Flow-based Attribution in Graphical Models: A Recursive Shapley Approach.
Raghav SingalGeorge MichailidisHoiyi NgPublished in: ICML (2021)
Keyphrases
- graphical models
- belief propagation
- probabilistic model
- random variables
- bayesian networks
- probabilistic graphical models
- probabilistic inference
- approximate inference
- structure learning
- map inference
- belief networks
- markov networks
- conditional random fields
- exact inference
- game theory
- factor graphs
- conditional independence
- statistical inference
- graph structure
- flow field
- information extraction
- exponential family
- loopy belief propagation
- structural learning
- undirected graphical models
- relational data
- lower bound