Hierarchical MixUp Multi-label Classification with Imbalanced Interdisciplinary Research Proposals

TL;DR

This paper introduces H-MixUp, a hierarchical multi-label classification framework that effectively models interdisciplinary research proposals by addressing hierarchical labels, heterogeneous textual semantics, and data imbalance using advanced mixup techniques and neural architectures.

Contribution

It presents a novel hierarchical mixup multi-label classification framework combining transformer, GCN, and multiple mixup strategies to handle complex interdisciplinary proposal data.

Findings

Improved classification accuracy on interdisciplinary proposals

Effective handling of hierarchical label structures

Balanced performance across data imbalance scenarios

Abstract

Funding agencies are largely relied on a topic matching between domain experts and research proposals to assign proposal reviewers. As proposals are increasingly interdisciplinary, it is challenging to profile the interdisciplinary nature of a proposal, and, thereafter, find expert reviewers with an appropriate set of expertise. An essential step in solving this challenge is to accurately model and classify the interdisciplinary labels of a proposal. Existing methodological and application-related literature, such as textual classification and proposal classification, are insufficient in jointly addressing the three key unique issues introduced by interdisciplinary proposal data: 1) the hierarchical structure of discipline labels of a proposal from coarse-grain to fine-grain, e.g., from information science to AI to fundamentals of AI. 2) the heterogeneous semantics of various main textual parts that play different roles in a proposal; 3) the number of proposals is imbalanced between non-interdisciplinary and interdisciplinary research. Can we simultaneously address the three issues in understanding the proposal's interdisciplinary nature? In response to this question, we propose a hierarchical mixup multiple-label classification framework, which we called H-MixUp. H-MixUp leverages a transformer-based semantic information extractor and a GCN-based interdisciplinary knowledge extractor for the first and second issues. H-MixUp develops a fused training method of Wold-level MixUp, Word-level CutMix, Manifold MixUp, and Document-level MixUp to address the third issue.