025-07-28

Computer VisionPlaying GamesRobots

Continuous control in the context of playing games, especially within artificial intelligence (AI) and machine learning (ML), refers to the ability to make a series of smooth, ongoing adjustments or actions to control a game or a simulation. This is in contrast to discrete control, where the actions are limited to a set of specific, distinct choices. Continuous control is crucial in environments where precision, timing, and the magnitude of actions matter, such as driving a car in a racing game, controlling a character in a simulation, or managing the flight of an aircraft in a flight simulator.

Description from the archive archive 2025-07-28.

Benchmarks archive 2025-07-28

73 leaderboard tables shown for this task, 73 with rows (a “benchmark” on this site is a table with at least one row, as on /sota), ordered by row count. “Best model” is the first row in the archive's own order at snapshot; nothing is re-ranked here and metric direction is not recorded in the archive. PwC's Trend sparklines are not in the archive, so that column is omitted. 10 shown of 73 until expanded.

Syntology 9 newer papers were judged to report on this task's benchmarks so far (Claude Sonnet 4.5, reading their tables: a model's judgement, not a result; 0 placed): cartpole.balance_sparse (1 judged, 0 placed), cartpole.swingup (1 judged, 0 placed), cheetah.run (3 judged, 0 placed), finger.turn_hard (1 judged, 0 placed), walker.stand (2 judged, 0 placed), walker.walk (2 judged, 0 placed), Acrobot (1 judged, 0 placed), Ant (2 judged, 0 placed), Hopper (3 judged, 0 placed), humanoid.run (2 judged, 0 placed), Mountain Car (1 judged, 0 placed), quadruped.run (1 judged, 0 placed), quadruped.walk (2 judged, 0 placed), reacher.easy (1 judged, 0 placed), reacher.hard (1 judged, 0 placed), walker.run (1 judged, 0 placed). Drawn from 8,886 of 18,637 newer papers checked so far (733 more had nothing to place; 41 could not be read).

