arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2108.00633cs.AI
收藏

使用学习的二值化神经网络基准进行规划,用于MaxSAT评估2021

Planning with Learned Binarized Neural Networks Benchmarks for MaxSAT Evaluation 2021

Monash University(莫纳什大学);University of Toronto(多伦多大学);KU Leuven(鲁汶大学);University of Copenhagen(哥本哈根大学)

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

摘要:

本文简要介绍了学习自动规划问题,其中状态转换函数是二进制神经网络(BNN)的形式,提出了一个通用的MaxSAT编码这个问题,并描述了四个域,即:导航,库存控制,系统管理员和Cellda,这是提交的基准MaxSAT评估2021。

英文摘要:

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.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