Large Search Model: Redefining Search Stack in the Era of LLMs
Topics: AI (Deep Learning), LLMO, Retrieval Augmented Generation (RAG), SGE
This opinion paper, titled “Large Search Model: Redefining Search Stack in the Era of LLMs,” introduces a new conceptual framework for search engines using Large Language Models (LLMs). The framework proposes unifying search tasks into one LLM that formulates tasks as text generation problems, enhancing search quality and simplifying the search stack. The paper discusses the current search engine architecture, the benefits of LLMs, the challenges of implementing this framework, and preliminary experiments showing its feasibility.