DatasetBest model (first row in archive order)PaperCodeSyntologyCompare
PyBullet Ant (8 rows) SAC gSDE Smooth Exploration for Robotic Reinforcement Learning code Syntology community repositories only · 1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified Compare
PyBullet HalfCheetah (8 rows) SAC Smooth Exploration for Robotic Reinforcement Learning code Syntology community repositories only · 1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified Compare
PyBullet Hopper (8 rows) SAC gSDE Smooth Exploration for Robotic Reinforcement Learning code Syntology community repositories only · 1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified Compare
PyBullet Walker2D (8 rows) SAC gSDE Smooth Exploration for Robotic Reinforcement Learning code Syntology community repositories only · 1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified Compare
Lunar Lander (OpenAI Gym) (5 rows) SAC Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement... code Syntology community repositories only · 91 ran (of which 61 constructed an object rather than computing a result; 73 with no instrument failure: 3 honoured, 1 violated, 69 with no contract checked; 18 where Syntology's instrument failed) · 57 unverified Compare
DeepMind Cheetah Run (Images) (4 rows) DreamerV1 — — — Compare
cartpole.balance_sparse (2 rows) SMuZero Learning and Planning in Complex Action Spaces code — Compare
cartpole.swingup (2 rows) SMuZero Learning and Planning in Complex Action Spaces code — Compare
cheetah.run (2 rows) SMuZero Learning and Planning in Complex Action Spaces code — Compare
DeepMind Cup Catch (Images) (2 rows) DrQ Image Augmentation Is All You Need: Regularizing Deep... code Syntology official: harvested, nothing ran · 6 ran (of which 3 constructed an object rather than computing a result; 5 with no instrument failure: 1 honoured, 1 violated, 3 with no contract checked; 1 where Syntology's instrument failed) · 4 unverified Compare
DeepMind Walker Walk (Images) (2 rows) DrQ Image Augmentation Is All You Need: Regularizing Deep... code Syntology official: harvested, nothing ran · 6 ran (of which 3 constructed an object rather than computing a result; 5 with no instrument failure: 1 honoured, 1 violated, 3 with no contract checked; 1 where Syntology's instrument failed) · 4 unverified Compare
finger.turn_hard (2 rows) SMuZero Learning and Planning in Complex Action Spaces code — Compare
walker.stand (2 rows) SMuZero Learning and Planning in Complex Action Spaces code — Compare
walker.walk (2 rows) SMuZero Learning and Planning in Complex Action Spaces code — Compare
2D Walker (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Acrobot (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Acrobot (limited sensors) (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Acrobot (noisy observations) (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
acrobot.swingup (1 row) SMuZero Learning and Planning in Complex Action Spaces code — Compare
Acrobot (system identifications) (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Ant (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Ant + Gathering (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Ant + Maze (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Ball in cup, catch (DMControl500k) (1 row) CURL CURL: Contrastive Unsupervised Representations for Reinforcement Learning code Syntology official (archive's flag): 6 ran · 7 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 5 where Syntology's instrument failed) · 1 unverified Compare
Ball in cup, catch (DMControl100k) (1 row) CURL CURL: Contrastive Unsupervised Representations for Reinforcement Learning code Syntology official (archive's flag): 6 ran · 7 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 5 where Syntology's instrument failed) · 1 unverified Compare
ball_in_cup.catch (1 row) SMuZero Learning and Planning in Complex Action Spaces code — Compare
Cart-Pole Balancing (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Cart-Pole Balancing (limited sensors) (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Cart-Pole Balancing (noisy observations) (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Cart-Pole Balancing (system identifications) (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Cart Pole (OpenAI Gym) (1 row) MAC Mean Actor Critic code — Compare
cartpole.balance (1 row) SMuZero Learning and Planning in Complex Action Spaces code — Compare
Cartpole, swingup (DMControl500k) (1 row) CURL CURL: Contrastive Unsupervised Representations for Reinforcement Learning code Syntology official (archive's flag): 6 ran · 7 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 5 where Syntology's instrument failed) · 1 unverified Compare
Cartpole, swingup (DMControl100k) (1 row) CURL CURL: Contrastive Unsupervised Representations for Reinforcement Learning code Syntology official (archive's flag): 6 ran · 7 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 5 where Syntology's instrument failed) · 1 unverified Compare
cartpole.swingup_sparse (1 row) SMuZero Learning and Planning in Complex Action Spaces code — Compare
Cheetah, run (DMControl500k) (1 row) CURL CURL: Contrastive Unsupervised Representations for Reinforcement Learning code Syntology official (archive's flag): 6 ran · 7 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 5 where Syntology's instrument failed) · 1 unverified Compare
Cheetah, run (DMControl100k) (1 row) CURL CURL: Contrastive Unsupervised Representations for Reinforcement Learning code Syntology official (archive's flag): 6 ran · 7 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 5 where Syntology's instrument failed) · 1 unverified Compare
Double Inverted Pendulum (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Finger, spin (DMControl500k) (1 row) CURL CURL: Contrastive Unsupervised Representations for Reinforcement Learning code Syntology official (archive's flag): 6 ran · 7 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 5 where Syntology's instrument failed) · 1 unverified Compare
Finger, spin (DMControl100k) (1 row) CURL CURL: Contrastive Unsupervised Representations for Reinforcement Learning code Syntology official (archive's flag): 6 ran · 7 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 5 where Syntology's instrument failed) · 1 unverified Compare
finger.spin (1 row) SMuZero Learning and Planning in Complex Action Spaces code — Compare
finger.turn_easy (1 row) SMuZero Learning and Planning in Complex Action Spaces code — Compare
fish.swim (1 row) MuZero Unplugged Online and Offline Reinforcement Learning by Planning with a Learned Model code — Compare
Full Humanoid (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Half-Cheetah (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Hopper (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
hopper.hop (1 row) SMuZero Learning and Planning in Complex Action Spaces code — Compare
hopper.stand (1 row) SMuZero Learning and Planning in Complex Action Spaces code — Compare
humanoid.run (1 row) MuZero Unplugged Online and Offline Reinforcement Learning by Planning with a Learned Model code — Compare
Inverted Pendulum (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Inverted Pendulum (limited sensors) (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Inverted Pendulum (system identifications) (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Inverted Pendulum (noisy observations) (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
manipulator.insert_ball (1 row) MuZero Unplugged Online and Offline Reinforcement Learning by Planning with a Learned Model code — Compare
manipulator.insert_peg (1 row) MuZero Unplugged Online and Offline Reinforcement Learning by Planning with a Learned Model code — Compare
Mountain Car (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Mountain Car (limited sensors) (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Mountain Car (noisy observations) (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Mountain Car (system identifications) (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
pendulum.swingup (1 row) SMuZero Learning and Planning in Complex Action Spaces code — Compare
quadruped.run (1 row) SMuZero Learning and Planning in Complex Action Spaces code — Compare
quadruped.walk (1 row) SMuZero Learning and Planning in Complex Action Spaces code — Compare
Reacher, easy (DMControl500k) (1 row) CURL CURL: Contrastive Unsupervised Representations for Reinforcement Learning code Syntology official (archive's flag): 6 ran · 7 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 5 where Syntology's instrument failed) · 1 unverified Compare
Reacher, easy (DMControl100k) (1 row) CURL CURL: Contrastive Unsupervised Representations for Reinforcement Learning code Syntology official (archive's flag): 6 ran · 7 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 5 where Syntology's instrument failed) · 1 unverified Compare
reacher.easy (1 row) SMuZero Learning and Planning in Complex Action Spaces code — Compare
reacher.hard (1 row) SMuZero Learning and Planning in Complex Action Spaces code — Compare
Simple Humanoid (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Swimmer (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Swimmer + Gathering (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Swimmer + Maze (1 row) TRPO Benchmarking Deep Reinforcement Learning for Continuous Control code Syntology community repositories only · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
walker.run (1 row) SMuZero Learning and Planning in Complex Action Spaces code — Compare
Walker, walk (DMControl500k) (1 row) CURL CURL: Contrastive Unsupervised Representations for Reinforcement Learning code Syntology official (archive's flag): 6 ran · 7 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 5 where Syntology's instrument failed) · 1 unverified Compare
Walker, walk (DMControl100k) (1 row) CURL CURL: Contrastive Unsupervised Representations for Reinforcement Learning code Syntology official (archive's flag): 6 ran · 7 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 5 where Syntology's instrument failed) · 1 unverified Compare

