LFRAG: Layout-oriented Fine-grained Retrieval-Augmented Generation on Multimodal Document Understanding

Collection of recent research papers (arXiv, 25 May 2026) covering advances in retrieval-augmented generation (RAG), multimodal document understanding, recommender systems, model alignment/optimization, and auditing generative search engines. Key works: LFRAG—block-level, layout-aware retrieval and LFDocQA benchmark for fine-grained multimodal RAG; RAG4Outcome—RAG for prognostic prediction in chronic osteomyelitis using multimodal clinical data; “AI-Friendly LaTeX”—pipeline for turning LaTeX source into RAG-friendly chunks; AKT-Rec and large-scale generative recommender studies addressing long