We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CV

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computer Vision and Pattern Recognition

Title: NBDT: Neural-Backed Decision Trees

Abstract: Deep learning is being adopted in settings where accurate and justifiable predictions are required, ranging from finance to medical imaging. While there has been recent work providing post-hoc explanations for model predictions, there has been relatively little work exploring more directly interpretable models that can match state-of-the-art accuracy. Historically, decision trees have been the gold standard in balancing interpretability and accuracy. However, recent attempts to combine decision trees with deep learning have resulted in models that (1) achieve accuracies far lower than that of modern neural networks (e.g. ResNet) even on small datasets (e.g. MNIST), and (2) require significantly different architectures, forcing practitioners pick between accuracy and interpretability. We forgo this dilemma by creating Neural-Backed Decision Trees (NBDTs) that (1) achieve neural network accuracy and (2) require no architectural changes to a neural network. NBDTs achieve accuracy within 1% of the base neural network on CIFAR10, CIFAR100, TinyImageNet, using recently state-of-the-art WideResNet; and within 2% of EfficientNet on ImageNet. This yields state-of-the-art explainable models on ImageNet, with NBDTs improving the baseline by ~14% to 75.30% top-1 accuracy. Furthermore, we show interpretability of our model's decisions both qualitatively and quantitatively via a semi-automatic process. Code and pretrained NBDTs can be found at this https URL
Comments: 14 pages, 9 figures
Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE)
Cite as: arXiv:2004.00221 [cs.CV]
  (or arXiv:2004.00221v1 [cs.CV] for this version)

Submission history

From: Alvin Wan [view email]
[v1] Wed, 1 Apr 2020 04:04:03 GMT (3656kb,D)
[v2] Thu, 11 Jun 2020 22:33:20 GMT (5183kb,D)
[v3] Thu, 28 Jan 2021 03:06:26 GMT (9783kb,D)

Link back to: arXiv, form interface, contact.