video-games

Atari Games

3260 papers • 126 benchmarks • 313 datasets

The Atari 2600 Games task (and dataset) involves training an agent to achieve high game scores. ( Image credit: Playing Atari with Deep Reinforcement Learning )

(Image credit: Papersgraph)

Benchmarks

These leaderboards are used to track progress in video-games

Trend

Dataset

Best Model

Actions

Atari 2600 Freeway

Atari 2600 Breakout

Atari 2600 Q*Bert

Libraries

i

Use these libraries to find video-games models and implementations

chainer/chainerrl

11 papers 1,097

Datasets

Arcade Learning Environment

DQN Replay Dataset

Atari Grand Challenge

Atari-HEAD

AtariARI

RLU

Subtasks

Montezuma's Revenge

Most implemented papers

Playing Atari with Deep Reinforcement Learning

K. Kavukcuoglu, Daan Wierstra, David Silver, Volodymyr Mnih, Alex Graves, Ioannis Antonoglou, Martin A. Riedmiller•Wed Dec 18 2013

This work presents the first deep learning model to successfully learn control policies directly from high-dimensional sensory input using reinforcement learning, which outperforms all previous approaches on six of the games and surpasses a human expert on three of them.

13135

Content

Atari 2600 Q*Bert

Atari 2600 Seaquest

Atari 2600 Space Invaders

Atari 2600 Frostbite

Atari 2600 Pong

Atari 2600 Venture

Atari 2600 Gravitar

Atari 2600 Alien

Atari 2600 Crazy Climber

Atari 2600 Private Eye

Atari 2600 Amidar

Atari 2600 Asterix

Atari 2600 Beam Rider

Atari 2600 Enduro

Atari 2600 Battle Zone

Atari 2600 Kangaroo

Atari 2600 Montezuma's Revenge

Atari 2600 Ms. Pacman

Atari 2600 Assault

Atari 2600 Bank Heist

Atari 2600 Boxing

Atari 2600 Centipede

Atari 2600 Chopper Command

Atari 2600 Demon Attack

Atari 2600 James Bond

Atari 2600 Krull

Atari 2600 Bowling

Atari 2600 Fishing Derby

Atari 2600 HERO

Atari 2600 Road Runner

Atari 2600 Time Pilot

Atari 2600 Tutankham

Atari 2600 Asteroids

Atari 2600 Double Dunk

Atari 2600 Gopher

Atari 2600 Ice Hockey

Atari 2600 Kung-Fu Master

Atari 2600 Name This Game

Atari 2600 Tennis

Atari 2600 Up and Down

Atari 2600 Atlantis

Atari 2600 Robotank

Atari 2600 Star Gunner

Atari 2600 Video Pinball

Atari 2600 Berzerk

Atari 2600 River Raid

Atari 2600 Wizard of Wor

Atari 2600 Zaxxon

Atari 2600 Skiing

Atari 2600 Pitfall!

Atari 2600 Defender

Atari 2600 Phoenix

Atari 2600 Solaris

Atari-57

Atari 2600 Yars Revenge

Atari 2600 Surround

Atari games

atari game

Atari 2600 Elevator Action

Atari 2600 Pooyan

Atari 2600 Carnival

Atari 2600 Journey Escape

NervanaSystems/coach

7 papers 2,233

chandar-lab/RLHive

7 papers 90

michaelnny/deep_rl_zoo

7 papers 55

hill-a/stable-baselines

6 papers 3,740

tensorlayer/RLzoo

5 papers 565

DLR-RM/stable-baselines3

4 papers 5,169

toni-sm/skrl

4 papers 183

floringogianu/atari-agents

3 papers 68

atavakol/action-hypergraph-networks

3 papers 18

Sheepsody/Batched-Impala-PyTorch

3 papers 12

google/dopamine

2 papers 10,036

google-research/batch_rl

2 papers 424

arnomoonens/yarll

2 papers 84

ulstu/robotics_ml

2 papers 35

ulstu/ml

2 papers 35

robintyh1/icml2021-pengqlambda

2 papers 15

ku2482/rltorch

2 papers 14

0

Paper Graph

Deep Reinforcement Learning with Double Q-Learning

David Silver, H. V. Hasselt, A. Guez•Mon Sep 21 2015

This paper proposes a specific adaptation to the DQN algorithm and shows that the resulting algorithm not only reduces the observed overestimations, as hypothesized, but that this also leads to much better performance on several games.

