Training Quantized Neural Networks to Global Optimality via Semidefinite Programming.
Burak BartanMert PilanciPublished in: ICML (2021)
Keyphrases
- semidefinite programming
- semidefinite
- global optimality
- neural network
- training process
- linear programming
- convex programming
- interior point methods
- globally optimal
- global optimization
- primal dual
- kernel matrix
- maximum margin
- training set
- training examples
- training samples
- supervised learning
- convex sets
- nonlinear programming
- kernel methods
- linear program
- graphical models
- np hard
- training data
- learning algorithm