基于层级软提示交互融合的少样本事件方面类别检测方法

doi:10.19678/j.issn.1000-3428.0069414

计算机工程 ›› 2025, Vol. 51 ›› Issue (9): 120-128. doi: 10.19678/j.issn.1000-3428.0069414

基于层级软提示交互融合的少样本事件方面类别检测方法

艾传鲜¹^,²^,³, 郭军军¹^,²^,*(), 尹兆良¹^,²^,³

1. 昆明理工大学信息工程与自动化学院, 云南昆明 650500
2. 昆明理工大学云南省人工智能重点实验室, 云南昆明 650500
3. 国家计算机网络应急技术处理协调中心云南分中心, 云南昆明 650000

收稿日期:2024-02-23 修回日期:2024-04-12 出版日期:2025-09-15 发布日期:2024-06-25
通讯作者: 郭军军
基金资助:
国家重点研发计划(202301AT070444); 国家自然科学基金(62366025); 云南省科技厅自然科学基金(202202AE090008-3)

Method for Event Aspect Category Detection in Few Shot Scenarios via Hierarchical Soft Prompt Interaction Fusion

AI Chuanxian¹^,²^,³, GUO Junjun¹^,²^,*(), YIN Zhaoliang¹^,²^,³

1. Faculty of Information Engineering and Automation, Kunming University of Science and Technology, Kunming 650500, Yunnan, China
2. Yunnan Key Laboratory of Artificial Intelligence, Kunming University of Science and Technology, Kunming 650500, Yunnan, China
3. Yunnan Branch, National Computer Network Emergency Response Technical Team/Coordination Center of China, Kunming 650000, Yunnan, China

Received:2024-02-23 Revised:2024-04-12 Online:2025-09-15 Published:2024-06-25
Contact: GUO Junjun

摘要/Abstract

摘要：

事件方面类别检测(ACD)旨在识别出给定事件文本中的方面类别, 相关研究可以辅助获取不同领域和事件文本中的关键信息, 特别是在社交媒体舆情研究中具有实用价值。在社交媒体舆情事件发展前期, 事件文本标记数据稀缺, 如何基于少量标记数据实现准确的事件方面检测是一个亟待解决的问题。提出一种基于预训练模型的多层级软提示交互融合少样本事件方面类别检测方法, 基于预训练构建多层级的软提示模板, 分别与预训练模型进行层级语义表征和交互融合, 并自适应地融合多层级的提示表征, 从而提升少样本事件方面类别检测的效果。在自构建中文社交媒体少样本数据集和英文数据集上进行实验, 实验结果证明所提方法明显优于其他基线方法, 此外, 消融实验和可视化结果验证了所提多层级提示交互融合模块的有效性。

关键词: 少样本, 提示学习, 软提示, 方面类别检测, 多层级提示交互融合

Abstract:

Event Aspect Category Detection (ACD) aims to identify the aspect categories present in event text. Data need to be collected from various fields and textual events, particularly when researching public opinion on social media. The first phase of social media opinion events lacks sufficient data for labeling event text. The pressing issue is precisely detecting event aspects using a limited amount of labeled data. This paper presents a novel method for event ACD with limited samples. This method utilizes a pre-trained model to construct soft prompt templates, performs hierarchical semantic characterization and interaction fusion, and adaptively combines multilayer prompt characterizations. The objective is to enhance the accuracy of event ACD using limited samples. Experiments on a self-constructed Chinese social media dataset and English dataset demonstrate that the proposed method is significantly superior to other baseline methods. Further ablation experiments and visualizations confirm the effectiveness of the proposed multilayer prompt interaction fusion module.

Key words: few-shot, prompt learning, soft prompt, Aspect Category Detection (ACD), multilayer prompt interaction fusion

艾传鲜, 郭军军, 尹兆良. 基于层级软提示交互融合的少样本事件方面类别检测方法[J]. 计算机工程, 2025, 51(9): 120-128.

AI Chuanxian, GUO Junjun, YIN Zhaoliang. Method for Event Aspect Category Detection in Few Shot Scenarios via Hierarchical Soft Prompt Interaction Fusion[J]. Computer Engineering, 2025, 51(9): 120-128.

