Bipol: Multi-axes Evaluation of Bias with Explainability in BenchmarkDatasetsShow others and affiliations
2023 (English)In: Proceedings of Recent Advances in Natural Language Processing / [ed] Galia Angelova, Maria Kunilovskaya and Ruslan Mitkov, Incoma Ltd. , 2023, p. 1-10Conference paper, Published paper (Refereed)
Abstract [en]
We investigate five English NLP benchmark datasets (on the superGLUE leaderboard) and two Swedish datasets for bias, along multiple axes. The datasets are the following: Boolean Question (Boolq), CommitmentBank (CB), Winograd Schema Challenge (WSC), Winogender diagnostic (AXg), Recognising Textual Entailment (RTE), Swedish CB, and SWEDN. Bias can be harmful and it is known to be common in data, which ML models learn from. In order to mitigate bias in data, it is crucial to be able to estimate it objectively. We use bipol, a novel multi-axes bias metric with explainability, to estimate and explain how much bias exists in these datasets. Multilingual, multi-axes bias evaluation is not very common. Hence, we also contribute a new, large Swedish bias-labeled dataset (of 2 million samples), translated from the English version and train the SotA mT5 model on it. In addition, we contribute new multi-axes lexica for bias detection in Swedish. We make the codes, model, and new dataset publicly available.
Place, publisher, year, edition, pages
Incoma Ltd. , 2023. p. 1-10
Series
International conference Recent advances in natural language processing, E-ISSN 2603-2813 ; 2023
National Category
Language Technology (Computational Linguistics)
Research subject
Machine Learning
Identifiers
URN: urn:nbn:se:ltu:diva-103097DOI: 10.26615/978-954-452-092-2_001Scopus ID: 2-s2.0-85179181932OAI: oai:DiVA.org:ltu-103097DiVA, id: diva2:1815962
Conference
International Conference Recent Advances In Natural Language Processing (RANLP 2023), Varna, Bulgaria, September 4-6, 2023
Note
ISBN for host publication: 978-954-452-092-2
2023-11-302023-11-302024-11-20Bibliographically approved