CUED at ProbSum 2023: Hierarchical Ensemble of Summarization Models

Potsawee Manakul, Yassir Fathullah, Adian Liusie, Vyas Raina, Vatsal Raina +1 more
6/8/2023
cs.CL

No official code found. Best implementation: potsawee/hierarchical_ensemble_summ (3 stars, MIT, last commit 2023-08-09).

Abstract

In this paper, we consider the challenge of summarizing patients' medical progress notes in a limited data setting. For the Problem List Summarization (shared task 1A) at the BioNLP Workshop 2023, we demonstrate that Clinical-T5 fine-tuned to 765 medical clinic notes outperforms other extractive, abstractive and zero-shot baselines, yielding reasonable baseline systems for medical note summarization. Further, we introduce Hierarchical Ensemble of Summarization Models (HESM), consisting of token-level ensembles of diverse fine-tuned Clinical-T5 models, followed by Minimum Bayes Risk (MBR) decoding. Our HESM approach lead to a considerable summarization performance boost, and when evaluated on held-out challenge data achieved a ROUGE-L of 32.77, which was the best-performing system at the top of the shared task leaderboard.

View on arXivView PDF

Code Implementations(1)

MIT

Hierarchical Ensemble of Summarization Models

30Python, Jupyter NotebookJun 5, 20233 years agoMIT

Similar Papers

Cite this paper

@article{manakul2023cued,
  title  = {CUED at ProbSum 2023: Hierarchical Ensemble of Summarization Models},
  author = {Potsawee Manakul and Yassir Fathullah and Adian Liusie and Vyas Raina and Vatsal Raina and Mark Gales},
  year   = {2023},
  eprint = {2306.05317},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2306.05317v1}
}

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/2306.05317.svg)](https://codeofpaper.com/paper/2306.05317)

Discussion