Maximal Marginal Relevance

Maximal Marginal Relevance (MMR) is a technique used to select a subset of information, often text passages or sentences, that balances relevance to a given query with diversity among the selected items. Current research focuses on improving MMR's application in retrieval augmented generation (RAG) for large language models, addressing limitations like the fixed context window size and the parameter sensitivity of traditional MMR implementations. This work aims to enhance the efficiency and effectiveness of information retrieval and summarization tasks, impacting areas such as question answering, multi-document summarization, and legal document processing.