Print Email Facebook Twitter Training Data Augmentation for Detecting Adverse Drug Reactions in User-Generated Content Title Training Data Augmentation for Detecting Adverse Drug Reactions in User-Generated Content Author Mesbah, S. (TU Delft Web Information Systems) Yang, J. (Amazon) Sips, R.H.J. (myTomorrows) Valle Torre, M. (TU Delft Web Information Systems) Lofi, C. (TU Delft Web Information Systems) Bozzon, A. (TU Delft Human-Centred Artificial Intelligence; TU Delft Web Information Systems) Houben, G.J.P.M. (TU Delft Web Information Systems) Date 2019-11-03 Abstract Social media provides a timely yet challenging data source for adverse drug reaction (ADR) detection. Existing dictionary-based, semi-supervised learning approaches are intrinsically limited by the coverage and maintainability of laymen health vocabularies. In this paper, we introduce a data augmentation approach that leverages variational autoencoders to learn high-quality data distributions from a large unlabeled dataset, and subsequently, to automatically generate a large labeled training set from a small set of labeled samples. This allows for efficient social-media ADR detection with low training and re-training costs to adapt to the changes and emergence of informal medical laymen terms. An extensive evaluation performed on Twitter and Reddit data shows that our approach matches the performance of fully-supervised approaches while requiring only 25% of training data. Subject NLPMLData augmentation To reference this document use: http://resolver.tudelft.nl/uuid:27dca22e-c9db-4780-81e3-0cd96a16a451 DOI https://doi.org/10.18653/v1/D19-1239 Embargo date 2020-06-01 Source International conference on Empirical Methods in Natural Language Processing (EMNLP) Event 2019 Conference on Empirical Methods in Natural Language Processing and 9th International Joint Conference on Natural Language Processing, 2019-11-03 → 2019-11-07, Hong Kong, China Bibliographical note Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public. Part of collection Institutional Repository Document type conference paper Rights © 2019 S. Mesbah, J. Yang, R.H.J. Sips, M. Valle Torre, C. Lofi, A. Bozzon, G.J.P.M. Houben Files PDF D19_1239.pdf 830.28 KB Close viewer /islandora/object/uuid:27dca22e-c9db-4780-81e3-0cd96a16a451/datastream/OBJ/view