DeepAI AI Chat
Log In Sign Up

Interior Point Methods with Adversarial Networks

by   Rafid Mahmood, et al.
York University

We present a new methodology, called IPMAN, that combines interior point methods and generative adversarial networks to solve constrained optimization problems with feasible sets that are non-convex or not explicitly defined. Our methodology produces ϵ-optimal solutions and demonstrates that, when there are multiple global optima, it learns a distribution over the optimal set. We apply our approach to synthetic examples to demonstrate its effectiveness and to a problem in radiation therapy treatment optimization with a non-convex feasible set.


Understanding Overparameterization in Generative Adversarial Networks

A broad class of unsupervised deep learning methods such as Generative A...

Stochastic Constrained DRO with a Complexity Independent of Sample Size

Distributionally Robust Optimization (DRO), as a popular method to train...

Optimization with Non-Differentiable Constraints with Applications to Fairness, Recall, Churn, and Other Goals

We show that many machine learning goals, such as improved fairness metr...

Solving a class of non-convex min-max games using adaptive momentum methods

Adaptive momentum methods have recently attracted a lot of attention for...

Integration of AI and mechanistic modeling in generative adversarial networks for stochastic inverse problems

The problem of finding distributions of input parameters for determinist...

An Unsupervised Deep Unrolling Framework for Constrained Optimization Problems in Wireless Networks

In wireless network, the optimization problems generally have complex co...