03 Sep 2024
学术报告：Robust and Efficient Retrieval-Augmented LLM
时间：2024年9月3日，星期三
地点：湖北省武汉市，武汉大学，计算机学院B405会议室
联系人：李清安 qingan@whu.edu.cn
| Time | Title & Speaker |
|---|---|
| 14:00 - 15:30 | Robust and Efficient Retrieval-Augmented LLM Dr. Shangyu Wu, City University of Hong Kong (香港城市大学) |
Large language models (LLMs) have demonstrated great success in various fields, benefiting from their huge amount of parameters that store knowledge. However, LLMs still suffer from several key issues, such as hallucination problems, knowledge update issues, and lacking domain-specific expertise. The appearance of retrieval-augmented generation (RAG), which leverages an external knowledge database to augment LLMs, makes up those drawbacks of LLMs. This talk will first give a complete review of significant RAG techniques, including retrievers and retrieval fusions. This talk will introduce our two works about making a good trade-off between model performance and model efficiency by incorporating retrieval-based augmentations.
大型语言模型（LLMs）在各个领域都取得了巨大成功，这得益于它们庞大的参数量，能够存储知识。然而，LLMs仍然存在一些问题，例如幻觉问题、知识更新问题以及缺乏特定领域的专业知识。检索增强生成（RAG）的出现，利用外部知识数据库来增强LLMs，弥补了LLMs的这些不足。本次演讲将首先全面回顾重要的RAG技术，包括检索器和检索融合。本次演讲将介绍我们的两项工作，它们通过引入基于检索的增强，实现了模型性能和模型效率之间的良好权衡。