Structure-Aware Legal Case Retrieval

The School of EECS is hosting the following HDR Progress Review 3 Seminar:

Structure-Aware Legal Case Retrieval

Speaker: Yanran Tang

Abstract:

Legal case retrieval (LCR) is a specialised and essential task that aims to identify relevant cases given a query case. For legal practitioners, such as judges and lawyers, retrieval systems offer a far more efficient alternative to manually searching through large volumes of legal documents. Existing LCR approaches can be broadly categorised into two streams: statistical retrieval models, which rely on term frequency similarity, and neural models, which encode cases into dense representations for nearest neighbour search.

However, existing approaches often underutilise domain-specific legal knowledge that is essential for accurately determining case relevance. To address this limitation, this seminar presents a structure-aware perspective on LCR, grounded in four key aspects: intra-case structural relations, structural legal semantic representations, inter-case connectivity relationships, and legal structural knowledge distillation. By explicitly incorporating these elements, the proposed approaches aim to improve representation learning capability, retrieval accuracy, and retrieval efficiency in LCR systems.

Bio:

Yanran Tang is a PhD candidate at the School of Electrical Engineering and Computer Science, The University of Queensland. She holds both LL.B. and LL.M. degrees. Her research focuses on structural representation learning in the legal domain, with an emphasis on improving legal case retrieval.