神经网络与逻辑规则的结合

IF 0.8 Q4 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE
Lujiang Zhang
{"title":"神经网络与逻辑规则的结合","authors":"Lujiang Zhang","doi":"10.1142/s1469026823500153","DOIUrl":null,"url":null,"abstract":"How to utilize symbolic knowledge in deep learning is an important problem. Deep neural networks are flexible and powerful, while symbolic knowledge has the virtue of interpretability and intuitiveness. It is necessary to combine the two together to inject symbolic knowledge into neural networks. We propose a novel approach to combine neural networks with logic rules. In this approach, task-specific supervised learning and policy-based reinforcement learning are performed alternately to train a neural model until convergence. The basic idea is to use supervised learning to train a deep model and use reinforcement learning to propel the deep model to meet logic rules. In the process of the policy gradient reinforcement learning, if a predicted output of a deep model meets all logical rules, the deep model is given a positive reward, otherwise, it is given a negative reward. By maximizing the expected rewards, the deep model can be gradually adjusted to meet logical constraints. We conduct experiments on the tasks of named entity recognition. The experimental results demonstrate the effectiveness of our method.","PeriodicalId":45994,"journal":{"name":"International Journal of Computational Intelligence and Applications","volume":" ","pages":""},"PeriodicalIF":0.8000,"publicationDate":"2023-03-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Combining Neural Networks with Logic Rules\",\"authors\":\"Lujiang Zhang\",\"doi\":\"10.1142/s1469026823500153\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"How to utilize symbolic knowledge in deep learning is an important problem. Deep neural networks are flexible and powerful, while symbolic knowledge has the virtue of interpretability and intuitiveness. It is necessary to combine the two together to inject symbolic knowledge into neural networks. We propose a novel approach to combine neural networks with logic rules. In this approach, task-specific supervised learning and policy-based reinforcement learning are performed alternately to train a neural model until convergence. The basic idea is to use supervised learning to train a deep model and use reinforcement learning to propel the deep model to meet logic rules. In the process of the policy gradient reinforcement learning, if a predicted output of a deep model meets all logical rules, the deep model is given a positive reward, otherwise, it is given a negative reward. By maximizing the expected rewards, the deep model can be gradually adjusted to meet logical constraints. We conduct experiments on the tasks of named entity recognition. The experimental results demonstrate the effectiveness of our method.\",\"PeriodicalId\":45994,\"journal\":{\"name\":\"International Journal of Computational Intelligence and Applications\",\"volume\":\" \",\"pages\":\"\"},\"PeriodicalIF\":0.8000,\"publicationDate\":\"2023-03-27\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"International Journal of Computational Intelligence and Applications\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1142/s1469026823500153\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q4\",\"JCRName\":\"COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Journal of Computational Intelligence and Applications","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1142/s1469026823500153","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE","Score":null,"Total":0}
引用次数: 0

摘要

如何在深度学习中利用符号知识是一个重要的问题。深度神经网络是灵活而强大的,而符号知识具有可解释性和直观性的优点。有必要将两者结合起来,为神经网络注入符号知识。我们提出了一种将神经网络与逻辑规则相结合的新方法。在这种方法中,任务特定的监督学习和基于策略的强化学习交替执行,以训练神经模型直到收敛。其基本思想是使用监督学习来训练深度模型,并使用强化学习来推动深度模型满足逻辑规则。在策略梯度强化学习过程中,如果深度模型的预测输出满足所有逻辑规则,则给予深度模型正奖励,否则给予负奖励。通过最大化预期回报,可以逐步调整深度模型以满足逻辑约束。我们对命名实体识别的任务进行了实验。实验结果证明了我们方法的有效性。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
Combining Neural Networks with Logic Rules
How to utilize symbolic knowledge in deep learning is an important problem. Deep neural networks are flexible and powerful, while symbolic knowledge has the virtue of interpretability and intuitiveness. It is necessary to combine the two together to inject symbolic knowledge into neural networks. We propose a novel approach to combine neural networks with logic rules. In this approach, task-specific supervised learning and policy-based reinforcement learning are performed alternately to train a neural model until convergence. The basic idea is to use supervised learning to train a deep model and use reinforcement learning to propel the deep model to meet logic rules. In the process of the policy gradient reinforcement learning, if a predicted output of a deep model meets all logical rules, the deep model is given a positive reward, otherwise, it is given a negative reward. By maximizing the expected rewards, the deep model can be gradually adjusted to meet logical constraints. We conduct experiments on the tasks of named entity recognition. The experimental results demonstrate the effectiveness of our method.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
CiteScore
2.90
自引率
0.00%
发文量
25
期刊介绍: The International Journal of Computational Intelligence and Applications, IJCIA, is a refereed journal dedicated to the theory and applications of computational intelligence (artificial neural networks, fuzzy systems, evolutionary computation and hybrid systems). The main goal of this journal is to provide the scientific community and industry with a vehicle whereby ideas using two or more conventional and computational intelligence based techniques could be discussed. The IJCIA welcomes original works in areas such as neural networks, fuzzy logic, evolutionary computation, pattern recognition, hybrid intelligent systems, symbolic machine learning, statistical models, image/audio/video compression and retrieval.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信