GIFT: Generalizable Interaction-aware Functional Tool Affordances without Labels


Dylan Turpin (University of Toronto),
Liquan Wang (University of Toronto),
Stavros Tsogkas (University of Toronto),
Sven Dickinson (University of Toronto),
Animesh Garg (University of Toronto, Vector Institute, NVIDIA)
Paper Website
Paper #060
Interactive Poster Session III Interactive Poster Session VI

0d 00h 00m

0d 00h 00m


Abstract

Tool use requires reasoning about the fit between an object’s affordances and the demands of a task. Visual affordance learning can benefit from goal-directed interaction experience, but current techniques rely on human labels or expert demonstrations to generate this data. In this paper, we describe a method that grounds affordances in physical interactions instead, thus removing the need for human labels or expert policies. We use an efficient sampling-based method to generate successful trajectories that provide contact data, which are then used to reveal affordance representations. Our framework, GIFT, operates in two phases: first, we discover visual affordances from goal-directed interaction with a set of procedurally generated tools; second, we train a model to predict new instances of the discovered affordances on novel tools in a self-supervised fashion. In our experiments, we show that GIFT can leverage a sparse keypoint representation to predict grasp and interaction points to accommodate multiple tasks, such as hooking, reaching, and hammering. GIFT outperforms baselines on all tasks and matches a human oracle on two of three tasks using novel tools.

Spotlight Presentation

Previous Paper Paper Website Next Paper