Non-Parametric Memory Guidance for Multi-Document Summarization

Florian Baud, Alex Aussem
11/14/2023
cs.CL

No official code found. Best implementation: florianbaud/retrieval-augmented-mds (0 stars, last commit 2024-06-07).

Abstract

Multi-document summarization (MDS) is a difficult task in Natural Language Processing, aiming to summarize information from several documents. However, the source documents are often insufficient to obtain a qualitative summary. We propose a retriever-guided model combined with non-parametric memory for summary generation. This model retrieves relevant candidates from a database and then generates the summary considering the candidates with a copy mechanism and the source documents. The retriever is implemented with Approximate Nearest Neighbor Search (ANN) to search large databases. Our method is evaluated on the MultiXScience dataset which includes scientific articles. Finally, we discuss our results and possible directions for future work.

View on arXivView PDF

Code Implementations(1)

florianbaud/retrieval-augmented-mdsCommunity100%
00Shell, PythonJan 20, 20232 years ago

Similar Papers

Summarization-Based Document IDs for Generative Retrieval with Language Models

84% match1 repo2023

ODSum: New Benchmarks for Open Domain Multi-Document Summarization

83% match1 repo112023

LMGQS: A Large-scale Dataset for Query-focused Summarization

83% match2023

Hybrid Long Document Summarization using C2F-FAR and ChatGPT: A Practical Study

83% match2023

GreekT5: A Series of Greek Sequence-to-Sequence Models for News Summarization

83% match1 repo22023

DiffuSum: Generation Enhanced Extractive Summarization with Diffusion

82% match1 repo202023

Cite this paper

@article{baud2023nonparametric,
  title  = {Non-Parametric Memory Guidance for Multi-Document Summarization},
  author = {Florian Baud and Alex Aussem},
  year   = {2023},
  eprint = {2311.10760},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2311.10760v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2311.10760.svg)](https://codeofpaper.com/paper/2311.10760)

Discussion