Self-Play RL

RL: Reinforcement Learning

Follow this topic Ask about this topic

Momentum

6 papers in the last four weeks, up 50% on the four weeks before. 0.1% of all new papers.

Jul 13Week of Sep 28

Latest papers 53

All topics

Faynt: Scaling and Optimizing Policies for Competitive Melee
Oct 1, 2026Ali Janati, Nikita Kuzmin, Rohit Swamy +1Reinforcement Learning Self-Play RL
Frisson Labs
Regularized policy gradient with learned mixtures of Gaussians for games with continuous actions
Sep 29, 2026Ondřej Kubíček, Viliam Lisý, Tuomas SandholmGame-Playing Agents Self-Play RL
Czech Technical University in Prague · Carnegie Mellon University · Artificial Intelligence Center
Learning to Optimize through Solver-Grounded Self-Play
Sep 28, 2026Xia Jiang, Yaoxin Wu, Chenyu Zhou +3Self-Play RL Natural Language to Optimization Modeling
Eindhoven University of Technology · Shanghai Jiaotong University
Steering Equilibrium Selection in Regularized Self-Play via the Reference Policy
Sep 17, 2026Luis LealNash Equilibrium Self-Play RL
The Surprising Effectiveness of Approximate Value Iteration in Self-Play
Sep 8, 2026Raphael Boige, Amine Boumaza, Bruno ScherrerGame-Playing Agents Self-Play RL
Université de Lorraine, CNRS, Inria, LORIA, F-54000 Nancy, France
SQL-Zero: Self-Evolving Text-to-SQL
Sep 7, 2026Daniel Machado Pedrozo, Julia Soares Dollis, Bryan Lincoln Marques de Oliveira +3Reinforcement Learning Self-Play RL
Instituto de Informática, Universidade Federal de Goiás Goiânia, GO, Brazil
Local Updates, Global Learning (LUGL): Playing Games with non-incremental Learners
Sep 3, 2026David Milec, Spyridon Samothrakis, Michael Fairbank +1Gradient-Boosted Decision Trees Game-Playing Agents
Faculty of Electrical Engineering, Czech Technical University in Prague, Prague, Czech Republic · School of Computer Science and Electronic Engineering, University of Essex, Colchester, United Kingdom · Maastricht University, the Netherlands
What Emerges and What Breaks in Self-Play Driving
Aug 31, 2026Laur Sisask, Ardi Tampuu, Tambet MatiisenReward Hacking Self-Play RL
University of Tartu, Tartu, Estonia
PokaiTrainer: Scaling Equilibrium Search to Competitive Pokémon VGC
Aug 29, 2026Max YuGame-Playing Agents Self-Play RL
Independent Researcher
SPADE: Self-Play in Adaptive Synthetic Executable Environments
Aug 19, 2026Bo Liu, Simon Yu, Yiding Jiang +15Self-Play RL LLM Agent Self-Improvement
University of Washington · Stanford University · Northeastern University +6
IB-RL: Isolated Bilateral Reinforcement Learning for Strategic Dialogue Agents
Aug 7, 2026Senhao Wang, Chenghao Cai, Haitao Hu +4Automated Negotiation Self-Play RL
Dcar Inc. ByteDance
Self-Play Meets Skill Evolution: Self-Evolving Search Agents that Pose, Solve, and Remember
Jul 31, 2026Zenghuang Fu, Zhaoyang Li, Qiuyuan Ai +6Self-Play RL Agent Skill Learning
University of Chinese Academy of Sciences · Institute of Automation, Chinese Academy of Sciences · Peking University +4
Pictura: Perspective-View Self-Play at Scale for Driving
Jul 28, 2026Yuan Yin, Elias Ramzi, Marc Lafon +8Self-Play RL RL for Autonomous Driving
valeo.ai, Paris, France · Valeo Brain, Créteil, France · Sorbonne Université, CNRS, ISIR, F-75005 Paris, France
From RLVR to RLSVR: Task Transformation Induces Self-Verifiable Rewards for Open-Ended LLM Self-Improvement
Jul 26, 2026Qinsi Wang, Jing Shi, Huazheng Wang +8Self-Play RL Reinforcement Learning with Verifiable Rewards
Duke University · Adobe Inc. · Oregon State University +3
TerraZero: Procedural Driving Simulation for Zero-Demonstration Self-Play at Scale
Jul 14, 2026Zhouchonghao Wu, Akshay Rangesh, Weixin Li +4Self-Play RL RL for Autonomous Driving
1Applied Intuition · University of California, Berkeley
World Models as Adversaries: Multi-Agent Self-Play Fine-Tuning for Robust Motion Planning
Jul 12, 2026Tong Nie, Yuewen Mei, Junlin He +3Adversarial Training Self-Play RL
The Hong Kong Polytechnic University · Tongji University · McGill University +1
AlphaZero in Sparsely Rewarded Games: Limits and Auxiliary Supervision
Jul 9, 2026Brent Kong, Tejas Ram, Tony Yue YuGame-Playing Agents Self-Play RL
Department of Mathematics · California Institute of Technology · Department of Computer Science
A Gold-Standard Study of What Makes a Lightweight Game-Playing Agent Strong
Jul 7, 2026Nima Kelidari, Mohammadsaeed Haghi, Mahdi SalmaniReinforcement Learning Game-Playing Agents
University of Southern California
Anchored Self-Play for Code Repair
Jul 3, 2026Caroline Choi, Zeyneb Kaya, Shirley Wu +3RL for Code Generation Automated Program Repair
Department of Computer Science, Stanford University, Stanford, CA, USA.
How Much Due Diligence Before You Bid? Learning in Intractable Takeover Auctions
Jun 28, 2026Zain NaboulsiSelf-Play RL Imperfect-Information Games
Principal AI Engineer, Sparq
EMAgnet: Parameter-Space EMA Regularization for Policy Gradient Self-Play in Large Games
Jun 22, 2026Tristan Maidment, JB Lanier, Chase McDonald +5Self-Play RL Imperfect-Information Games
1Riot Games · 3New York University · University of California, Irvine
Superhuman AI for Generals.io Using Self-Play Reinforcement Learning
Jun 22, 2026Matej Straka, Viliam Lisý, Martin SchmidGame-Playing Agents Self-Play RL
Department of Applied Mathematics, Charles University, Prague, Czech Republic · Department of Computer Science, Faculty of Electrical Engineering, Czech Technical University in Prague, Czech Republic · Charles University +1
GARIP: A Running-Average Moving Reference for Last-Iterate Self-Play in Two-Player Zero-Sum Games
Jun 21, 2026Can SavcıNash Equilibrium Last-Iterate Convergence
Independent Researcher İzmir, Türkiye
Scaling Self-Play for End-to-End Driving
Jun 17, 2026Luke Rowe, Roger Girgis, Rodrigue de Schaetzen +6Self-Play RL End-to-End Autonomous Driving
Mila · Université de Montréal · Polytechnique Montréal +4
Human-like autonomy emerges from self-play and a pinch of human data
Jun 11, 2026Daphne Cornelisse, Julian Hunt, Zixu Zhang +4Self-Play RL Autonomous Driving
1NYU Tandon School of Engineering · 2NYU Courant · 3Princeton University +2
CoPark: Learning Reactive Parking via Self-Play
Jun 2, 2026Jiarong Wei, Yanxing Chen, Sinuo Song +3Multi-Agent Coordination Self-Play RL
Department of Computer Science, University of Freiburg, Germany · CARIAD SE, Germany · Technical University of Munich, Germany
S-SPPO: Semantic-Calibrated Self-Play Preference Optimization
Jun 1, 2026Xiwen Chen, Wenhui Zhu, Jingjing Wang +13LLM Alignment Self-Play RL
Morgan Stanley, USA · Arizona State University, USA · Clemson University, USA +5
SCOPE: Self-Play via Co-Evolving Policies for Open-Ended Tasks
May 29, 2026Wai-Chung Kwan, Aryo Pradipta Gema, Joshua Ong Jun Leang +1Open-Ended Generation Self-Play RL
University of Edinburgh · Imperial College London
Self-Play Reinforcement Learning under Imperfect Information in Big 2
May 21, 2026Aalok PatwaGame-Playing Agents Self-Play RL
University of Pennsylvania
Survive or Collapse: The Asymmetric Roles of Data Gating and Reward Grounding in Self-Play RL
May 21, 2026Sophia Xiao Pu, Zhaotian Weng, Chengzhi Liu +4Reinforcement Learning Self-Play RL
University of California, Santa Barbara · Cisco Research

