📄 中文摘要
HybridRAG是一种新颖且实用的检索增强生成（RAG）框架，旨在提高聊天机器人的响应准确性和速度。该框架首先通过光学字符识别（OCR）和布局分析技术，处理包含复杂布局（如文本、表格和图形）的原始非结构化PDF文档，并将其转换为层次化文本块。随后，HybridRAG在查询时利用预生成的问答对这些文本块进行检索和生成，从而克服了传统RAG方法在处理非结构化数据时的局限性。这一方法的提出为在现实世界的聊天机器人应用中提供了更强的适用性和效率。
HybridRAG：基于预生成问答的原始非结构化文档的实用LLM聊天机器人框架
📄 English Summary
HybridRAG: A Practical LLM-based ChatBot Framework based on Pre-Generated Q&A over Raw Unstructured Documents
HybridRAG is a novel and practical Retrieval-Augmented Generation (RAG) framework aimed at enhancing the accuracy and speed of chatbot responses. The framework first ingests raw, unstructured PDF documents with complex layouts (including text, tables, and figures) through Optical Character Recognition (OCR) and layout analysis, converting them into hierarchical text chunks. Subsequently, HybridRAG utilizes pre-generated Q&A to perform retrieval and generation on these text chunks at query time, addressing the limitations of traditional RAG methods when dealing with unstructured data. This approach offers greater applicability and efficiency for real-world chatbot applications.