English

New Methods & Metrics for LFQA tasks

Computation and Language 2021-12-28 v1 Machine Learning

Abstract

Long-form question answering (LFQA) tasks require retrieving the documents pertinent to a query, using them to form a paragraph-length answer. Despite considerable progress in LFQA modeling, fundamental issues impede its progress: i) train/validation/test dataset overlap, ii) absence of automatic metrics and iii) generated answers not being "grounded" in retrieved documents. This work addresses every one these critical bottlenecks, contributing natural language inference/generation (NLI/NLG) methods and metrics that make significant strides to their alleviation.

Keywords

Cite

@article{arxiv.2112.13432,
  title  = {New Methods & Metrics for LFQA tasks},
  author = {Suchismit Mahapatra and Vladimir Blagojevic and Pablo Bertorello and Prasanna Kumar},
  journal= {arXiv preprint arXiv:2112.13432},
  year   = {2021}
}

Comments

8 pages, 8 figures

R2 v1 2026-06-24T08:31:59.438Z