English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Lack of factual correctness is an issue that still plagues state-of-the-artsummarization systems despite their impressive progress on generating seeminglyfluent summaries. In this paper, we show that factual inconsistency can becaused by irrelevant parts of the input text, which act as confounders. To thatend, we leverage information-theoretic measures of causal effects to quantifythe amount of confounding and precisely quantify how they affect thesummarization performance. Based on insights derived from our theoreticalresults, we design a simple multi-task model to control such confounding byleveraging human-annotated relevant sentences when available. Crucially, wegive a principled characterization of data distributions where such confoundingcan be large thereby necessitating the use of human annotated relevantsentences to generate factual summaries. Our approach improves faithfulnessscores by 20\% over strong baselines on AnswerSumm\citep{fabbri2021answersumm}, a conversation summarization dataset where lackof faithfulness is a significant issue due to the subjective nature of thetask. Our best method achieves the highest faithfulness score while alsoachieving state-of-the-art results on standard metrics like ROUGE and METEOR.We corroborate these improvements through human evaluation.

Improving Faithfulness of Abstractive Summarization by Controlling Confounding Effect of Irrelevant Sentences