SOTAVerified

Atari Games

The Atari 2600 Games task (and dataset) involves training an agent to achieve high game scores.

( Image credit: Playing Atari with Deep Reinforcement Learning )

Papers

Recently AddedMost HypedMost ActiveNeeds VerificationMost Verified

Showing 326–350 of 625 papers

TitleStatusHype
A Convergent Variant of the Boltzmann Softmax Operator in Reinforcement Learning—0
Action Q-Transformer: Visual Explanation in Deep Reinforcement Learning with Encoder-Decoder Model using Action Query—0
Activation by Interval-wise Dropout: A Simple Way to Prevent Neural Networks from Plasticity Loss—0
Adapting Auxiliary Losses Using Gradient Similarity—0
Adapting Behaviour for Learning Progress—0
Adaptive N-step Bootstrapping with Off-policy Data—0
Adaptive Q-Network: On-the-fly Target Selection for Deep Reinforcement Learning—0
Addressing Action Oscillations through Learning Policy Inertia—0
A Decentralized Policy Gradient Approach to Multi-task Reinforcement Learning—0
A Deep Learning Approach for Joint Video Frame and Reward Prediction in Atari Games—0
Adventurer: Exploration with BiGAN for Deep Reinforcement Learning—0
Adversary A3C for Robust Reinforcement Learning—0
Agent-Aware Training for Agent-Agnostic Action Advising in Deep Reinforcement Learning—0
AGIL: Learning Attention from Human for Visuomotor Tasks—0
A Human Mixed Strategy Approach to Deep Reinforcement Learning—0
A Mathematical Theory of Deep Convolutional Neural Networks for Feature Extraction—0
An advantage actor-critic algorithm for robotic motion planning in dense and dynamic scenarios—0
Analysing Results from AI Benchmarks: Key Indicators and How to Obtain Them—0
Analysis of Q-learning with Adaptation and Momentum Restart for Gradient Descent—0
An Approach to Partial Observability in Games: Learning to Both Act and Observe—0
An Entropy Regularization Free Mechanism for Policy-based Reinforcement Learning—0
A new Potential-Based Reward Shaping for Reinforcement Learning Agent—0
An initial attempt of combining visual selective attention with deep reinforcement learning—0
APF+: Boosting adaptive-potential function reinforcement learning methods with a W-shaped network for high-dimensional games—0
Approximate Shielding of Atari Agents for Safe Exploration—0
Show:102550
← PrevPage 14 of 25Next →
All datasetsAtari 2600 BreakoutAtari 2600 FreewayAtari 2600 Q*BertAtari 2600 SeaquestAtari 2600 VentureAtari 2600 Space InvadersAtari 2600 FrostbiteAtari 2600 GravitarAtari 2600 PongAtari 2600 Private EyeAtari 2600 Montezuma's RevengeAtari 2600 Alien

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1GDI-H3(200M frames)Score864—Unverified
2GDI-I3(200M frames)Score864—Unverified
3GDI-I3Score864—Unverified
4GDI-H3Score864—Unverified
5Bootstrapped DQNScore855—Unverified
6FQFScore854.2—Unverified
7R2D2Score837.7—Unverified
8Ape-XScore800.9—Unverified
9Agent57Score790.4—Unverified
10IMPALA (deep)Score787.34—Unverified
#ModelMetricClaimedVerifiedStatus
1GDI-H3(200M frames)Score34—Unverified
2TRPO-hashScore34—Unverified
3IQNScore34—Unverified
4NoisyNet-DuelingScore34—Unverified
5QR-DQN-1Score34—Unverified
6Go-ExploreScore34—Unverified
7GDI-I3Score34—Unverified
8GDI-H3Score34—Unverified
9ASL DDQNScore33.9—Unverified
10Bootstrapped DQNScore33.9—Unverified
#ModelMetricClaimedVerifiedStatus
1Agent57Score580,328.14—Unverified
2QR-DQN-1Score572,510—Unverified
3R2D2Score408,850—Unverified
4IMPALA (deep)Score351,200.12—Unverified
5Ape-XScore302,391.3—Unverified
6A2C + SILScore104,975.6—Unverified
7MuZero (Res2 Adam)Score94,906.25—Unverified
8DreamerV2Score94,688—Unverified
9MuZeroScore72,276—Unverified
10DNAScore52,398—Unverified
#ModelMetricClaimedVerifiedStatus
1GDI-H3(200M frames)Score1,000,000—Unverified
2GDI-H3Score1,000,000—Unverified
3Agent57Score999,997.63—Unverified
4R2D2Score999,996.7—Unverified
5MuZeroScore999,976.52—Unverified
6MuZero (Res2 Adam)Score999,659.18—Unverified
7GDI-I3Score943,910—Unverified
8Ape-XScore392,952.3—Unverified
9C51 noopScore266,434—Unverified
10Duel noopScore50,254.2—Unverified