Early Discovery of Disappearing Entities in Microblogs

38609

Authors

nan

Abstract

AbstractWe make decisions by reacting to changes in the real world, particularly the emergence and disappearance of impermanent entities such as restaurants, services, and events. Because we want to avoid missing out on opportunities or making fruitless actions after those entities have disappeared, it is important to know when entities disappear as early as possible. We thus tackle the task of detecting disappearing entities from microblogs where various information is shared timely. The major challenge is detecting uncertain contexts of disappearing entities from noisy microblog posts. To collect such disappearing contexts, we design time-sensitive distant supervision, which utilizes entities from the knowledge base and time-series posts. Using this method, we actually build large-scale Twitter datasets of disappearing entities. To ensure robust detection in noisy environments, we refine pretrained word embeddings for the detection model on microblog streams in a timely manner. Experimental results on the Twitter datasets confirmed the effectiveness of the collected labeled data and refined word embeddings; the proposed method outperformed a baseline in terms of accuracy, and more than 70% of the detected disappearing entities in Wikipedia are discovered earlier than the update on Wikipedia, with the average lead-time is over one month.

Keywords

nan

Full Content

Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics Volume 1: Long Papers, pages 4507–4520 July 9-14, 2023 ©2023 Association for Computational Linguistics Early Discovery of Disappearing Entities in Microblogs Satoshi Akasaki∗ Yahoo Japan Corporation sakasaki@yahoo-corp.jp Naoki Yoshinaga Institute of Industrial Science, The University of Tokyo ynaga@iis.u-tokyo.ac.jp Masashi Toyoda Institute of Industrial Science, The University of Tokyo toyoda@iis.u-tokyo.ac.jp Abstract We make decisions by reacting to changes in the real world, particularly the emergence and disappearance of impermanent entities such as restaurants, services, and events. Because we want to avoid missing out on opportunities or making fruitless actions after those entities have disappeared, it is important to know when entities disappear as early as possible. We thus tackle the task of detecting disappearing enti- ties from microblogs where various informa- tion is shared timely. The major challenge is detecting uncertain contexts of disappearing en- tities from noisy microblog posts. To collect such disappearing contexts, we design time- sensitive distant supervision, which utilizes en- tities from the knowledge base and time-series posts. Using this method, we actually build large-scale Twitter datasets of disappearing en- tities1 for English and Japanese. To ensure robust detection in noisy environments, we re- fine pretrained word embeddings for the detec- tion model on microblog streams in a timely manner. Experimental results on the Twitter datasets confirmed the effectiveness of the col- lected labeled data and refined word embed- dings; the proposed method outperformed a baseline in terms of accuracy, and more than 70% of the detected disappearing entities in Wikipedia are discovered earlier than the up- date on Wikipedia, with the average lead-time is over one month. 1 Introduction Our daily actions depend on the state of the real world and its changes, especially changes in the entities available to us. Among the various changes, the beginning of entities, or emerging entities (Akasaki et al., 2019) such as new songs, movies, and events, are useful for understanding ∗Part of this work was carried out during the period the author was at the University of Tokyo. 1We release the datasets (tweet IDs) used in the experi- ments to promote reproducibility. URL I'm so sad to hear that Dave Laing has died. Dave was a very accomplished music industry journalist. Can't believe Google+ is being shut down. It's like when they just pulled Google Friends Connect… Here's your Demolition Day Planner for Martin Tower. A brief, stray shower can't be ruled out… Red Bull Air Race World Championship will not continue after 2019. URL Pristin to disband after 2 years promoting as a group K-Netz express how raged they are… Westfield State University Athletics - Dave Laing Named Little East Coach of the Year URL Don't forget you can follow our posts and leave feedback about recent visits on our Google+ page! PHOTOS!! Martin Tower‘s mosaic artwork goes on public display URL URL Pilot lineup for 2018 Red Bull Air Race World Championship: Britain is back! URL Happy new year everyone! I hope this year will …

Content truncated to 500 words

Labeled Text

AbstractWe make decisions by reacting to changes in the real world, particularly the emergence and disappearance of impermanent entities such as restaurants, services, and events. Because we want to avoid missing out on opportunities or making fruitless actions after those entities have disappeared, it is important to know when entities disappear as early as possible. We thus tackle the task of detecting disappearing entities from microblogs where various information is shared timely. The major challenge is detecting uncertain contexts of disappearing entities from noisy microblog posts. To collect such disappearing contexts, we design time-sensitive distant supervision, which utilizes entities from the knowledge base and time-series posts. Using this method, we actually build large-scale Twitter datasets of disappearing entities. To ensure robust detection in noisy environments, we refine pretrained word embeddings for the detection model on microblog streams in a timely manner. Experimental results on the Twitter datasets confirmed the effectiveness of the collected labeled data and refined word embeddings; the proposed method outperformed a baseline in terms of accuracy, and more than 70% of the detected disappearing entities in Wikipedia are discovered earlier than the update on Wikipedia, with the average lead-time is over one month.

Publication Details

Paper ID:
38609
Year:
2023
Publication:
TIPSTER
Publication Type:
conf
DOI:
nan