Robust Attribution Regularization

An emerging problem in trustworthy machine learning is to train models that produce robust interpretations for their predictions. We take a step towards solving this problem through the lens of axiomatic attribution of neural networks. Our theory is grounded in the recent work, Integrated Gradients (IG), in axiomatically attributing a neural network's output change to its input change. We propose training objectives in classic robust optimization models to achieve robust IG attributions. Our objectives give principled generalizations of previous objectives designed for robust predictions, and they naturally degenerate to classic soft-margin training for one-layer neural networks. We also generalize previous theory and prove that the objectives for different robust optimization models are closely related. Experiments demonstrate the effectiveness of our method, and also point to intriguing problems which hint at the need for better optimization techniques or better neural network architectures for robust attribution training.

[1]  Yann LeCun,et al.  The mnist database of handwritten digits , 2005 .

[2]  Ankur Taly,et al.  Axiomatic Attribution for Deep Networks , 2017, ICML.

[3]  Johannes Stallkamp,et al.  Man vs. computer: Benchmarking machine learning algorithms for traffic sign recognition , 2012, Neural Networks.

[4]  Andrew Zisserman,et al.  Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps , 2013, ICLR.

[5]  Andrew Zisserman,et al.  A Visual Vocabulary for Flower Classification , 2006, 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06).

[6]  J. Zico Kolter,et al.  Provable defenses against adversarial examples via the convex outer adversarial polytope , 2017, ICML.

[7]  M. Kendall.  A NEW MEASURE OF RANK CORRELATION , 1938 .

[8]  Daniel Kuhn,et al.  Data-driven distributionally robust optimization using the Wasserstein metric: performance guarantees and tractable reformulations , 2015, Mathematical Programming.

[9]  Yuan Yu,et al.  TensorFlow: A system for large-scale machine learning , 2016, OSDI.

[10]  D. Luenberger.  Optimization by Vector Space Methods , 1968 .

[11]  Roland Vollgraf,et al.  Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms , 2017, ArXiv.

[12]  Aleksander Madry,et al.  Towards Deep Learning Models Resistant to Adversarial Attacks , 2017, ICLR.

[13]  Aleksander Madry,et al.  Adversarial Robustness as a Prior for Learned Representations , 2019 .

[14]  Harris Drucker,et al.  Improving generalization performance using double backpropagation , 1992, IEEE Trans. Neural Networks.

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

[16]  Andrew Slavin Ross,et al.  Improving the Adversarial Robustness and Interpretability of Deep Neural Networks by Regularizing their Input Gradients , 2017, AAAI.

[17]  Aleksander Madry,et al.  Robustness May Be at Odds with Accuracy , 2018, ICLR.

[18]  Abubakar Abid,et al.  Interpretation of Neural Networks is Fragile , 2017, AAAI.

[19]  Laurent El Ghaoui,et al.  Robust Optimization , 2021, ICORES.

[20]  John C. Duchi,et al.  Certifying Some Distributional Robustness with Principled Adversarial Training , 2017, ICLR.