QABISAR: Query-Article Bipartite Interactions for Statutory Article Retrieval

TL;DR

QABISAR tackles statutory article retrieval by modeling many-to-many query–article relationships through a query–article bipartite graph enriched with statute hierarchy. It combines a dense bi-encoder for efficient retrieval with a graph encoder that captures multi-faceted interactions via graph attention, and uses knowledge distillation to transfer graph-derived semantics to unseen queries. The two-stage training, along with KD, yields superior retrieval performance on the BSARD dataset and highlights the value of integrating structured statute topology with bipartite interactions. This approach enhances recall while paving the way for practical SAR systems that can support accessible legal information and pre-fetching for downstream QA tasks.

Abstract

In this paper, we introduce QABISAR, a novel framework for statutory article retrieval, to overcome the semantic mismatch problem when modeling each query-article pair in isolation, making it hard to learn representation that can effectively capture multi-faceted information. QABISAR leverages bipartite interactions between queries and articles to capture diverse aspects inherent in them. Further, we employ knowledge distillation to transfer enriched query representations from the graph network into the query bi-encoder, to capture the rich semantics present in the graph representations, despite absence of graph-based supervision for unseen queries during inference. Our experiments on a real-world expert-annotated dataset demonstrate its effectiveness.

Figures (2)

• Figure 1: Ablation Study on QABISAR
• Figure 2: Effect of various distillation strategies