The adam algorithm has been very successful for solving non-convex optimization problems that appear in deep learning. Are there ways to extend adam to solve constrained optimization problems? Among the papers citing Adam, there are papers based on an augmented Lagrangian approach. Maybe, they are dealing with constraints by constructing an augmented Lagrangian. However, there are too many such papers. Some guidance will be useful.

  • 1
    $\begingroup$ Adam could be used out-of-the-box with any method that recasts a constrained optimization problem as an unconstrained problem (e.g. Lagrange multipliers, reparameterization, etc.). There's nothing specific to Adam at all in this kind of approach. $\endgroup$ – user20160 Aug 21 at 0:52
  • $\begingroup$ Thanks a lot for the clarification. $\endgroup$ – haripkannan Aug 21 at 6:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.