Too Legal; Didn't Read (TLDR): Summarization of Court Opinions
2023 Intermountain Engineering, Technology and Computing (IETC)
Abstract
Court opinions are dense, lengthy, and often inaccessible to non-lawyers. We explore automatic summarization of US court opinions using transformer-based language models and evaluate the resulting summaries for faithfulness, coverage, and readability. Our approach combines extractive selection with abstractive generation to produce summaries that retain legally salient holdings while remaining accessible to general readers.