Adversarial Object Rearrangement in Constrained Environments with Heterogeneous Graph Neural Networks.
Xibai LouHoujian YuRoss WorobelYang YangChanghyun ChoiPublished in: IROS (2023)
Keyphrases
- neural network
- highly dynamic
- pattern recognition
- object model
- random walk
- d objects
- heterogeneous environments
- object level
- dynamic environments
- back propagation
- directed graph
- complex objects
- artificial neural networks
- partial occlusion
- weighted graph
- graph representation
- graph model
- physical objects
- graph structure
- graph theory
- bipartite graph
- fuzzy logic
- object segmentation
- multi layer
- graph databases
- graph partitioning
- spatial relationships
- multilayer perceptron
- graph matching
- shortest path