# From Real World to Logic and Back: Learning Generalizable Relational Concepts For Long Horizon Robot Planning

arXiv: https://arxiv.org/abs/2402.11871 · HTML: https://mlsanity.com/paper/2402.11871/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 19 Feb 2024.

## Built on by (most cited first, 9 of 9)

- [World Model for Robot Learning: A Comprehensive Survey](https://mlsanity.com/paper/2605.00080/) (2026)
- [Hybrid Diffusion for Simultaneous Symbolic and Continuous Planning](https://mlsanity.com/paper/2509.21983/) (2025)
- [Autonomous Option Invention for Continual Hierarchical Reinforcement Learning and Planning](https://mlsanity.com/paper/2412.16395/) (2024)
- [Learning Lifted Action Models from Unsupervised Visual Traces](https://mlsanity.com/paper/2604.19043/) (2026)
- [Autonomous Assessment of Generalizability of AI Agent Capabilities](https://mlsanity.com/paper/2512.16733/) (2025)
- [Learning Bilevel Policies over Symbolic World Models for Long-Horizon Planning](https://mlsanity.com/paper/2605.15975/) (2026)
- [STRIPS-WM: Learning Grounded Propositional STRIPS-style World Models from Images](https://mlsanity.com/paper/2606.06832/) (2026)
- [PDDL-ART: Autonomous Symbolic Abstraction From Demonstration For Long-Horizon Robotic Manipulation Using Vision-Language Models](https://mlsanity.com/paper/2608.17146/) (2026)
- [TANDEM: Task and Motion Planning with As-Needed Demonstrations for Efficient Vision-Language-Action Model Fine-tuning](https://mlsanity.com/paper/2609.28314/) (2026)

## Abstract (arXiv)

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× 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.