Translating the Law

Overview

• Description: Translating the Law is an exploration of NLP and deep learning techniques to summarise and simplify UK Supreme Court cases. The main frameworks and tools used were:

1. Scraping: SCRAPY
2. Summarisation model: BART
3. Q&A model: BERT
4. Case search: Elasticsearch
5. Clustering: top2vec
6. Frontend: Streamlit
7. Containerisation & composition: Docker and GCP

The main data source was the UK Supreme Court

Examples

• The front end

• An example model generated summary

• The search page

• The summarisation page