What Objects Enable, Not What They Are: Functional Latent Spaces for Affordance Reasoning

In three linesA4D introduces a functional latent space organized around object affordances rather than appearance to improve robot planning generalization. The system achieves 94% accuracy on existing affordances (+15 points vs SOTA) and 90% on new affordances using only 10% of original training data, with 100x faster inference.

Summary generated by Claude — human-verified