On the Performance of Deep Generative Models of Realistic SAT Instances.
Iván GarzónPablo MesejoJesús Giráldez-CruPublished in: SAT (2022)
Keyphrases
- generative model
- sat instances
- sat solvers
- randomly generated
- satisfiability problem
- probabilistic model
- sat problem
- discriminative models
- deep belief networks
- discriminative learning
- computational properties
- em algorithm
- conditional random fields
- prior knowledge
- semi supervised
- phase transition
- stochastic local search
- generative and discriminative models
- hierarchical hidden markov models
- object categories
- np complete
- parameter learning
- constraint satisfaction problems
- expectation maximization
- objective function
- semi supervised learning
- information extraction
- lower bound
- search algorithm