Syntology column: samples harvested from the paper's repositories and executed on synthesized fixtures; “ran” is not a correctness claim and does not order the table. Each line reads “N ran (of which C constructed an object rather than computing a result; K with no instrument failure: H honoured, V violated, P with no contract checked; I where Syntology's instrument failed) · U unverified”; the instrument figure counts failures of Syntology's instrument, not of the code. When the archive marks a repository official for the row's paper, the line starts with that repository's state (“official” is the archive's flag, not a verdict on who wrote the code; “community repositories only” when every sample that ran came from a community repository, “official: no sample here; runs from other or unrecorded repositories” when some came from a repository the paper names or has in its text, or from none recorded); hover it for the repositories the samples that ran came from. A dash means no Syntology record for that paper, not a recorded non-run. Each “ran” line links to that paper's sample list. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. For agents: get_harvested_code_for_paper(arxiv_id) lists each paper's samples; how to connect.

Libraries

Not in the archive: the export carries no per-task library table, so there is nothing to show at snapshot 2025-07-28.

Datasets archive 2025-07-28

10 datasets whose archive record lists this task, ordered by the archive's paper count.

Subtasks archive 2025-07-28

3 subtasks in the archive's task tree.

Parent tasks archive 2025-07-28

Papers archive 2025-07-28

archive papers tagged: 1,161 · with a code link: 494 · where Syntology ran a sample: 193 (158 with a run with no instrument failure, 35 where every run was a failure of Syntology's instrument) Syntology

Show: all tagged papersonly where code ran (193 of 1,161 tagged: 158 with a run with no instrument failure, 35 where every run was a failure of Syntology's instrument)

Page 1 of 12: papers 1 to 100 of 1,161, in archive order: by repositories listed in the archive (most first), then newest first, not by stars (the archive holds no stars, so PwC's “Social” and “Latest” sorts cannot be reproduced). Papers that list no repository come after every paper that lists one.

Papers without a page here are shown as plain text. A Syntology line reads “N ran (of which C constructed an object rather than computing a result; K with no instrument failure: H honoured, V violated, P with no contract checked; I where Syntology's instrument failed) · U unverified”; the figure “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. When the archive marks a repository official for the paper, the line starts with that repository's state (the archive's flag, not a verdict on who wrote the code); hover it for the repositories the samples that ran came from. The first sentence of the abstract is shown for the first 30 papers; every abstract is on its paper's page.

Syntology lines on 60 of the papers shown; no Syntology record for the others (a paper without an arXiv id cannot be joined to the graph, and absence from the graph layer is not a recorded non-run). “Ran” means the sample executed on a synthesized fixture, not that the paper's result was reproduced; each line links to that paper's sample list. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. For agents: get_harvested_code_for_paper(arxiv_id) lists each paper's samples; how to connect.

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