Exploring Over-smoothing in Graph Attention Networks from the Markov Chain Perspective.
Weichen ZhaoChenguang WangCongying HanTiande GuoPublished in: FAIML (2023)
Keyphrases
- markov chain
- random walk
- transition matrix
- stationary distribution
- steady state
- finite state
- transition probabilities
- state space
- monte carlo
- markov process
- monte carlo simulation
- markov model
- monte carlo method
- graph structure
- directed graph
- social networks
- markov chain monte carlo
- reinforcement learning
- neural network
- single server
- gibbs sampling
- network structure
- gibbs sampler