https://www.ecice06.com/CN/Y2025/V51/I9/120

图/表 12

图1 本文方法结构

Fig.1 Structure of the proposed method

图2 自建数据集方面类别分布

Fig.2 Distribution of aspect categories in self-constructed dataset

图3 不同提示长度对模型性能的影响

Fig.3 Impact of different prompt lengths on model performance

图4 不同样本数量对模型性能的影响

Fig.4 Impact of different numbers of shots on model performance

图5 可视化分析

Fig.5 Visualization and analysis

参考文献 32

1	PONTIKI M, GALANIS D, PAPAGEORGIOU H, et al. SemEval-2016 task 5: aspect based sentiment analysis[C]//Proceedings of International Workshop on Semantic Evaluation. San Diego, USA: [s. n. ], 2016: 1-10.
2	HU M T, ZHAO S W, GUO H L, et al. Multi-label few-shot learning for aspect category detection[EB/OL]. [2023-12-12]. https://arxiv.org/abs/2105.14174v1.
3	KIRANGE D K , DESHMUKH R R , KIRANGE M D K . Aspect based sentiment analysis SemEval-2014 task 4. Asian Journal of Computer Science & Information Technology, 2014, 4 (8): 1- 10.
4	JANG H , REMPEL E , ROE I , et al. Tracking public attitudes toward COVID-19 vaccination on tweets in Canada: using aspect-based sentiment analysis. Journal of Medical Internet Research, 2022, 24 (3): e35016. doi: 10.2196/35016
5	SAMUEL J , ALI G G M N , RAHMAN M M , et al. COVID-19 public sentiment insights and machine learning for tweets classification. Information, 2020, 11 (6): 314. doi: 10.3390/info11060314
6	ZHAO F, SHEN Y C, WU Z, et al. Label-driven denoising framework for multi-label few-shot aspect category detection[EB/OL]. [2023-12-12]. https://arxiv.org/abs/2210.04220v1.
7	WANG Z Y, IWAIHARA M. Few-shot multi-label aspect category detection utilizing prototypical network with sentence-level weighting and label augmentation[M]//STRAUSS C, AMAGASA T, KOTSIS G, et al. Database and expert systems applications. Berlin, Germany: Springer, 2023: 363-377.
8	DEVLIN J, CHANG M W, LEE K, et al. BERT: pre-training of deep bidirectional transformers for language understanding[EB/OL]. [2023-12-12]. http://arxiv.org/abs/1810.04805.
9	LESTER B, AL-RFOU R, CONSTANT N, et al. The power of scale for parameter-efficient prompt tuning[EB/OL]. [2023-12-12]. https://arxiv.org/abs/2104.08691v2.
10	LIU P F , YUAN W Z , FU J L , et al. Pre-train, prompt, and predict: a systematic survey of prompting methods in natural language processing. ACM Computing Surveys, 2023, 55 (9): 1- 35.
11	顾勋勋, 刘建平, 邢嘉璐, 等. 文本分类中Prompt Learning方法研究综述. 计算机工程与应用, 2024, 60 (11): 50- 61.
	GU X X , LIU J P , XING J L , et al. A summary of the research on Prompt Learning methods in text classification. Computer Engineering and Applications, 2024, 60 (11): 50- 61.
12	LIU X, JI K X, FU Y C, et al. P-tuning v2: prompt tuning can be comparable to fine-tuning universally across scales and tasks[EB/OL]. [2023-12-12]. https://arxiv.org/abs/2110.07602v3.
13	ZHANG N Y, LI L Q, CHEN X, et al. Differentiable prompt makes pre-trained language models better few-shot learners[EB/OL]. [2023-12-12]. https://arxiv.org/abs/2108.13161v7.
14	SEOH R, BIRLE I, TAK M, et al. Open aspect target sentiment classification with natural language prompts[EB/OL]. [2023-12-12]. https://arxiv.org/abs/2109.03685v1.
15	MIKOLOV T, CHEN K, CORRADO G, et al. Efficient estimation of word representations in vector space[EB/OL]. [2023-12-12]. https://arxiv.org/abs/1301.3781v3.
16	TOH Z, SU J. NLANGP: supervised machine learning system for aspect category classification and opinion target extraction[C]//Proceedings of the 9th International Workshop on Semantic Evaluation (SemEval 2015). Stroudsburg, USA: Association for Computational Linguistics, 2015: 1-10.
17	SAEIDI M, BOUCHARD G, LIAKATA M, et al. SentiHood: targeted aspect based sentiment analysis dataset for urban neighbourhoods[C]//Proceedings of the 26th International Conference on Computational Linguistics: Technical Papers. Osaka, Japan: The COLING 2016 Organizing Committee, 2016: 1546-1556.
18	TANG D Y, QIN B, LIU T. Document modeling with gated recurrent neural network for sentiment classification[C]//Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing. Stroudsburg, USA: Association for Computational Linguistics, 2015: 1-10.
19	WANG W Y, PAN S J, DAHLMEIER D, et al. Recursive neural conditional random fields for aspect-based sentiment analysis[EB/OL]. [2023-12-12]. https://arxiv.org/abs/1603.06679v3.
20	LIANG B, DU J C, XU R F, et al. Context-aware embedding for targeted aspect-based sentiment analysis[EB/OL]. [2023-12-12]. https://arxiv.org/abs/1906.06945v1.
21	PETERS M E, NEUMANN M, IYYER M, et al. Deep contextualized word representations[C]//Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1(Long Papers). New Orleans, USA: Association for Computational Linguistics, 2018: 2227-2237.
22	RADFORD A, NARASIMHAN K, SALIMANS T, et al. Improving language understanding by generative pre-training[EB/OL]. [2023-12-12]. https://www.mikecaptain.com/resources/pdf/GPT-1.pdf.
23	VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. New York, USA: ACM Press, 2017: 6000-6010.
24	KALYAN K S, RAJASEKHARAN A, SANGEETHA S. AMMUS: a survey of transformer-based pretrained models in natural language processing[EB/OL]. [2023-12-12]. https://arxiv.org/abs/2108.05542v2.
25	SCHOUTEN K, FRASINCAR F, JONG F. COMMIT-p1wp3: a co-occurrence based approach to aspect-level sentiment analysis[C]//Proceedings of the 8th International Workshop on Semantic Evaluation (SemEval 2014). Dublin, Ireland: Association for Computational Linguistics, 2014: 203-207.
26	GHADERY E, MOVAHEDI S, FAILI H, et al. An unsupervised approach for aspect category detection using soft cosine similarity measure[EB/OL]. [2023-12-12]. https://arxiv.org/abs/1812.03361v2.
27	BROWN T B, MANN B, RYDER N, et al. Language models are few-shot learners[EB/OL]. [2023-12-12]. http://arxiv.org/abs/2005.14165.
28	LIU X , ZHENG Y N , DU Z X , et al. GPT understands, too. AI Open, 2024, 5, 208- 215. doi: 10.1016/j.aiopen.2023.08.012
29	SONG Y S , WANG T , CAI P Y , et al. A comprehensive survey of few-shot learning: evolution, applications, challenges, and opportunities. ACM Computing Surveys, 2023, 55 (13s): 1- 40.
30	李子成, 常晓琴, 李雅梦, 等. 基于联合学习的少样本多类别情感分类方法. 北京大学学报(自然科学版), 2023, 59 (1): 57- 64.
	LI Z C , CHANG X Q , LI Y M , et al. A joint learning approach to few-shot learning for multi-category sentiment classification. Acta Scientiarum Naturalium Universitatis Pekinensis, 2023, 59 (1): 57- 64.
31	李睿凡, 魏志宇, 范元涛, 等. 增强提示学习的少样本文本分类方法. 北京大学学报(自然科学版), 2024, 60 (1): 1- 12.
	LI R F , WEI Z Y , FAN Y T , et al. Enhanced prompt learning for few-shot text classification method. Acta Scientiarum Naturalium Universitatis Pekinensis, 2024, 60 (1): 1- 12.
32	GAO T, FISCH A, CHEN D. Making pre-trained language models better few-shot learners[EB/OL]. [2023-12-12]. http://arxiv.org/abs/2012.15723.

