Adversarial Policies Beat Superhuman Go AIs

Authors: Tony T. Wang, Adam Gleave, Tom Tseng, Kellin Pelrine, Nora Belrose, Joseph Miller, Michael D. Dennis, Yawen Duan, Viktor Pogrebniak, Sergey Levine, Stuart Russell
Published: 2022-11-01 | Updated: 2023-07-13

Source: https://arxiv.org/abs/2211.00241

PDF: https://arxiv.org/pdf/2211.00241

Labels Predicted by AI

学習の改善 DNN IP保護手法 悪意のある行為者の分類

Please note that these labels were automatically added by AI. Therefore, they may not be entirely accurate.
For more details, please see the About the Literature Database page.

Abstract

We attack the state-of-the-art Go-playing AI system KataGo by training adversarial policies against it, achieving a >97 running at superhuman settings. Our adversaries do not win by playing Go well. Instead, they trick KataGo into making serious blunders. Our attack transfers zero-shot to other superhuman Go-playing AIs, and is comprehensible to the extent that human experts can implement it without algorithmic assistance to consistently beat superhuman AIs. The core vulnerability uncovered by our attack persists even in KataGo agents adversarially trained to defend against our attack. Our results demonstrate that even superhuman AI systems may harbor surprising failure modes. Example games are available https://goattack.far.ai/.