Proto: Program-Guided Transformer for Program-Guided Tasks

Authors

Zelin Zhao, The Chinese University of Hong Kong
Karan Samel, The Chinese University of Hong KongFollow
Binghong Chen, Georgia Tech, GeorgiaFollow
Le Song, Biomap & Mohamed bin Zayed University of Artificial IntelligenceFollow

Document Type

Article

Publication Title

arXiv

Abstract

Programs, consisting of semantic and structural information, play an important role in the communication between humans and agents. Towards learning general program executors to unify perception, reasoning, and decision making, we formulate program-guided tasks which require learning to execute a given program on the observed task specification. Furthermore, we propose Program-Guided Transformer (ProTo), which integrates both semantic and structural guidance of a program by leveraging cross-attention and masked self-attention to pass messages between the specification and routines in the program. ProTo executes a program in a learned latent space and enjoys stronger representation ability than previous neural-symbolic approaches. We demonstrate that ProTo significantly outperforms the previous state-of-the-art methods on GQA visual reasoning and 2D Minecraft policy learning datasets. Additionally, ProTo demonstrates better generalization to unseen, complex, and human-written programs. © 2021, CC BY.

DOI

10.48550/arXiv.2110.00804

Publication Date

10-16-2022

Comments

Preprints: arXiv

Recommended Citation

Z. Zhao, K. Samel, B. Chen, L. Song, "ProTo: Program-guided transformer for program-guided tasks," 2021, arXiv:2110.00804