Explicitly Imposing Constraints in Deep Networks via Conditional Gradients Gives Improved Generalization and Faster Convergence.
Sathya N. RaviTuan DinhVishnu Suresh LokhandeVikas SinghPublished in: AAAI (2019)
Keyphrases
- faster convergence
- convergence speed
- step size
- global optimum
- convergence rate
- computer networks
- global constraints
- genetic algorithm
- linear constraints
- constraint satisfaction
- complex networks
- global optimization
- improved algorithm
- conditional distributions
- heterogeneous networks
- network design
- constrained optimization
- constraint programming
- search space
- artificial neural networks
- feature space
- social networks