采用平衡函数的大规模多标签文本分类

doi:10.3778/j.issn.1002-8331.2209-0472

摘要/Abstract

摘要： 大规模多标签文本分类是自然语言处理领域的一项挑战性任务。该任务存在标签数据长尾分布的情况，在这种情况下，模型学习尾部标签分类能力不佳，导致模型的整体分类效果不理想。为解决以上问题，提出采用平衡函数的大规模多标签文本分类方法。该方法使用BERT预训练模型对文本进行词嵌入处理，进一步使用预训练模型中多层编码器的拼接输出作为文本向量表示，获取了丰富的文本语义信息，提高了模型收敛速度。最后采用平衡函数针对预测标签的训练损失赋予不同的衰减权重，提高了方法在尾部标签分类上的学习能力。在Eurlex-4K和Wiki10-31K数据集上的实验结果表明，评价指标P@1、P@3和P@5上分别达到86.95%、74.12%、61.43%和88.57%、77.46%、67.90%。

关键词: 自然语言处理, 大规模多标签文本分类, BERT, 平衡函数, 深度学习

Abstract: Extreme multi-label text classification is a challenging task in the field of natural language processing. In this task, there is a long-tailed distribution situation of labeled data. In this situation, model has a poor ability to learn tail labels classification, which results the overall classification effect is not good. In order to address the above problems, an extreme multi-label text classification method based on balance function is proposed. Firstly, the BERT pre-training model is used for word embedding. Further, the concatenated output of the multi-layer encoder in the pre-trained model is used as the text vector representation to obtain richer text semantic information and improves the model convergence speed. Finally, the balance function is used to assign different attenuation weights to the training losses of different prediction labels, which improves the learning ability of the method on tail label classification. The experimental results on Eurlex-4K and Wiki10-31K datasets show that the evaluation indicators P@1, P@3 and P@5 respectively reach 86.95%, 74.12%, 61.43% and 88.57%, 77.46% and 67.90%.

Key words: natural language processing (NLP), extreme multi-label text classification, BERT, balance function, deep learning

陈钊鸿, 洪智勇, 余文华, 张昕. 采用平衡函数的大规模多标签文本分类[J]. 计算机工程与应用, 2024, 60(4): 163-172.

CHEN Zhaohong, HONG Zhiyong, YU Wenhua, ZHANG Xin. Extreme Multi-Label Text Classification Based on Balance Function[J]. Computer Engineering and Applications, 2024, 60(4): 163-172.

