Umeå universitets logga

umu.sePublikationer
Ändra sökning
RefereraExporteraLänk till posten
Permanent länk

Direktlänk
Referera
Referensformat
  • apa
  • ieee
  • vancouver
  • Annat format
Fler format
Språk
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Annat språk
Fler språk
Utmatningsformat
  • html
  • text
  • asciidoc
  • rtf
PromptStream: self-supervised news story discovery using topic-aware article representations
Umeå universitet, Teknisk-naturvetenskapliga fakulteten, Institutionen för datavetenskap.ORCID-id: 0000-0002-6791-8284
Umeå universitet, Teknisk-naturvetenskapliga fakulteten, Institutionen för datavetenskap. Aeterna Labs, Umeå, Sweden.ORCID-id: 0000-0002-4366-7863
Aeterna Labs, Umeå, Sweden.
2024 (Engelska)Ingår i: Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024) / [ed] Nicoletta Calzolari; Min-Yen Kan; Veronique Hoste; Alessandro Lenci; Sakriani Sakti; Nianwen Xue, ELRA Language Resource Association , 2024, s. 13222-13232Konferensbidrag, Publicerat paper (Refereegranskat)
Abstract [en]

Given the importance of identifying and monitoring news stories within the continuous flow of news articles, this paper presents PromptStream, a novel method for unsupervised news story discovery. In order to identify coherent and comprehensive stories across the stream, it is crucial to create article representations that incorporate as much topic-related information from the articles as possible. PromptStream constructs these article embeddings using cloze-style prompting. These representations continually adjust to the evolving context of the news stream through self-supervised learning, employing a contrastive loss and a memory of the most confident article-story assignments from the most recent days. Extensive experiments with real news datasets highlight the notable performance of our model, establishing a new state of the art. Additionally, we delve into selected news stories to reveal how the model’s structuring of the article stream aligns with story progression.

Ort, förlag, år, upplaga, sidor
ELRA Language Resource Association , 2024. s. 13222-13232
Serie
International conference on computational linguistics, ISSN 2951-2093
Nyckelord [en]
news story discovery, online clustering, data stream, contrastive learning, cloze-style prompting, article embedding
Nationell ämneskategori
Data- och informationsvetenskap
Forskningsämne
datorlingvistik; datalogi
Identifikatorer
URN: urn:nbn:se:umu:diva-222533ISI: 001612977900380Scopus ID: 2-s2.0-85195905320ISBN: 978-2-493814-10-4 (tryckt)OAI: oai:DiVA.org:umu-222533DiVA, id: diva2:1845888
Konferens
The 2024 joint international conference on computational linguistics, language resources and evaluation (LREC-COLING 2024), Torino, Italy, May 20-25, 2024
Anmärkning

Also part of series: LREC proceedings, ISSN: 2522-2686

Tillgänglig från: 2024-03-20 Skapad: 2024-03-20 Senast uppdaterad: 2026-11-12Bibliografiskt granskad
Ingår i avhandling
1. Deep learning for news topic identification in limited supervision and unsupervised settings
Öppna denna publikation i ny flik eller fönster >>Deep learning for news topic identification in limited supervision and unsupervised settings
2024 (Engelska)Doktorsavhandling, sammanläggning (Övrigt vetenskapligt)
Alternativ titel[sv]
Djup maskininlärning för identifiering av nyhetsämnen i inlärningssituationer med begränsad eller ingen övervakning
Abstract [en]

In today's world, following news is crucial for decision-making and staying informed. With the growing volume of daily news, automated processing is essential for timely insights and in aiding individuals and corporations in navigating the complexities of the information society. Another use of automated processing is contextual advertising, which addresses privacy concerns associated with cookie-based advertising by placing ads solely based on web page content, without tracking users or their online behavior. Therefore, accurately determining and categorizing page content is crucial for effective ad placements. The news media, heavily reliant on advertising to sustain operations, represent a substantial market for contextual advertising strategies.

Inspired by these practical applications and the advancements in deep learning over the past decade, this thesis mainly focuses on using deep learning for categorizing news articles into topics of varying granularity. Considering the dynamic nature of these applications and the limited availability of relevant labeled datasets for training models, the thesis emphasizes developing methods that can be trained effectively using unlabeled or partially labeled data. It proposes semi-supervised text classification models for categorizing datasets into predefined coarse-grained topics, where only a few labeled examples exist for each topic, while the majority of the dataset remains unlabeled. Furthermore, to better explore coarse-grained topics within news archives and streams and overcome the limitations of predefined topics in text classification the thesis suggests deep clustering approaches that can be trained in unsupervised settings. 

Moreover, to address the identification of fine-grained topics, the thesis introduces a novel story discovery model for monitoring event-based topics in multi-source news streams. Given that online news reporting often incorporates diverse modalities like text, images, video, and audio to convey information, the thesis finally initiates an investigation into the synergy between textual and visual elements in news article analysis. To achieve this objective, a text-image dataset was annotated, and a baseline was established for event-topic discovery in multimodal news streams. While primarily intended for news monitoring and contextual advertising, the proposed models can, more generally, be regarded as novel approaches in semi-supervised text classification, deep clustering, and news story discovery. Comparison with state-of-the-art baseline models demonstrates their effectiveness in addressing the respective objectives.

Ort, förlag, år, upplaga, sidor
Umeå: Umeå University, 2024. s. 62
Serie
Report / UMINF, ISSN 0348-0542 ; 24.04
Nyckelord
Topic Identification, Data Clustering, News Stream Clustering, Semi-Supervised Learning, Unsupervised Learning, Event Topics, News Stories, Multimodal News, Document Classification, Document Clustering, Deep Learning, Deep Clustering, Pre-trained Language Models
Nationell ämneskategori
Datavetenskap (datalogi)
Identifikatorer
urn:nbn:se:umu:diva-222534 (URN)9789180703420 (ISBN)9789180703437 (ISBN)
Disputation
2024-04-16, MIT.A.121, MIT-huset, Umeå, 13:15 (Engelska)
Opponent
Handledare
Tillgänglig från: 2024-03-26 Skapad: 2024-03-20 Senast uppdaterad: 2024-03-22Bibliografiskt granskad

Open Access i DiVA

fulltext(587 kB)74 nedladdningar
Filinformation
Filnamn FULLTEXT01.pdfFilstorlek 587 kBChecksumma SHA-512
7f599b09e4643a906584fed6fac940ff009a10c77e50778ffe4f49af2169e8eaec99420a5d15f4519156fb969e2b557b648f5bc1ac38c51461c8b471aa417c0f
Typ fulltextMimetyp application/pdf

Övriga länkar

ScopusPublisher's full text

Person

Hatefi, ArezooEklund, Anton

Sök vidare i DiVA

Av författaren/redaktören
Hatefi, ArezooEklund, Anton
Av organisationen
Institutionen för datavetenskap
Data- och informationsvetenskap

Sök vidare utanför DiVA

GoogleGoogle Scholar
Totalt: 77 nedladdningar
Antalet nedladdningar är summan av nedladdningar för alla fulltexter. Det kan inkludera t.ex tidigare versioner som nu inte längre är tillgängliga.

isbn
urn-nbn

Altmetricpoäng

isbn
urn-nbn
Totalt: 400 träffar
RefereraExporteraLänk till posten
Permanent länk

Direktlänk
Referera
Referensformat
  • apa
  • ieee
  • vancouver
  • Annat format
Fler format
Språk
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Annat språk
Fler språk
Utmatningsformat
  • html
  • text
  • asciidoc
  • rtf