Visual Affordance Prediction for Guiding Robot Exploration
This addresses the challenge of efficient robot exploration in manipulation tasks, though it appears incremental as it builds on existing Transformer and VQ-VAE methods.
The paper tackles the problem of enabling robots to understand possible interactions with objects by developing a visual affordance prediction approach that infers plausible future states from images, showing compositional generalization to diverse objects beyond training data.
Motivated by the intuitive understanding humans have about the space of possible interactions, and the ease with which they can generalize this understanding to previously unseen scenes, we develop an approach for learning visual affordances for guiding robot exploration. Given an input image of a scene, we infer a distribution over plausible future states that can be achieved via interactions with it. We use a Transformer-based model to learn a conditional distribution in the latent embedding space of a VQ-VAE and show that these models can be trained using large-scale and diverse passive data, and that the learned models exhibit compositional generalization to diverse objects beyond the training distribution. We show how the trained affordance model can be used for guiding exploration by acting as a goal-sampling distribution, during visual goal-conditioned policy learning in robotic manipulation.