Neural Prototype Trees for Interpretable Fine-grained Image Recognition
| Autor: | Christin Seifert, Meike Nauta, Ron van Bree |
|---|---|
| Přispěvatelé: | Datamanagement & Biometrics, Digital Society Institute |
| Jazyk: | angličtina |
| Rok vydání: | 2020 |
| Předmět: |
FOS: Computer and information sciences
Computer Science - Machine Learning Binary tree Artificial neural network business.industry Computer science Computer Science - Artificial Intelligence Computer Vision and Pattern Recognition (cs.CV) Deep learning Decision tree Medizin 22/2 OA procedure Computer Science - Computer Vision and Pattern Recognition Ensemble learning Machine Learning (cs.LG) Tree (data structure) Informatik Artificial Intelligence (cs.AI) Computer vision Node (circuits) Pruning (decision trees) Artificial intelligence business |
| Zdroj: | Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 14933-14943 STARTPAGE=14933;ENDPAGE=14943;TITLE=Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) CVPR |
| Popis: | Prototype-based methods use interpretable representations to address the black-box nature of deep learning models, in contrast to post-hoc explanation methods that only approximate such models. We propose the Neural Prototype Tree (ProtoTree), an intrinsically interpretable deep learning method for fine-grained image recognition. ProtoTree combines prototype learning with decision trees, and thus results in a globally interpretable model by design. Additionally, ProtoTree can locally explain a single prediction by outlining a decision path through the tree. Each node in our binary tree contains a trainable prototypical part. The presence or absence of this learned prototype in an image determines the routing through a node. Decision making is therefore similar to human reasoning: Does the bird have a red throat? And an elongated beak? Then it's a hummingbird! We tune the accuracy-interpretability trade-off using ensemble methods, pruning and binarizing. We apply pruning without sacrificing accuracy, resulting in a small tree with only 8 learned prototypes along a path to classify a bird from 200 species. An ensemble of 5 ProtoTrees achieves competitive accuracy on the CUB-200- 2011 and Stanford Cars data sets. Code is available at https://github.com/M-Nauta/ProtoTree Accepted to IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2021. 11 pages, and 9 pages supplementary |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|