基于多特征和跨模态知识蒸馏的鱼病命名实体识别  

Named Entity Recognition of Fish Disease Based on Multi-feature and Cross-Modal Knowledge Distillation

作　　者：沈志成 陈明[1,2] SHEN Zhicheng;CHEN Ming(School of Information,Shanghai Ocean University,Shanghai 201306,China;Key Laboratory of Fishery Information,Ministry of Agriculture and Rural Affairs,Shanghai 201306,China)

机构地区：[1]上海海洋大学信息学院,上海201306 [2]农业农村部渔业信息重点实验室,上海201306

出　　处：《数据采集与处理》2025年第1期230-246,共17页Journal of Data Acquisition and Processing

基　　金：广东省重点领域研发计划项目(2021B0202070001)。

摘　　要：为解决多模态鱼病知识缺乏合理安排的问题,同时降低知识蒸馏过程的冗余数据,从而部署存储低、样本小、精度高的识别模型,提出一种基于多特征协同预测-跨模态多头蒸馏的方法,命名为FSFDAI-TMRD。在多特征协同预测方面,重点改进了原多任务多特征协同预测架构。首先使用更细粒度的BMES(Begin-middle-end-single)法代替原工作中BIO(Begin-inside-outside)法的粗略标注,其次修改原架构的联合概率分布计算公式,使得模型可以更好地识别嵌套名词实体。在跨模态多头蒸馏方面,本文运用了跨模态注意力机制。首先计算合并、拆分和点积后的多头关系矩阵,其次利用相对熵进行知识蒸馏,使得模型可以更好地对齐异构师生间的中间特征。同时,本文还应用了双仿射注意力机制及对抗性权重扰动函数等方法,加强学习语义语音和字形词义等多特征知识。与主流模型相比,本文方法的精确率、召回率和F1值分别提升了0.45%、3.96%和2.28%,并且存储优化比例提高3.01%,模型参数规模缩小94.86%。In order to solve the lack of reasonable arrangement of multi-modal fish disease knowledge,and at the same time reduce the redundant data in the knowledge distillation process,so as to deploy a recognition model with low storage,small samples,and high accuracy,this paper proposes a new method,named as FSFDAI-TMRD.In terms of multi-feature collaborative prediction,this paper focuses on improving the original multi-feature collaborative multi-feature prediction architecture of multi-tasks.Firstly,the finer-grained begin-middle-end-single(BMES)method is used instead of the rough labeling of the begin-inside-outside(BIO)method in the original work.Secondly,the formula for calculating the joint probability distribution of the original architecture is modified,so that the model can better recognize the nested noun entities.In terms of cross-modal multi-head distillation,this paper proposes to employ a crossmodal attention mechanism.Firstly,it calculates the multi-head relationship matrix after merging,splitting,and dot product,and secondly,it utilizes the relative entropy for knowledge distillation,so that the model can better align the intermediate features between heterogeneous teachers and students.Meanwhile,this paper also applies the biaffine attention and adversarial weight perturbation function to enhance the learning of multi-feature knowledge such as semantic phonology and word form word meaning.Compared with the mainstream model,the precision,recall and F1 value of the FSFDAI-TMRD method are improved by 0.45%,3.96%and 2.28%,respectively.The storage optimization ratio is improved by 3.01%and the model parameter size is reduced by 94.86%.

关 键 词：知识蒸馏 鱼病命名实体识别 对抗训练 双仿射注意力机制 模型压缩 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]