Robust Autonomy Emerges from Self-Play

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Cusumano-Towner, Marco, Hafner, David, Hertzberg, Alex, Huval, Brody, Petrenko, Aleksei, Vinitsky, Eugene, Wijmans, Erik, Killian, Taylor, Bowers, Stuart, Sener, Ozan, Krähenbühl, Philipp, Koltun, Vladlen
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866913679544418304
author Cusumano-Towner, Marco
Hafner, David
Hertzberg, Alex
Huval, Brody
Petrenko, Aleksei
Vinitsky, Eugene
Wijmans, Erik
Killian, Taylor
Bowers, Stuart
Sener, Ozan
Krähenbühl, Philipp
Koltun, Vladlen
author_facet Cusumano-Towner, Marco
Hafner, David
Hertzberg, Alex
Huval, Brody
Petrenko, Aleksei
Vinitsky, Eugene
Wijmans, Erik
Killian, Taylor
Bowers, Stuart
Sener, Ozan
Krähenbühl, Philipp
Koltun, Vladlen
contents Self-play has powered breakthroughs in two-player and multi-player games. Here we show that self-play is a surprisingly effective strategy in another domain. We show that robust and naturalistic driving emerges entirely from self-play in simulation at unprecedented scale -- 1.6~billion~km of driving. This is enabled by Gigaflow, a batched simulator that can synthesize and train on 42 years of subjective driving experience per hour on a single 8-GPU node. The resulting policy achieves state-of-the-art performance on three independent autonomous driving benchmarks. The policy outperforms the prior state of the art when tested on recorded real-world scenarios, amidst human drivers, without ever seeing human data during training. The policy is realistic when assessed against human references and achieves unprecedented robustness, averaging 17.5 years of continuous driving between incidents in simulation.
format Preprint
id arxiv_https___arxiv_org_abs_2502_03349
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Robust Autonomy Emerges from Self-Play
Cusumano-Towner, Marco
Hafner, David
Hertzberg, Alex
Huval, Brody
Petrenko, Aleksei
Vinitsky, Eugene
Wijmans, Erik
Killian, Taylor
Bowers, Stuart
Sener, Ozan
Krähenbühl, Philipp
Koltun, Vladlen
Machine Learning
Artificial Intelligence
Robotics
Self-play has powered breakthroughs in two-player and multi-player games. Here we show that self-play is a surprisingly effective strategy in another domain. We show that robust and naturalistic driving emerges entirely from self-play in simulation at unprecedented scale -- 1.6~billion~km of driving. This is enabled by Gigaflow, a batched simulator that can synthesize and train on 42 years of subjective driving experience per hour on a single 8-GPU node. The resulting policy achieves state-of-the-art performance on three independent autonomous driving benchmarks. The policy outperforms the prior state of the art when tested on recorded real-world scenarios, amidst human drivers, without ever seeing human data during training. The policy is realistic when assessed against human references and achieves unprecedented robustness, averaging 17.5 years of continuous driving between incidents in simulation.
title Robust Autonomy Emerges from Self-Play
topic Machine Learning
Artificial Intelligence
Robotics
url https://arxiv.org/abs/2502.03349