Twitxy Saved

Fangzheng Tian

@DanielTian97

A PhD student at University of Glasgow. Working on Information Retrieval and Natural Language Processing.
Glasgow, Scotland Joined November 2021
114 Following 53 Followers 43 Posts

Latest posts 5

Fangzheng Tian @DanielTian97
This paper is an initial discussion about the correlation between RAG context’s topical relevance and their ‘utility’ for the downstream task. The experiment framework in this paper can be used for further, more in-depth investigations. Link: doi.org/10.48550/arXiv….
arxiv.org Is Relevance Propagated from Retriever to Generator in RAG? Retrieval Augmented Generation (RAG) is a framework for incorporating external knowledge, usually in the form of a set of documents retrieved from a collection, as a part of a prompt to a large...
Topical Relevance is *not* what all you need for improved RAG performance.
We show this for abstractive query-based summarization on TREC DL.
Work by student @DanielTian97 co-supervised by @craig_macdonald
To appear: #ecir2025
Arxiv: arxiv.org/abs/2502.15025
View quoted post on X