Judge Embeddings: Vector Representations of Legal Belief

Ash, Elliott and Chen, Daniel L. (2019) Judge Embeddings: Vector Representations of Legal Belief. Computational Analysis of Law. (In Press)

Full text not available from this repository.
Official URL: http://tse-fr.eu/pub/32431


Recent work in natural language processing represents language objects (words and documents) as dense vectors that encode the relations between those objects (Blei, 2012; Mikolov et al., 2013). These methods have recently been adapted to the analysis of human social behavior (e.g. Caliskan et al., 2017). This paper explores the vectorization of legal beliefs, with the goal of understanding judicial reasoning and the causal impacts of law. We illustrate the usefulness of these vectors in three ways. First, we show that they recover intuitive institutional connections between judges. Second, we show the vectors can be used as features in a decision prediction task. Third, we show that they can be used to measure implicit bias by judges toward women and racial minorities.

Item Type: Article
Language: English
Date: 2019
Refereed: Yes
Divisions: TSE-R (Toulouse)
Site: UT1
Date Deposited: 22 May 2018 08:56
Last Modified: 24 Jan 2019 00:17
OAI ID: oai:tse-fr.eu:32431
URI: http://publications.ut-capitole.fr/id/eprint/25832

Actions (login required)

View Item View Item