Neural Constraint Satisfaction: Hierarchical Abstraction for Combinatorial Generalization in Object Rearrangement
원문 초록
Object rearrangement is a challenge for embodied agents because solving these tasks requires generalizing across a combinatorially large set of configurations of entities and their locations. Worse, the representations of these entities are unknown and must be inferred from sensory percepts. We present a hierarchical abstraction approach to uncover these underlying entities and achieve combinatorial generalization from unstructured visual inputs. By constructing a factorized transition graph over clusters of entity representations inferred from pixels, we show how to learn a correspondence between intervening on states of entities in the agent's model and acting on objects in the environment. We use this correspondence to develop a method for control that generalizes to different numbers and configurations of objects, which outperforms current offline deep RL methods when evaluated on simulated rearrangement tasks.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
PARSEC: Preference Adaptation for Robotic Object Rearrangement from Scene Context
ManiSkill-HAB: A Benchmark for Low-Level Manipulation in Home Rearrangement Tasks
3DGS-CD: 3D Gaussian Splatting-based Change Detection for Physical Object Rearrangement
CORE4D: A 4D Human-Object-Human Interaction Dataset for Collaborative Object REarrangement
Reinforcement Learning via Auxiliary Task Distillation
Grasp, See, and Place: Efficient Unknown Object Rearrangement with Policy Structure Prior
토론
이 논문과 연결된 포럼 글이 아직 없습니다.