[1]	逯暄, 景路琪, 彭甫镕. 基于增量学习的结直肠息肉分割方法[J]. 计算机工程, 2025, 51(7): 284-293.
[2]	杨冬菊, 黄俊涛. 基于大语言模型的中文科技文献标注方法[J]. 计算机工程, 2024, 50(9): 113-120.
[3]	周炫余, 吴莲华, 郑勤华, 肖天星, 王紫璇, 张思敏. 联合语义提示和记忆增强的弱监督跳绳视频异常检测方法[J]. 计算机工程, 2024, 50(7): 87-95.
[4]	曾碧卿, 陈鹏飞, 姚勇涛. 融合思维链和低秩自适应微调的方面情感三元组抽取[J]. 计算机工程, 2024, 50(7): 53-62.
[5]	邓远飞, 李加伟, 蒋运承. 基于知识注入提示学习的专利短语相似度计算[J]. 计算机工程, 2024, 50(4): 294-302.
[6]	陈志强, 仇瑜, 朱宇, 王晓英. 基于先验知识引导提示学习的自监督分类法补全[J]. 计算机工程, 2024, 50(12): 151-162.
[7]	李鸿鹏, 马博, 杨雅婷, 王磊, 王震, 李晓. 基于槽位语义增强提示学习的篇章级事件抽取方法[J]. 计算机工程, 2023, 49(9): 23-31.
[8]	张博旭, 蒲智, 程曦. 基于提示学习的维吾尔语文本分类研究[J]. 计算机工程, 2023, 49(6): 292-299,313.
[9]	王雨燕, 赵佳鹏, 时金桥, 申立艳, 刘洪梦, 杨燕燕. 暗网网页用户身份信息聚合方法[J]. 计算机工程, 2023, 49(11): 187-194, 210.
[10]	胡彬, 王晓军, 张雷. 一种半监督对抗鲁棒模型无关元学习方法[J]. 计算机工程, 2022, 48(12): 112-118.
[11]	樊星, 刘占文, 林杉, 窦瑞娟. 融合人眼特征与深度学习的疲劳驾驶检测模型[J]. 计算机工程, 2021, 47(8): 243-250.

选择文件类型/文献管理软件名称

选择包含的内容

基于层级软提示交互融合的少样本事件方面类别检测方法

Method for Event Aspect Category Detection in Few Shot Scenarios via Hierarchical Soft Prompt Interaction Fusion

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 12

参考文献 32

相关文章 11

编辑推荐

Metrics

本文评价

模态框（Modal）标题

选择文件类型/文献管理软件名称

选择包含的内容

基于层级软提示交互融合的少样本事件方面类别检测方法

Method for Event Aspect Category Detection in Few Shot Scenarios via Hierarchical Soft Prompt Interaction Fusion

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 12

参考文献 32

相关文章 11

编辑推荐

Metrics

本文评价