LH-Mix: Local Hierarchy Correlation Guided Mixup over Hierarchical Prompt Tuning
This work addresses hierarchical text classification, a domain-specific problem in natural language processing, with an incremental improvement over existing prompt tuning methods.
The paper tackles hierarchical text classification by addressing the limitation of ignoring implicit sibling/peer correlations in local hierarchies, proposing LH-Mix which integrates local hierarchies into prompts and applies a novel Mixup guided by local hierarchy correlation. The model achieves remarkable performance on three widely-used datasets, though no concrete numbers are provided in the abstract.
Hierarchical text classification (HTC) aims to assign one or more labels in the hierarchy for each text. Many methods represent this structure as a global hierarchy, leading to redundant graph structures. To address this, incorporating a text-specific local hierarchy is essential. However, existing approaches often model this local hierarchy as a sequence, focusing on explicit parent-child relationships while ignoring implicit correlations among sibling/peer relationships. In this paper, we first integrate local hierarchies into a manual depth-level prompt to capture parent-child relationships. We then apply Mixup to this hierarchical prompt tuning scheme to improve the latent correlation within sibling/peer relationships. Notably, we propose a novel Mixup ratio guided by local hierarchy correlation to effectively capture intrinsic correlations. This Local Hierarchy Mixup (LH-Mix) model demonstrates remarkable performance across three widely-used datasets.