Welcome to the new version of CaltechAUTHORS. Login is currently restricted to library staff. If you notice any issues, please email coda@library.caltech.edu
Published August 20, 2017 | Published + Submitted
Book Section - Chapter Open

Learning-Based Abstractions for Nonlinear Constraint Solving

Abstract

We propose a new abstraction refinement procedure based on machine learning to improve the performance of nonlinear constraint solving algorithms on large-scale problems. The proposed approach decomposes the original set of constraints into smaller subsets, and uses learning algorithms to propose sequences of abstractions that take the form of conjunctions of classifiers. The core procedure is a refinement loop that keeps improving the learned results based on counterexamples that are obtained from partial constraints that are easy to solve. Experiments show that the proposed techniques significantly improved the performance of state-of-the-art constraint solvers on many challenging benchmarks. The mechanism is capable of producing intermediate symbolic abstractions that are also important for many applications and for understanding the internal structures of hard constraint solving problems.

Additional Information

© 2017 International Joint Conferences on Artificial Intelligence. This work was partially supported by STARnet, a Semiconductor Research Corporation program, sponsored by MARCO and DARPA and in part by Toyota InfoTechnology Center and NSF CPS1446725. The authors would also like to thank Joel W. Burdick for helpful input.

Attached Files

Published - 0083.pdf

Submitted - 3f140e46-a9be-4750-afeb-eeddb3a821ff.out.pdf

Files

0083.pdf
Files (598.0 kB)
Name Size Download all
md5:e129599a7513802c88aded1749cfa122
304.2 kB Preview Download
md5:b223a09cfa5c45c420bf86aef6db6308
293.8 kB Preview Download

Additional details

Created:
August 19, 2023
Modified:
October 25, 2023