Incorporating hierarchy into text encoder
Web(2024) demonstrate that encoding the holistic la-bel structure directly by a structure encoder can further improve performance. Following this re-search, a bunch of models try to study how the hierarchy should interact with the text. BothChen et al.(2024) andChen et al.(2024) embed word and label hierarchy jointly in a same space.Deng WebHierarchical text classification (HTC) aims to cate-gorize a text into a set of labels with a structured class hierarchy (commonly modeled as a tree) (Silla and Freitas,2011). HTC is a multi-label text clas-sification problem, where the classification result corresponds to one or more paths of the hierarchy (Zhou et al.,2024).
Incorporating hierarchy into text encoder
Did you know?
WebMay 5, 2024 · Existing methods encode label hierarchy in a global view, where label hierarchy is treated as the static hierarchical structure containing all labels. Since global hierarchy is static and... WebJan 1, 2024 · In this paper, we propose a L abel-based A ttention for H ierarchical Multi-label Text C lassification Neural N etwork ( LA-HCN), where the novel label-based attention …
WebThis work has been accepted as the long paper "Incorporating Hierarchy into Text Encoder: a Contrastive Learning Approach for Hierarchical Text Classification" in ACL 2024. … Web本文提出的模型叫做Hierarchy-Guided Contrastive Learning Representation（HGCLR） for HTC，主要包括四个模块： 「Text Encoder」 ：主要负责text的编码（CLS的），这里文本的编码会用于预测本样本的所有label，得到一个output并计算loss，这个text encoder也是最终用于预测阶段的模型。 「Graph Encoder」 ：主要是利用Graphormer对标签层次结构进 …
WebSep 24, 2024 · Incorporating Hierarchy into Text Encoder: a Contrastive Learning Approach for Hierarchical Text Classification 1 code implementation • ACL 2024 Hierarchical text classification is a challenging subtask of multi-label classification due to its complex label hierarchy. Classification Contrastive Learning +3 65 Paper Code Webrectly embed the hierarchy into a text en-coder. During training, HGCLR constructs positive samples for input text under the guid-ance of the label hierarchy. By pulling to-gether the …
WebDue to the limitation of text encoding length by the popular encoders (e.g., BERT), existing studies [5–7] usually cut a long document into segments, and then take each segment as the input. This will lead to the macro discourse structure and generation model can not be fully utilized. To solve this issue, we
Web2 days ago · Instead of modeling them separately, in this work, we propose Hierarchy-guided Contrastive Learning (HGCLR) to directly embed the hierarchy into a text encoder. During … shanks red lightningWebMar 8, 2024 · Hierarchical text classification is a challenging subtask of multi-label classification due to its complex label hierarchy. Existing methods encode text and label hierarchy separately and mix their representations for classification, where the hierarchy remains unchanged for all input text. shanks refrigeration marion paWebMar 9, 2024 · The Hierarchy-guided Contrastive Learning (HGCLR) [ 24] directly embeds the hierarchy into a text encoder. In [ 25 ], authors propose a contrastive learning method to obtain representations for text classification based on monolingual embeddings of BERT. shanks red filmWeb2024. Incorporating hierarchy into text encoder: a contrastive learning approach for hierarchical text classification. Z Wang, P Wang, L Huang, X Sun, H Wang. arXiv preprint arXiv:2203.03825. , 2024. 5. 2024. Zero-shot Cross-lingual Transfer of Prompt-based Tuning with a Unified Multilingual Prompt. shanks red hair piratesWebMay 5, 2024 · To model semantic information along with hierarchical information, Zhou et al. proposes hierarchy-aware multi-label attention. Chen et al. reformulates it as a matching … polymicrobial infection icd 10 codeWebits complex label hierarchy. Existing meth-ods encode text and label hierarchy separately andmixtheirrepresentationsforclassication, where the hierarchy remains unchanged for all … shanks refuseWebDec 25, 2024 · This paper proposes a contrastive learning approach for HTC through generating high quality positive samples using the taxonomic hierarchy. The hierarchy is modeled with Graphormer and trained end-to-end with positive sample generation under a contrastive loss and a classification loss. Comments Wow, amazing! It’s hard to believe … polymicrobial infection treatment