Legal Document Similarity-Metric Learning Task

A Legal Document Similarity-Metric Learning Task is a domain-specific document similarity-metric learning task focused on learning a legal document similarity model that can assess the semantic similarity between legal documents.

• Context:
  • Input: Paired Legal Documents such as similar and dissimilar cases.
  • output: A Legal Document Similarity Model that outputs a score quantifying the degree of semantic similarity.
  • measure: legal experts' judgment of document similarity.
  • Both textual similarity and citation network similarity provide useful signals.
  • It can incorporate knowledge of legal domain into the similarity metric.
  • It can enable Legal Information Retrieval like case law search.
  • ...
• Example(s):
  • A Legal Case Similarity Metric Learning Task to learn a legal case similarity model based on relevant vs non-relevant cases.
  • A Legal Statute Similarity Metric Learning Task to learn a statute similarity model based on related vs unrelated statutes.
  • A Contract Similarity Metric Learning Task to learn a contract similarity model based on similar and dissimilar contract pairs.
  • A Contract Clause Similarity Metric Learning Task to learn a contract clause similarity model based on related and unrelated contract clauses.
  • A Patent Document Similarity Metric Learning Task to learn a patent document similarity model based on related vs unrelated patent documents.
  • A Legal Term Similarity Metric Learning Task to learn a legal term similarity model based on related vs unrelated legal terms.
  • A Legal Question Similarity Metric Learning Task to learn a legal question similarity model based on duplicate and non-duplicate legal questions.
  • A Legal Named Entity Similarity Metric Learning Task to learn a legal named entity similarity model based on related vs unrelated legal named entities.
  • ...
• Counter-Example(s):
  • Scientific Paper Similarity Metric Learning Task - focuses on academic papers rather than legal documents.
  • General Document Similarity Learning - not specialized to legal domain.
  • Text Classification Task - predicts categories rather than learning text similarities.
  • ...
• See: Legal Information Retrieval, Legal Text Mining.

---

---

References

2022

• (Bhattacharya et al., 2022) ⇒ Prajjwal Bhattacharya, Kripabandhu Ghosh, Arpan Pal, and Saptarshi Ghosh. (2022). “Legal case document similarity: You need both network and text." In: Information Processing & Management 58.4: 102580.
  • QUOTE: "...We incorporate domain knowledge for legal document similarity into Hier-... legal document similarity. Both textual and network similarity provide important signals for legal case similarity..."
  • NOTE: Examines combining textual similarity with citation network patterns for effective Legal Case Document Similarity Metric Learning.

2022

• (Bi et al., 2022) ⇒ Shuyang Bi, Zaeem Ali, Miao Wang, Tianming Wu, and Guoqing Qi. (2022). “Learning heterogeneous graph embedding for Chinese legal document similarity." In: Knowledge-Based Systems 244: 108860.
  • QUOTE: "...Measuring the similarity between legal documents to find prior documents from a massive collection that are similar to a current document is an essential component in legal assistant..."
  • NOTE: Proposes graph embedding techniques to learn legal document similarity for applications like prior case retrieval.

2020

• (Bhattacharya et al., 2020a) ⇒ Prajjwal Bhattacharya, Kripabandhu Ghosh, Arpan Pal, and Saptarshi Ghosh. (2020). “Methods for computing legal document similarity: A comparative study." arXiv preprint arXiv:2005.06032.
  • ABSTRACT: Computing similarity between two legal documents is an important and challenging task in the domain of Legal Information Retrieval. Finding similar legal documents has many applications in downstream tasks, including prior-case retrieval, recommendation of legal articles, and so on. Prior works have proposed two broad ways of measuring similarity between legal documents - analyzing the precedent citation network, and measuring similarity based on textual content similarity measures. But there has not been a comprehensive comparison of these existing methods on a common platform. In this paper, we perform the first systematic analysis of the existing methods. In addition, we explore two promising new similarity computation methods - one text-based and the other based on network embeddings, which have not been considered till now.
  • NOTE: Compares different approaches to learn Legal Document Similarity combining textual signals and citation patterns.

2020

• (Wagh & Anand, 2020) ⇒ Rahul S. Wagh and Deepali Anand. (2020). “Legal document similarity: a multi-criteria decision-making perspective." In: PeerJ Computer Science 6: e317.
  • QUOTE: "...present the legal document similarity... We specifically focus on the problem of finding the similarity..."
  • NOTE: Proposes a multi-criteria approach to Legal Document Similarity Metric Learning combining different signals.

---