TUdatalib : Benchmarking the Attribution Quality of Vision Models

Zur Kurzanzeige

dc.contributor.author	Hesse, Robin
dc.contributor.author	Schaub-Meyer, Simone
dc.contributor.author	Roth, Stefan
dc.date.accessioned	2025-04-03T12:11:16Z
dc.date.available	2025-04-03T12:11:16Z
dc.date.issued	2024-12
dc.identifier.uri	https://tudatalib.ulb.tu-darmstadt.de/handle/tudatalib/4531
dc.description	Attribution maps are one of the most established tools to explain the functioning of computer vision models. They assign importance scores to input features, indicating how relevant each feature is for the prediction of a deep neural network. While much research has gone into proposing new attribution methods, their proper evaluation remains a difficult challenge. In this work, we propose a novel evaluation protocol that overcomes two fundamental limitations of the widely used incremental-deletion protocol, i.e., the out-of-domain issue and lacking inter-model comparisons. This allows us to evaluate 23 attribution methods and how different design choices of popular vision backbones affect their attribution quality. We find that intrinsically explainable models outperform standard models and that raw attribution values exhibit a higher attribution quality than what is known from previous work. Further, we show consistent changes in the attribution quality when varying the network design, indicating that some standard design choices promote attribution quality.	de_DE
dc.language.iso	en	de_DE
dc.relation	IsDescribedBy;arXiv;2407.11910
dc.rights	Apache License 2.0
dc.rights.uri	https://www.apache.org/licenses/LICENSE-2.0
dc.subject	deep learning	de_DE
dc.subject	interpretability	de_DE
dc.subject	explainable artificial intelligence	de_DE
dc.subject	evaluation	de_DE
dc.subject.classification	4.43-05 Bild- und Sprachverarbeitung, Computergraphik und Visualisierung, Human Computer Interaction, Ubiquitous und Wearable Computing	de_DE
dc.subject.ddc	004
dc.title	Benchmarking the Attribution Quality of Vision Models	de_DE
dc.type	Software	de_DE
dc.type	Model	de_DE
tud.project	EC/H2020 \| 866008 \| RED	de_DE
tud.project	HMWK \| 500/10.001-(00111) \| 3AI - TP Roth	de_DE
tud.project	HMWK \| 500/10.001-(00012) \| TAM - TP Roth	de_DE
tud.project	HMWK \| 500/10.001-(00111) \| 3AI-NWG Schaub-Meyer	de_DE
tud.unit	TUDa