Oct 1, 2026, cs.LGJ/K move · Enter open · S save

Faynt: Scaling and Optimizing Policies for Competitive Melee

Ali Janati, Nikita Kuzmin, Rohit Swamy, Charles Niu

Frisson Labs

Reinforcement Learning Self-Play RL Transformer Imitation Learning

Open paper AskPDF arXiv

We introduce Faynt, a family of 10M- and 75M-parameter Transformer policies for Super Smash Bros. Melee, each controlling all 26 characters with a single checkpoint. After reinforcement learning (RL), the 10M wins 240 of 244 same-character games (98.4%) against fourteen specialist and multi-character releases on their supported rosters, with a winning record against every release. These opponents retain 21- or 24-frame action delays; Faynt uses no added delay, and we have not isolated the effect of this difference. In a separate evaluation against a privately supplied zero-delay Slippi-AI model, the 10M wins all 68 games across two conditioning settings. We study architecture, optimization, scaling, and hyperparameter transfer to guide pretraining on approximately 840,000 human replays. Post-training combines rank- and outcome-based curricula, 75M-to-10M distillation, and RL restricted to Fox mirror matches. On the initial 152-game benchmark, the supervised 10M wins 69.7% of games, compared with 45.4% for the pretrained 75M, despite higher overall held-out controller-prediction loss. The weighted validation loss used for supervised checkpoint selection agrees with the win-rate ordering of all four pretrained and supervised policies. After supervised post-training, both models take less damage per minute, build larger early leads, and win more often after losing the first life. Optimized inference on recorded game states averages 5.2 ms per decision for the 10M and 8.7 ms for the 75M on an NVIDIA T4, excluding emulator execution and communication. We open-source the weights, both benchmark suites, and a platform for automated model tournaments.

Momentum

Related topics

Latest papers 53