Learning Hidden Subgoals under Temporal Ordering Constraints in Reinforcement Learning

TL;DR

This paper introduces LSTOC, a reinforcement learning algorithm that learns hidden subgoals and their temporal orderings using contrastive learning, enhancing efficiency and generalization in complex tasks with temporal constraints.

Contribution

LSTOC is the first method to simultaneously learn hidden subgoals and their temporal orderings in RL using a contrastive learning approach and subgoal trees.

Findings

LSTOC outperforms baseline methods in various environments.

It improves sample efficiency in discovering subgoals.

It generalizes well to unseen tasks.

Abstract

In real-world applications, the success of completing a task is often determined by multiple key steps which are distant in time steps and have to be achieved in a fixed time order. For example, the key steps listed on the cooking recipe should be achieved one-by-one in the right time order. These key steps can be regarded as subgoals of the task and their time orderings are described as temporal ordering constraints. However, in many real-world problems, subgoals or key states are often hidden in the state space and their temporal ordering constraints are also unknown, which make it challenging for previous RL algorithms to solve this kind of tasks. In order to address this issue, in this work we propose a novel RL algorithm for {\bf l}earning hidden {\bf s}ubgoals under {\bf t}emporal {\bf o}rdering {\bf c}onstraints (LSTOC). We propose a new contrastive learning objective which can effectively learn hidden subgoals (key states) and their temporal orderings at the same time, based on first-occupancy representation and temporal geometric sampling. In addition, we propose a sample-efficient learning strategy to discover subgoals one-by-one following their temporal order constraints by building a subgoal tree to represent discovered subgoals and their temporal ordering relationships. Specifically, this tree can be used to improve the sample efficiency of trajectory collection, fasten the task solving and generalize to unseen tasks. The LSTOC framework is evaluated on several environments with image-based observations, showing its significant improvement over baseline methods.