TY  - CONF
A1  - Klöser, Lars
A1  - Büsgen, André
A1  - Kohl, Philipp
A1  - Kraft, Bodo
A1  - Zündorf, Albert
A2  - Conte, Donatello
A2  - Fred, Ana
A2  - Gusikhin, Oleg
A2  - Sansone, Carlo
T1  - Explaining relation classification models with semantic extents
T2  - DeLTA 2023: Deep Learning Theory and Applications
N2  - 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.
KW  - Relation classification
KW  - Natural language processing
KW  - Natural language understanding
KW  - Information extraction
KW  - Trustworthy artificial intelligence
Y1  - 2023
UR  - https://opus.bibliothek.fh-aachen.de/opus4/frontdoor/index/index/docId/10720
SN  - 978-3-031-39058-6 (Print)
SN  - 978-3-031-39059-3 (Online)
N1  - 4th International Conference, DeLTA 2023, Rome, Italy, July 13–14, 2023.
SP  - 189
EP  - 208
PB  - Springer
CY  - Cham
ER  -