Explaining relation classification models with semantic extents
- In recent years, the development of large pretrained language models, such as BERT and GPT, significantly improved information extraction systems on various tasks, including relation classification. State-of-the-art systems are highly accurate on scientific benchmarks. A lack of explainability is currently a complicating factor in many real-world applications. Comprehensible systems are necessary to prevent biased, counterintuitive, or harmful decisions. We introduce semantic extents, a concept to analyze decision patterns for the relation classification task. Semantic extents are the most influential parts of texts concerning classification decisions. Our definition allows similar procedures to determine semantic extents for humans and models. We provide an annotation tool and a software framework to determine semantic extents for humans and models conveniently and reproducibly. Comparing both reveals that models tend to learn shortcut patterns from data. These patterns are hard to detect with current interpretability methods, such as input reductions. Our approach can help detect and eliminate spurious decision patterns during model development. Semantic extents can increase the reliability and security of natural language processing systems. Semantic extents are an essential step in enabling applications in critical areas like healthcare or finance. Moreover, our work opens new research directions for developing methods to explain deep learning models.
Author: | Lars KlöserORCiD, André BüsgenORCiD, Philipp KohlORCiD, Bodo Kraft, Albert Zündorf |
---|---|
DOI: | https://doi.org/10.1007/978-3-031-39059-3_13 |
ISBN: | 978-3-031-39058-6 (Print) |
ISBN: | 978-3-031-39059-3 (Online) |
Parent Title (English): | Deep Learning Theory and Applications |
Publisher: | Springer |
Place of publication: | Cham |
Editor: | Donatello Conte, Ana Fred, Oleg Gusikhin, Carlo Sansone |
Document Type: | Conference Proceeding |
Language: | English |
Year of Completion: | 2023 |
Tag: | Information extraction; Natural language processing; Natural language understanding; Relation classification; Trustworthy artificial intelligence |
First Page: | 189 |
Last Page: | 208 |
Note: | 4th International Conference, DeLTA 2023, Rome, Italy, July 13–14, 2023. |
Link: | https://doi.org/10.1007/978-3-031-39059-3_13 |
Zugriffsart: | campus |
Institutes: | FH Aachen / Fachbereich Medizintechnik und Technomathematik |
collections: | Verlag / Springer |