Adversarial Object Rearrangement in Constrained Environments with Heterogeneous Graph Neural Networks.
Xibai LouHoujian YuRoss WorobelYang YangChanghyun ChoiPublished in: CoRR (2023)
Keyphrases
- neural network
- highly dynamic
- pattern recognition
- real world
- artificial neural networks
- d objects
- object model
- graph theory
- graph representation
- graph theoretic
- graph model
- directed graph
- topological information
- neural network model
- complex objects
- heterogeneous environments
- moving objects
- target object
- multi agent
- heterogeneous networks
- graph based algorithm
- object models
- activation function
- spanning tree
- multi layer
- weighted graph
- object segmentation
- graph structure
- object classes
- bipartite graph
- partial occlusion
- recurrent neural networks
- multilayer perceptron
- data objects
- random walk
- graph cuts
- graphical models
- viewpoint
- object recognition