Using Deep Image Priors to Generate Counterfactual Explanations

Through the use of carefully tailored convolutional neural network architectures, a deep image prior (DIP) can be used to obtain pre-images from latent representation encodings. Though DIP inversion has been known to be superior to conventional regularized inversion strategies such as total variation, such an over-parameterized generator is able to effectively reconstruct even images that are not in the original data distribution. This limitation makes it challenging to utilize such priors for tasks such as counterfactual reasoning, wherein the goal is to generate small, interpretable changes to an image that systematically leads to changes in the model prediction. To this end, we propose a novel regularization strategy based on an auxiliary loss estimator jointly trained with the predictor, which efficiently guides the prior to recover natural pre-images. Our empirical studies with a real-world ISIC skin lesion detection problem clearly evidence the effectiveness of the proposed approach in synthesizing meaningful counterfactuals. In comparison, we find that the standard DIP inversion often proposes visually imperceptible perturbations to irrelevant parts of the image, thus providing no additional insights into the model behavior.

[1]  Yoshua Bengio,et al.  Generative Adversarial Nets , 2014, NIPS.

[2]  Andrea Vedaldi,et al.  Deep Image Prior , 2017, International Journal of Computer Vision.

[3]  Rushil Anirudh,et al.  MimicGAN: Robust Projection onto Image Manifolds with Corruption Mimicking , 2020, International Journal of Computer Vision.

[4]  Thomas Brox,et al.  U-Net: Convolutional Networks for Biomedical Image Segmentation , 2015, MICCAI.

[5]  Ali Ahmed,et al.  Blind Image Deconvolution Using Deep Generative Priors , 2018, IEEE Transactions on Computational Imaging.

[6]  Thomas Brox,et al.  Inverting Visual Representations with Convolutional Networks , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[7]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[8]  Andrea Vedaldi,et al.  Visualizing Deep Convolutional Neural Networks Using Natural Pre-images , 2015, International Journal of Computer Vision.

[9]  Noel C. F. Codella,et al.  Skin Lesion Analysis Toward Melanoma Detection 2018: A Challenge Hosted by the International Skin Imaging Collaboration (ISIC) , 2019, ArXiv.

[10]  Chinmay Hegde,et al.  Solving Linear Inverse Problems Using Gan Priors: An Algorithm with Provable Guarantees , 2018, 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[11]  Minh N. Do,et al.  Semantic Image Inpainting with Deep Generative Models , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[12]  In So Kweon,et al.  Learning Loss for Active Learning , 2019, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[13]  Andrea Vedaldi,et al.  Understanding deep image representations by inverting them , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[14]  Rushil Anirudh,et al.  Unsupervised Audio Source Separation using Generative Priors , 2020, INTERSPEECH.

[15]  Harald Kittler,et al.  Descriptor : The HAM 10000 dataset , a large collection of multi-source dermatoscopic images of common pigmented skin lesions , 2018 .

[16]  Rushil Anirudh,et al.  Accurate and Robust Feature Importance Estimation under Distribution Shifts , 2020, AAAI.

[17]  Deepta Rajan,et al.  Calibrating Healthcare AI: Towards Reliable and Interpretable Deep Predictive Models , 2020, ArXiv.