From Real World to Logic and Back: Learning Generalizable Relational Concepts For Long Horizon Robot Planning
This addresses the challenge of interpretable, scalable, and generalizable robot planning systems for robotics, representing a novel method rather than an incremental improvement.
The paper tackles the problem of robots generalizing from limited experience to long-horizon tasks in unseen environments by enabling them to autonomously learn symbolic, relational concepts from raw demonstrations, achieving performance on par with hand-engineered models and scaling to execution horizons far beyond training while handling up to 18 times more objects than seen during learning.
Robots still lag behind humans in their ability to generalize from limited experience, particularly when transferring learned behaviors to long-horizon tasks in unseen environments. We present the first method that enables robots to autonomously invent symbolic, relational concepts directly from a small number of raw, unsegmented, and unannotated demonstrations. From these, the robot learns logic-based world models that support zero-shot generalization to tasks of far greater complexity than those in training. Our framework achieves performance on par with hand-engineered symbolic models, while scaling to execution horizons far beyond training and handling up to 18$\times$ more objects than seen during learning. The results demonstrate a framework for autonomously acquiring transferable symbolic abstractions from raw robot experience, contributing toward the development of interpretable, scalable, and generalizable robot planning systems. Project website and code: https://aair-lab.github.io/r2l-lamp.