8389 0

Paper Graph

Prioritized Experience Replay

David Silver, Ioannis Antonoglou, T. Schaul, John Quan•Tue Nov 17 2015

A framework for prioritizing experience, so as to replay important transitions more frequently, and therefore learn more efficiently, in Deep Q-Networks, a reinforcement learning algorithm that achieved human-level performance across many Atari games.

4113 0

Paper Graph

Dueling Network Architectures for Deep Reinforcement Learning

Nando de Freitas, H. V. Hasselt, Ziyun Wang, T. Schaul, Matteo Hessel, Marc Lanctot•Thu Nov 19 2015

This paper presents a new neural network architecture for model-free reinforcement learning that leads to better policy evaluation in the presence of many similar-valued actions and enables the RL agent to outperform the state-of-the-art on the Atari 2600 domain.

4115 0

Paper Graph

Asynchronous Methods for Deep Reinforcement Learning

K. Kavukcuoglu, David Silver, Volodymyr Mnih, Adrià Puigdomènech Badia, Mehdi Mirza, Alex Graves, T. Lillicrap, Tim Harley•Wed Feb 03 2016

A conceptually simple and lightweight framework for deep reinforcement learning that uses asynchronous gradient descent for optimization of deep neural network controllers and shows that asynchronous actor-critic succeeds on a wide variety of continuous motor control problems as well as on a new task of navigating random 3D mazes using a visual input.

9534 0

Paper Graph

Rainbow: Combining Improvements in Deep Reinforcement Learning

David Silver, H. V. Hasselt, Georg Ostrovski, T. Schaul, Matteo Hessel, Joseph Modayil, Will Dabney, Dan Horgan, Bilal Piot, M. G. Azar•Thu Oct 05 2017

This paper examines six extensions to the DQN algorithm and empirically studies their combination, showing that the combination provides state-of-the-art performance on the Atari 2600 benchmark, both in terms of data efficiency and final performance.

2459 0

Paper Graph

Trust Region Policy Optimization

P. Abbeel, S. Levine, Michael I. Jordan, John Schulman, Philipp Moritz•Wed Feb 18 2015

A method for optimizing control policies, with guaranteed monotonic improvement, by making several approximations to the theoretically-justified scheme, called Trust Region Policy Optimization (TRPO).

7595 0

Paper Graph

Evolution Strategies as a Scalable Alternative to Reinforcement Learning

I. Sutskever, Tim Salimans, Xi Chen, Jonathan Ho•Thu Mar 09 2017

This work explores 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, and highlights several advantages of ES as a blackbox optimization technique.

1677 0

Paper Graph

A Distributional Perspective on Reinforcement Learning

Marc G. Bellemare, Will Dabney, R. Munos•Sun Jul 16 2017

This paper argues for the fundamental importance of the value distribution: the distribution of the random return received by a reinforcement learning agent, and designs a new algorithm which applies Bellman's equation to the learning of approximate value distributions.

1725 0

Paper Graph

An Intriguing Failing of Convolutional Neural Networks and the CoordConv Solution

Alexander Sergeev, J. Lehman, J. Yosinski, Rosanne Liu, F. Such, Piero Molino, Eric Frank•Sat Jun 30 2018

Preliminary evidence that swapping convolution for CoordConv can improve models on a diverse set of tasks is shown, which works by giving convolution access to its own input coordinates through the use of extra coordinate channels without sacrificing the computational and parametric efficiency of ordinary convolution.

989 0

Paper Graph

Adding a benchmark result helps the community track progress.