Planning with learned binarized neural networks benchmarks for MaxSAT evaluation 2021

Buser Say, Scott Sanner, Jo Devriendt, Jakob Nordström, Peter Stuckey

Research output: Chapter in Book/Report/Conference proceedingChapter (Report)Research

Abstract

This document provides a brief introduction to learned automated planning problem where the state transition function is in the form of a binarized neural network (BNN), presents a general MaxSAT encoding for this problem, and describes the four domains, namely; Navigation, Inventory Control, System Administrator and Cellda, that are submitted as benchmarks for MaxSAT Evaluation 2021.
Original languageEnglish
Title of host publicationMaxSAT Evaluation 2021
Subtitle of host publicationSolver and Benchmark Descriptions
EditorsFahiem Bacchus, Jeremias Berg, Matti Järvisalo, Ruben Martins
Place of PublicationHelsinki Finland
PublisherUniversity of Helsinki
Pages32-36
Number of pages5
Publication statusPublished - 2021

Cite this