For my self-proposed final project in Natural Language Processing, I explored ways to make state-of-the-art NLP conventions more performant on legal text classification.

Legal text presents unique challenges for machine learning models trained primarily on general-purpose corpora. The vocabulary is specialized, the sentence structures are often convoluted, and the contextual meaning can depend heavily on domain-specific knowledge.

This project highlighted just how much work is needed in the application of ML to legal use cases—a gap that continues to be relevant as more legal tech startups attempt to apply AI to the practice of law.

• View the project on GitHub

🍅

← Back to all posts