Learning Affordances at Inference-Time for Vision-Language-Action Models

TL;DR

This paper introduces LITEN, a method enabling vision-language-action models to learn from past real-world robot experiences, improving planning and task success through iterative reasoning and assessment phases.

Contribution

LITEN connects low-level policies with high-level vision-language models for dynamic, inference-time learning from unstructured robot trajectories, enhancing task planning.

Findings

LITEN improves long-horizon task success.

Effective learning from unstructured robot data.

Iterative reasoning enhances planning accuracy.

Abstract

Solving complex real-world control tasks often takes multiple tries: if we fail at first, we reflect on what went wrong, and change our strategy accordingly to avoid making the same mistake. In robotics, Vision-Language-Action models (VLAs) offer a promising path towards solving complex control tasks, but lack the ability to contextually and dynamically readjust behavior when they fail to accomplish a task. In this work, we introduce Learning from Inference-Time Execution (LITEN), which connects a VLA low-level policy to a high-level VLM that conditions on past experiences by including them in-context, allowing it to learn the affordances and capabilities of the low-level VLA. Our approach iterates between a reasoning phase that generates and executes plans for the low-level VLA, and an assessment phase that reflects on the resulting execution and draws useful conclusions to be included in future reasoning contexts. Unlike similar approaches to self-refinement in non-robotics domains, LITEN must reflect on unstructured real-world robot trajectories (e.g., raw videos), which requires structured guiderails during assessment. Our experimental results demonstrate LITEN is able to effectively learn from past experience to generate plans that use high-affordance instructions to accomplish long-horizon tasks.