Classification Metrics for Image Explanations: Towards Building Reliable XAI-Evaluations
Fuente:
arXiv
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866916279639605248 |
|---|---|
| author | Fresz, Benjamin Lörcher, Lena Huber, Marco |
| author_facet | Fresz, Benjamin Lörcher, Lena Huber, Marco |
| contents | Decision processes of computer vision models - especially deep neural networks - are opaque in nature, meaning that these decisions cannot be understood by humans. Thus, over the last years, many methods to provide human-understandable explanations have been proposed. For image classification, the most common group are saliency methods, which provide (super-)pixelwise feature attribution scores for input images. But their evaluation still poses a problem, as their results cannot be simply compared to the unknown ground truth. To overcome this, a slew of different proxy metrics have been defined, which are - as the explainability methods themselves - often built on intuition and thus, are possibly unreliable. In this paper, new evaluation metrics for saliency methods are developed and common saliency methods are benchmarked on ImageNet. In addition, a scheme for reliability evaluation of such metrics is proposed that is based on concepts from psychometric testing. The used code can be found at https://github.com/lelo204/ClassificationMetricsForImageExplanations . |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_05068 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Classification Metrics for Image Explanations: Towards Building Reliable XAI-Evaluations Fresz, Benjamin Lörcher, Lena Huber, Marco Computer Vision and Pattern Recognition Artificial Intelligence Human-Computer Interaction Decision processes of computer vision models - especially deep neural networks - are opaque in nature, meaning that these decisions cannot be understood by humans. Thus, over the last years, many methods to provide human-understandable explanations have been proposed. For image classification, the most common group are saliency methods, which provide (super-)pixelwise feature attribution scores for input images. But their evaluation still poses a problem, as their results cannot be simply compared to the unknown ground truth. To overcome this, a slew of different proxy metrics have been defined, which are - as the explainability methods themselves - often built on intuition and thus, are possibly unreliable. In this paper, new evaluation metrics for saliency methods are developed and common saliency methods are benchmarked on ImageNet. In addition, a scheme for reliability evaluation of such metrics is proposed that is based on concepts from psychometric testing. The used code can be found at https://github.com/lelo204/ClassificationMetricsForImageExplanations . |
| title | Classification Metrics for Image Explanations: Towards Building Reliable XAI-Evaluations |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Human-Computer Interaction |
| url | https://arxiv.org/abs/2406.05068 |