Self-Supervised Behavior Cloned Transformers are Path Crawlers for Text Games

Ruoyao Wang, Peter Jansen


Abstract
In this work, we introduce a self-supervised behavior cloning transformer for text games, which are challenging benchmarks for multi-step reasoning in virtual environments. Traditionally, Behavior Cloning Transformers excel in such tasks but rely on supervised training data. Our approach auto-generates training data by exploring trajectories (defined by common macro-action sequences) that lead to reward within the games, while determining the generality and utility of these trajectories by rapidly training small models then evalauating their performance on unseen development games. Through empirical analysis, we show our method consistently uncovers generalizable training data, achieving about 90% performance of supervised systems across three benchmark text games.
Anthology ID:
2023.findings-emnlp.369
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2023
Month:
December
Year:
2023
Address:
Singapore
Editors:
Houda Bouamor, Juan Pino, Kalika Bali
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
5555–5565
Language:
URL:
https://aclanthology.org/2023.findings-emnlp.369
DOI:
10.18653/v1/2023.findings-emnlp.369
Bibkey:
Cite (ACL):
Ruoyao Wang and Peter Jansen. 2023. Self-Supervised Behavior Cloned Transformers are Path Crawlers for Text Games. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 5555–5565, Singapore. Association for Computational Linguistics.
Cite (Informal):
Self-Supervised Behavior Cloned Transformers are Path Crawlers for Text Games (Wang & Jansen, Findings 2023)
Copy Citation:
PDF:
https://aclanthology.org/2023.findings-emnlp.369.pdf