03715nam a22005055i 4500 978-3-031-51518-7 DE-He213 20251030111806.0 cr nn 008mamaa 240405s2024 sz | s |||| 0|eng d 9783031515187 978-3-031-51518-7 10.1007/978-3-031-51518-7 doi QA76.9.N38 UYQL bicssc COM073000 bisacsh UYQL thema Atanasova, Pepa. author. aut http://id.loc.gov/vocabulary/relators/aut 74766 Accountable and Explainable Methods for Complex Reasoning over Text [electronic resource] / by Pepa Atanasova. 1st ed. 2024. Cham : Springer Nature Switzerland : Imprint: Springer, 2024. XVIII, 199 p. 24 illus. in color. online resource. text txt rdacontent computer c rdamedia online resource cr rdacarrier text file PDF rda 1. Executive Summary -- Part I: Accountability for Complex Reasoning Tasks over Text -- 2. Fact Checking with Insufficient Evidence -- 3. Generating Label Cohesive and Well-Formed Adversarial Claims -- Part II: Explainability for Complex Reasoning Tasks over Text -- 4. Generating Fact Checking Explanations -- 5. Generating Fluent Fact Checking Explanations with Unsupervised Post-Editing -- 6. Multi-Hop Fact Checking of Political Claims -- Part III: Diagnostic Explainability Methods -- 7. A Diagnostic Study of Explainability Techniques for Text Classification -- 8. Diagnostics-Guided Explanation Generation -- 9. Recent Developments on Accountability and Explainability for Complex Reasoning Tasks. This thesis presents research that expands the collective knowledge in the areas of accountability and transparency of machine learning (ML) models developed for complex reasoning tasks over text. In particular, the presented results facilitate the analysis of the reasons behind the outputs of ML models and assist in detecting and correcting for potential harms. It presents two new methods for accountable ML models; advances the state of the art with methods generating textual explanations that are further improved to be fluent, easy to read, and to contain logically connected multi-chain arguments; and makes substantial contributions in the area of diagnostics for explainability approaches. All results are empirically tested on complex reasoning tasks over text, including fact checking, question answering, and natural language inference. This book is a revised version of the PhD dissertation written by the author to receive her PhD from the Faculty of Science, University of Copenhagen, Denmark. In 2023, it won the Informatics Europe Best Dissertation Award, granted to the most outstanding European PhD thesis in the field of computer science. Natural language processing (Computer science). 65131 Information storage and retrieval systems. 65568 Machine learning. 55059 Natural Language Processing (NLP). 66164 Information Storage and Retrieval. 65570 Machine Learning. 55059 SpringerLink (Online service) Springer Nature eBook Printed edition: 9783031515170 Printed edition: 9783031515194 https://doi.org/10.1007/978-3-031-51518-7 ZDB-2-SCS ZDB-2-SXCS 58964 58964