参考文献

[1] YANG Z, YANG D, DYER C, et al. Hierarchical attention networks for document classification[C]//Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, 2016: 1480-1489.
[2] AL-SABAHI K, ZUPING Z, NADHER M. A hierarchical structured self-attentive model for extractive document summarization (HSSAS)[J]. IEEE Access, 2018, 6: 24205-24212.
[3] CAMBRIA E, OLSHER D, RAJAGOPAL D. SenticNet 3: a common and common-sense knowledge base for cognition-driven sentiment analysis[C]//Proceedings of the 28th AAAI Conference on Artificial Intelligence, 2014.
[4] 刘鑫, 梅红岩, 王嘉豪, 等. 图神经网络推荐方法研究[J]. 计算机工程与应用, 2022, 58(10): 41-49.
LIU X, MEI H Y, WANG J H, et al. Research on graph neural network recommendation method[J]. Computer Engineering and Applications, 2022, 58(10): 41-49.
[5] AGRAWAL R, GUPTA A, PRABHU Y, et al. Multi-label learning with millions of labels: recommending advertiser bid phrases for web pages[C]//Proceedings of the 22nd International Conference on World Wide Web, 2013: 13-24.
[6] PARTALAS I, KOSMOPOULOS A, BASKIOTIS N, et al. LSHTC: a benchmark for large-scale text classification[J]. arXiv:1503.08581, 2015.
[7] 王岳, 李雅文, 李昂. 科技资源文本层次多标签分类方法[J]. 计算机工程与应用, 2023, 59(13): 92-98.
WANG Y, LI Y W, LI A. Academic resource text hierarchical multi-label classification[J]. Computer Engineering and Applications, 2023, 59(13): 92-98.
[8] LOZAMENCíA E, FüRNKRANZ J. Efficient pairwise multilabel classification for large-scale problems in the legal domain[C]//Proceedings of the Joint European Conference on Machine Learning and Knowledge Discovery in Databases. Berlin, Heidelberg: Springer, 2008: 50-65.
[9] DEVLIN J, CHANG M W, LEE K, et al. BERT: pre-training of deep bidirectional transformers for language understanding[J]. arXiv:1810.04805, 2018.
[10] ZUBIAGA A. Enhancing navigation on Wikipedia with social tags[J]. arXiv:1202. 5469, 2012.
[11] BABBAR R, SCH?LKOPF B. DiSMEC: distributed sparse machines for extreme multi-label classification[C]//Proceedings of the 10th ACM International Conference on Web Search and Data Mining, 2017: 721-729.
[12] BABBAR R, SCH?LKOPF B. Data scarcity, robustness and extreme multi-label classification[J]. Machine Learning, 2019, 108(8): 1329-1351.
[13] YEN I E H, HUANG X, RAVIKUMAR P, et al. PD-sparse: a primal and dual sparse approach to extreme multiclass and multilabel classification[C]//Proceedings of the 33rd International Conference on Machine Learning, 2016: 3069-3077.
[14] PRABHU Y, KAG A, HARSOLA S, et al. Parabel: partitioned label trees for extreme classification with application to dynamic search advertising[C]//Proceedings of the 2018 World Wide Web Conference, 2018: 993-1002.
[15] BHATIA K, JAIN H, KAR P, et al. Sparse local embeddings for extreme multi-label classification[C]//Advances in Neural Information Processing Systems 28, 2015.
[16] TAGAMI Y. AnnexML: approximate nearest neighbor search for extreme multi-label classification[C]//Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2017: 455-464.
[17] LIU J, CHANG W C, WU Y, et al. Deep learning for extreme multi-label text classification[C]//Proceedings of the 40th International ACM SIGIR Conference on Research and Development in Information Retrieval, 2017: 115-124.
[18] YOU R, ZHANG Z, WANG Z, et al. AttentionXML: label tree-based attention-aware deep model for high-performance extreme multi-label text classification[C]//Advances in Neural Information Processing Systems 32, 2019.
[19] CHANG W C, YU H F, ZHONG K, et al. Taming pretrained transformers for extreme multi-label text classification[C]//Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2020: 3163-3171.
[20] VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Advances in Neural Information Processing Systems 30, 2017.
[21] 黄伟, 刘贵全. MSML-BERT模型的层级多标签文本分类方法研究[J]. 计算机工程与应用, 2022, 58(15): 191-201.
HUANG W, LIU G Q. Study on hierarchical multi-label text classification method of msml-bert model[J]. Computer Engineering and Applications, 2022, 58(15): 191-201.
[22] CHARTE F, RIVERA A J, DEL JESUS M J, et al. Addressing imbalance in multilabel classification: measures and random resampling algorithms[J]. Neurocomputing, 2015, 163: 3-16.
[23] YANG Y, LIU X. A re-examination of text categorization methods[C]//Proceedings of the 22nd Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, 1999: 42-49.
[24] LIN T Y, GOYAL P, GIRSHICK R, et al. Focal loss for dense object detection[C]//Proceedings of the 2017 IEEE International Conference on Computer Vision, 2017: 2980-2988.
[25] CUI Y, JIA M, LIN T Y, et al. Class-balanced loss based on effective number of samples[C]//Proceedings of the 2019 IEEE/CVF Conference on Cmputer Vision and Pattern Recognition, 2019: 9268-9277.
[26] WU T, HUANG Q, LIU Z, et al. Distribution-balanced loss for multi-label classification in long-tailed datasets[C]//Proceedings of the 16th European Conference on Computer Vision. Cham: Springer, 2020: 162-178.
[27] MIKOLOV T, CHEN K, CORRADO G, et al. Efficient estimation of word representations in vector space[J]. arXiv: 1301. 3781, 2013.
[28] PENNINGTON J, SOCHER R, MANNING C D. GloVe: global vectors for word representation[C]//Proceedings of the 2014 Conference on Empirical Methods in Ntural Language Processing, 2014: 1532-1543.
[29] WYDMUCH M, JASINSKA K, KUZNETSOV M, et al. A no-regret generalization of hierarchical softmax to extreme multi-label classification[C]//Advances in Neural Information Processing Systems 31, 2018.
[30] KHANDAGALE S, XIAO H, BABBAR R. Bonsai: diverse and shallow trees for extreme multi-label classification[J]. Machine Learning, 2020, 109(11): 2099-2119.
[31] JIANG T, WANG D, SUN L, et al. LightXML: transformer with dynamic negative sampling for high-performance extreme multi-label text classification[C]//Proceedings of the 35th AAAI Conference on Artificial Intelligence, 2021: 7987-7994.
[32] ZHANG R, WANG Y S, YANG Y, et al. Long-tailed extreme multi-label text classification with generated pseudo label descriptions[J]. arXiv: 2204. 00958, 2022.
[33] WANG Q, SHU H, ZHU J. GUDN: a novel guide network for extreme multi-label text classification[J]. arXiv:2201. 11582, 2022.