Differentiable Optimization

Differentiable Optimization #

Differentiable optimization makes optimization layers differentiable so they can be embedded in neural networks, enabling end-to-end learning with optimization as a component.

Recent Literature #

1. OptNet: Differentiable Optimization as a Layer in Neural Networks ICML, 2017. paper, code

   Brandon Amos, J. Zico Kolter

2. Differentiation of Blackbox Combinatorial Solvers ICLR, 2020. paper, code

   Maria-Florina Balcan, Dan DeFreitas, Amit Levi, Segev Shlomovich

3. CombOptNet: Fit the Right NP-Hard Problem by Learning Integer Programming Constraints ICML, 2021. paper, code

   Minhan Han, Patrick Wilder, Valdinei Freire, Harikrishna Narasimhan, Andrew Perrault, Milind Tambe

4. Implicit Differentiation of Nonlinear Optimization Problems NeurIPS, 2021. paper, code

   Jean-Pierre Hespanha, Noureddine Elhadji Boularas, Daniel Cremers

5. Decision-Focused Learning in Games ICML, 2023. paper

   Yoann Thesot, Maxime Wabartha, Vincent François-Lavet

6. Learning to Prescribe with Differentiable Optimization ICML, 2023. paper, code

   Niki Zadeh, J. Zico Kolter, Brandon Amos