Deeper Summarisation: The Second Time Around

Teufel, Simone

doi:10.1007/978-3-319-75487-1_44

Cited by 2 publications

(3 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Traditional summarization approaches sought to provide more control over properties of generated summaries such as their informativeness (Jones, 1993;Carbonell & Goldstein, 1998;Nenkova & McKeown, 2011;Lloret, 2012;Teufel, 2016), non-redundancy (Carbonell & Goldstein, 1998), or discourse organization (Marcu, 1998;Christensen et al, 2013). However, more modern approaches mostly employ neural end-to-end models (Cheng & Lapata, 2016;Lewis et al, 2020;J.…”

Section: Informativeness Redundancy and Cohesionmentioning

confidence: 99%

On the Trade-off between Redundancy and Cohesiveness in Extractive Summarization

Cardenas,

Galle,

Cohen

2024

jair

View full text Add to dashboard Cite

Extractive summaries are usually presented as lists of sentences with no expected cohesion between them and with plenty of redundant information if not accounted for. In this paper, we investigate the trade-offs incurred when aiming to control for inter-sentential cohesion and redundancy in extracted summaries, and their impact on their informativeness. As case study, we focus on the summarization of long, highly redundant documents and consider two optimization scenarios, reward-guided and with no supervision. In the reward-guided scenario, we compare systems that control for redundancy and cohesiveness during sentence scoring. In the unsupervised scenario, we introduce two systems that aim to control all three properties --informativeness, redundancy, and cohesiveness-- in a principled way. Both systems implement a psycholinguistic theory that simulates how humans keep track of relevant content units and how cohesiveness and non-redundancy constraints are applied in short-term memory during reading. Extensive automatic and human evaluations reveal that systems optimizing for --among other properties-- cohesiveness are capable of better organizing content in summaries compared to systems that optimize only for redundancy, while maintaining comparable informativeness. We find that the proposed unsupervised systems manage to extract highly cohesive summaries across varying levels of document redundancy, although sacrificing informativeness in the process. Finally, we lay evidence as to how simulated cognitive processes impact the trade-off between the analysed summary properties.

show abstract

Section: Informativeness Redundancy and Cohesionmentioning

confidence: 99%

On the Trade-off between Redundancy and Cohesiveness in Extractive Summarization

Cardenas,

Galle,

Cohen

2024

jair

View full text Add to dashboard Cite

show abstract

“…Modern approaches mostly employ neural end-to-end models (Cheng & Lapata, 2016;Lewis et al, 2020;Zhang et al, 2020), meaning that crucial intermediate steps such as content planning or selection are not explicitly modeled, in contrast to traditional summarization approaches (Jones, 1993;Carbonell & Goldstein, 1998;Nenkova et al, 2011;Lloret, 2012;Teufel, 2016). Recent efforts have demonstrated that accounting for planning helps dealing with coherence of final summaries (Goldfarb-Tarrant et al, 2020;Sharma et al, 2019;Hua et al, 2021), whereas explicit content selection modules can be tailored to tackle problems such as factuality (Cao et al, 2018;Maynez et al, 2020), coverage (Kedzie et al, 2018;Puduppully et al, 2019;Wiseman et al, 2017), and redundancy Jia et al, 2021;Bi et al, 2021).…”

Section: Content Selection Redundancy and Length Controlmentioning

confidence: 99%

“…Even though recent advances in machine learning brought promising results -mostly involving increasingly larger neural networks-in all stages of the summarization pipeline, core challenges such as redundancy during content selection (Xiao & Carenini, 2020;Jia, Cao, Fang, Zhou, Fang, Liu, & Wang, 2021) and coherence of produced summaries (Cao, Wei, Li, & Li, 2018;Sharma, Li, & Wang, 2019;Hua, Sreevatsa, & Wang, 2021) remain critically open (Teufel, 2016;Gatt & Krahmer, 2018). Notably, Xiao and Carenini (2020) reported that modern extractive summarization systems are prone to produce highly redundant excerpts when redundancy is not explicitly accounted for.…”

Section: Introductionmentioning

confidence: 99%

On the Trade-off between Redundancy and Local Coherence in Summarization

Cardenas¹,

Gallé²,

Cohen³

2022

Preprint

View full text Add to dashboard Cite

Extractive summarization systems are known to produce poorly coherent and, if not accounted for, highly redundant text. In this work, we tackle the problem of summary redundancy in unsupervised extractive summarization of long, highly-redundant documents. For this, we leverage a psycholinguistic theory of human reading comprehension which directly models local coherence and redundancy. Implementing this theory, our system operates at the proposition level and exploits properties of human memory representations to rank similarly content units that are coherent and non-redundant, hence encouraging the extraction of less redundant final summaries. Because of the impact of the summary length on automatic measures, we control for it by formulating content selection as an optimization problem with soft constraints in the budget of information retrieved. Using summarization of scientific articles as a case study, extensive experiments demonstrate that the proposed systems extract consistently less redundant summaries across increasing levels of document redundancy, whilst maintaining comparable performance (in terms of relevancy and local coherence) against strong unsupervised baselines according to automated evaluations.

show abstract

Deeper Summarisation: The Second Time Around

Cited by 2 publications

References 32 publications

On the Trade-off between Redundancy and Cohesiveness in Extractive Summarization

On the Trade-off between Redundancy and Cohesiveness in Extractive Summarization

On the Trade-off between Redundancy and Local Coherence in Summarization

Contact Info

Product

Resources

About