End-to-End Learning for Constrained Optimization

In this thesis, we study the integration of constrained optimization algorithms with the training of deep neural networks. In particular, we we primarily interested in end-to-end trainable prediction and decision models composed of differentiable components. The use of such techniques spans several broad application areas, which we divide into two categories. When Learning to Optimize, the goal is to train neural networks to solve or aid in the solution of constrained optimization problems. In the Predict-Then-Optimize setting, the goal is to optimize decisions under uncertainty by estimating unknown coefficients in optimization problems from correlated data. Both frameworks stem from efforts to enhance optimization modeling technology for operations research and decision making tasks. This thesis contributes to both areas, and seeks to combine techniques from each to enhance the expressive and computational ability of models that learn to make decisions.