Learning Affordances at Inference-Time for Vision-Language-Action Models
This addresses the need for robots to adaptively learn from failures in complex real-world control, though it is incremental as it builds on existing self-refinement approaches by adapting them to unstructured robot trajectories.
The paper tackles the problem of Vision-Language-Action models lacking dynamic readjustment after failures in robotics tasks by introducing LITEN, which connects a low-level policy to a high-level VLM that learns affordances from past experiences, resulting in effective generation of high-affordance instructions for long-horizon tasks.
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.