Active Learning Efficiency Benchmark for Coreference Resolution Including Advanced Uncertainty Representations

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F23%3A00374834" target="_blank" >RIV/68407700:21230/23:00374834 - isvavai.cz</a>

Výsledek na webu

<a href="https://doi.org/10.1109/CISDS61173.2023.00016" target="_blank" >https://doi.org/10.1109/CISDS61173.2023.00016</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/CISDS61173.2023.00016" target="_blank" >10.1109/CISDS61173.2023.00016</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Active Learning Efficiency Benchmark for Coreference Resolution Including Advanced Uncertainty Representations

Popis výsledku v původním jazyce

Active learning is a powerful technique that accelerates model learning by iteratively expanding training data based on the model’s feedback. This approach has proven particularly relevant in natural language processing and other machine learning domains. While active learning has been extensively studied for conventional classification tasks, its application to more specialized tasks like neural coreference resolution has the potential for improvement. In our research, we present a significant advancement by applying active learning to the neural coreference problem, and setting a benchmark of 39% reduction in required annotations for training data. Simultaneously, it preserves performance compared to the original model trained on the full data. We compare various uncertainty sampling techniques along with Bayesian modifications of coreference resolution models, conducting a comprehensive analysis of annotation efforts. The results demonstrate that the best-performing techniques seek to maximize label annotation in previously chosen documents, showcasing their effectiveness and preserving performance.

Název v anglickém jazyce

Active Learning Efficiency Benchmark for Coreference Resolution Including Advanced Uncertainty Representations

Popis výsledku anglicky

Active learning is a powerful technique that accelerates model learning by iteratively expanding training data based on the model’s feedback. This approach has proven particularly relevant in natural language processing and other machine learning domains. While active learning has been extensively studied for conventional classification tasks, its application to more specialized tasks like neural coreference resolution has the potential for improvement. In our research, we present a significant advancement by applying active learning to the neural coreference problem, and setting a benchmark of 39% reduction in required annotations for training data. Simultaneously, it preserves performance compared to the original model trained on the full data. We compare various uncertainty sampling techniques along with Bayesian modifications of coreference resolution models, conducting a comprehensive analysis of annotation efforts. The results demonstrate that the best-performing techniques seek to maximize label annotation in previously chosen documents, showcasing their effectiveness and preserving performance.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

<a href="/cs/project/TL05000057" target="_blank" >TL05000057: Signál a šum v éře Žurnalistiky 5.0 - komparativní perspektiva novinářských žánrů automatizovaných obsahů</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2023

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

2023 2nd International Conference on Frontiers of Communications, Information System and Data Science

ISBN

979-8-3503-8147-4

ISSN

—

e-ISSN

—

Počet stran výsledku

8

Strana od-do

40-47

Název nakladatele

IEEE Computer Society

Místo vydání

Los Alamitos

Místo konání akce

Xi’an

Datum konání akce

24. 11. 2023

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—