HiLight: A Hierarchy-aware Light Global Model with Hierarchical Local ConTrastive Learning
This work addresses efficiency and stability problems in hierarchical text classification, which is important for applications like document categorization, but it is incremental as it builds on existing methods.
The paper tackles the scalability and collapse issues in hierarchical text classification by proposing HiLight, a lightweight global model that replaces the structure encoder with hierarchical local contrastive learning, achieving competitive performance on benchmark datasets.
Hierarchical text classification (HTC) is a special sub-task of multi-label classification (MLC) whose taxonomy is constructed as a tree and each sample is assigned with at least one path in the tree. Latest HTC models contain three modules: a text encoder, a structure encoder and a multi-label classification head. Specially, the structure encoder is designed to encode the hierarchy of taxonomy. However, the structure encoder has scale problem. As the taxonomy size increases, the learnable parameters of recent HTC works grow rapidly. Recursive regularization is another widely-used method to introduce hierarchical information but it has collapse problem and generally relaxed by assigning with a small weight (ie. 1e-6). In this paper, we propose a Hierarchy-aware Light Global model with Hierarchical local conTrastive learning (HiLight), a lightweight and efficient global model only consisting of a text encoder and a multi-label classification head. We propose a new learning task to introduce the hierarchical information, called Hierarchical Local Contrastive Learning (HiLCL). Extensive experiments are conducted on two benchmark datasets to demonstrate the effectiveness of our model.