Completed
Most important theoretical result
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Neural Network Verification as Piecewise Linear Optimization
Automatically move to the next video in the Classroom when playback concludes
- 1 Intro
- 2 Neural network verification
- 3 Key insights and approach
- 4 Optimization over a trained neural network
- 5 Fitting unknown functions to make predictions
- 6 Application: Deep reinforcement learning
- 7 Application: Designing DNA for protein binding
- 8 Neural networks in one slide
- 9 Most important theoretical result
- 10 MIP formulations for a single ReLU neuron
- 11 MIP formulation strength
- 12 Formulations for convex PWL functions
- 13 Network 1: Small network standard training
- 14 Propagation algorithms
- 15 Computational results
- 16 Extensions: Binarized and quantized networks