Legal Extractive Summarization of U.S. Court Opinions
This work addresses making U.S. court opinions more accessible to the general public, representing incremental progress in legal text summarization.
The paper tackled legal extractive summarization of U.S. court opinions using a dataset of 430K annotated opinions, finding that the reinforcement-learning-based MemSum model outperformed transformer-based models in automated metrics and was effective in capturing key points according to expert human evaluation.
This paper tackles the task of legal extractive summarization using a dataset of 430K U.S. court opinions with key passages annotated. According to automated summary quality metrics, the reinforcement-learning-based MemSum model is best and even out-performs transformer-based models. In turn, expert human evaluation shows that MemSum summaries effectively capture the key points of lengthy court opinions. Motivated by these results, we open-source our models to the general public. This represents progress towards democratizing law and making U.S. court opinions more accessible to the general public.