四川大学学报（自然科学版）2026，Vol.63Issue(3)：586-596,11.DOI:10.19907/j.0490-6756.250213
基于动态-分层-对抗协同优化的知识增强BERT文本分类模型
Classification:A triple enhancement framework combining dynamic,hierarchical,and adversarial mechanisms
摘要
Abstract
Pre-trained language models like BERT excel at capturing general linguistic patterns but often un-derperform in domain-specific text classification due to a lack of structured knowledge.To enhance reasoning capabilities in specialized domains,knowledge injection has emerged as a mainstream approach.However,excessive knowledge fusion may distort the original semantics of sentences,leading to Knowledge Noise(KN).To address domain-specific knowledge gaps and mitigate the impact of KN in text classification,we propose a triple-enhanced BERT framework that integrates a Knowledge-Enhanced Dynamic Attention(KEDA),a Hierarchical Knowledge Fusion Network(HKFN),and an Adversarial Knowledge Regularizer(AKR).Experimental results on seven diverse domain-specific corpora demonstrate that our model signifi-cantly outperforms existing baselines.关键词
知识增强/动态注意力/分层知识融合/知识正则化/文本分类Key words
knowledge-enhanced/dynamic attention/hierarchical knowledge fusion/knowledge regulariza-tion/text classification分类
信息技术与安全科学引用本文复制引用
孙豪,蒲亦非..基于动态-分层-对抗协同优化的知识增强BERT文本分类模型[J].四川大学学报（自然科学版）,2026,63(3):586-596,11.基金项目
国家自然科学基金面上项目(62171303) （62171303）
中国兵器装备集团(成都)火控技术中心项目(非密)(HK20-03) （成都）
国家重点研发项目(2018YFC0830300) （2018YFC0830300）
