RUN-CSP: Unsupervised Learning of Message Passing Networks for Binary Constraint Satisfaction Problems.
Jan TönshoffMartin RitzertHinrikus WolfMartin GrohePublished in: CoRR (2019)
Keyphrases
- message passing
- unsupervised learning
- belief propagation
- distributed systems
- singly connected
- constraint satisfaction problems
- shared memory
- markov random field
- sum product algorithm
- factor graphs
- probabilistic inference
- constraint satisfaction
- inference in graphical models
- approximate inference
- dimensionality reduction
- distributed shared memory
- object recognition
- supervised learning
- maximum likelihood