Numerical Goal-based Transformers for Practical Conditions

Seonghyun Kim, Samyeul Noh, Ingook Jang

Confirmation: I have read and confirm that at least one author will be attending the workshop in person if the submission is accepted

Keywords: goal-conditioned reinforcement learning, numerical goal-conditioned transformer, conservative reward estimation

TL;DR: This work represents attempts and results for numerical goal-based transformers to operate under practical conditions.

Abstract: Goal-conditioned reinforcement learning (GCRL) studies aim to apply trained agents in realistic environments. In particular, offline reinforcement learning is being studied as a way to reduce the cost of online interactions in GCRL. One such method is Decision Transformer (DT), which utilizes a numerical goal called "return-to-go" for superior performance. Since DT assumes an idealized environment, such as perfect knowledge of rewards, it is necessary to study an improved approach for real-world applications. In this work, we present various attempts and results for numerical goal-based transformers to operate under practical conditions.

Submission Number: 5