English

Ranking-based Fusion Algorithms for Extreme Multi-label Text Classification (XMTC)

Information Retrieval 2025-07-08 v1

Abstract

In the context of Extreme Multi-label Text Classification (XMTC), where labels are assigned to text instances from a large label space, the long-tail distribution of labels presents a significant challenge. Labels can be broadly categorized into frequent, high-coverage \textbf{head labels} and infrequent, low-coverage \textbf{tail labels}, complicating the task of balancing effectiveness across all labels. To address this, combining predictions from multiple retrieval methods, such as sparse retrievers (e.g., BM25) and dense retrievers (e.g., fine-tuned BERT), offers a promising solution. The fusion of \textit{sparse} and \textit{dense} retrievers is motivated by the complementary ranking characteristics of these methods. Sparse retrievers compute relevance scores based on high-dimensional, bag-of-words representations, while dense retrievers utilize approximate nearest neighbor (ANN) algorithms on dense text and label embeddings within a shared embedding space. Rank-based fusion algorithms leverage these differences by combining the precise matching capabilities of sparse retrievers with the semantic richness of dense retrievers, thereby producing a final ranking that improves the effectiveness across both head and tail labels.

Keywords

Cite

@article{arxiv.2507.03761,
  title  = {Ranking-based Fusion Algorithms for Extreme Multi-label Text Classification (XMTC)},
  author = {Celso França and Gestefane Rabbi and Thiago Salles and Washington Cunha and Leonardo Rocha and Marcos André Gonçalves},
  journal= {arXiv preprint arXiv:2507.03761},
  year   = {2025}
}
R2 v1 2026-07-01T03:47:10.235Z