Print Email Facebook Twitter SparCAssist Title SparCAssist: A Model Risk Assessment Assistant Based on Sparse Generated Counterfactuals Author Zhang, Zijian (L3S Research Center) Setty, Vinay (University of Stavanger; L3S Research Center) Anand, A. (TU Delft Web Information Systems; L3S Research Center) Date 2022 Abstract We introduce SparCAssist, a general-purpose risk assessment tool for the machine learning models trained for language tasks. It evaluates models' risk by inspecting their behavior on counterfactuals, namely out-of-distribution instances generated based on the given data instance. The counterfactuals are generated by replacing tokens in rational subsequences identified by ExPred, while the replacements are retrieved using HotFlip or the Masked-Language-Model-based algorithms. The main purpose of our system is to help the human annotators to assess the model's risk on deployment. The counterfactual instances generated during the assessment are the by-product and can be used to train more robust NLP models in the future. Subject counterfactual interpretationdata-annotation toolshuman-in-the-loop machine learninginterpretable machine learning To reference this document use: http://resolver.tudelft.nl/uuid:553f3377-f4bd-4b19-a0ee-974d9947f62b DOI https://doi.org/10.1145/3477495.3531677 Publisher Association for Computing Machinery (ACM) Embargo date 2023-07-01 ISBN 978-1-4503-8732-3 Source SIGIR 2022 - Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval Event 45th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2022, 2022-07-11 → 2022-07-15, Madrid, Spain Series SIGIR 2022 - Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval 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 © 2022 Zijian Zhang, Vinay Setty, A. Anand Files PDF 3477495.3531677.pdf 1.31 MB Close viewer /islandora/object/uuid:553f3377-f4bd-4b19-a0ee-974d9947f62b/datastream/OBJ/view