Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Chucri, Charbel, Azouz, Rami, Ott, Joachim
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2410.01736
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917793261158400
author Chucri, Charbel
Azouz, Rami
Ott, Joachim
author_facet Chucri, Charbel
Azouz, Rami
Ott, Joachim
contents Recent retrieval-augmented models enhance basic methods by building a hierarchical structure over retrieved text chunks through recursive embedding, clustering, and summarization. The most relevant information is then retrieved from both the original text and generated summaries. However, such approaches face limitations with dynamic datasets, where adding or removing documents over time complicates the updating of hierarchical representations formed through clustering. We propose a new algorithm to efficiently maintain the recursive-abstractive tree structure in dynamic datasets, without compromising performance. Additionally, we introduce a novel post-retrieval method that applies query-focused recursive abstractive processing to substantially improve context quality. Our method overcomes the limitations of other approaches by functioning as a black-box post-retrieval layer compatible with any retrieval algorithm. Both algorithms are validated through extensive experiments on real-world datasets, demonstrating their effectiveness in handling dynamic data and improving retrieval performance.
format Preprint
id arxiv_https___arxiv_org_abs_2410_01736
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Recursive Abstractive Processing for Retrieval in Dynamic Datasets
Chucri, Charbel
Azouz, Rami
Ott, Joachim
Computation and Language
Machine Learning
Recent retrieval-augmented models enhance basic methods by building a hierarchical structure over retrieved text chunks through recursive embedding, clustering, and summarization. The most relevant information is then retrieved from both the original text and generated summaries. However, such approaches face limitations with dynamic datasets, where adding or removing documents over time complicates the updating of hierarchical representations formed through clustering. We propose a new algorithm to efficiently maintain the recursive-abstractive tree structure in dynamic datasets, without compromising performance. Additionally, we introduce a novel post-retrieval method that applies query-focused recursive abstractive processing to substantially improve context quality. Our method overcomes the limitations of other approaches by functioning as a black-box post-retrieval layer compatible with any retrieval algorithm. Both algorithms are validated through extensive experiments on real-world datasets, demonstrating their effectiveness in handling dynamic data and improving retrieval performance.
title Recursive Abstractive Processing for Retrieval in Dynamic Datasets
topic Computation and Language
Machine Learning
url https://arxiv.org/abs/2410.01736