Training Machine Learning Models by Regularizing their Explanations

Abstract

Neural networks are among the most accurate supervised learning methods in use today. However, their opacity makes them difficult to trust in critical applications, especially when conditions in training may differ from those in practice. Recent efforts to develop explanations for neural networks and machine learning models more generally have produced tools to shed light on the implicit rules behind predictions. These tools can help us identify when models are right for the wrong reasons. However, they do not always scale to explaining predictions for entire datasets, are not always at the right level of abstraction, and most importantly cannot correct the problems they reveal. In this thesis, we explore the possibility of training machine learning models (with a particular focus on neural networks) using explanations themselves. We consider approaches where models are penalized not only for making incorrect predictions but also for providing explanations that are either inconsistent with domain knowledge or overly complex. These methods let us train models which can not only provide more interpretable rationales for their predictions but also generalize better when training data is confounded or meaningfully different from test data (even adversarially so).

Harvard CSE master's thesis; includes portions of arxiv:1703.03717 and arxiv:1711.09404

References in corpus (18)

• Towards A Rigorous Science of Interpretable Machine Learning
• Snorkel: Rapid Training Data Creation with Weak Supervision
• SmoothGrad: removing noise by adding noise
• Towards Deep Neural Network Architectures Robust to Adversarial Examples
• Certified Defenses against Adversarial Examples
• Understanding Neural Networks through Representation Erasure
• The Space of Transferable Adversarial Examples
• Deep k-Nearest Neighbors: Towards Confident, Interpretable and Robust Deep Learning
• Grad-CAM: Why did you say that?
• Defensive Distillation is Not Robust to Adversarial Examples
• The (Un)reliability of saliency methods
• Deep Learning for Case-Based Reasoning through Prototypes: A Neural Network that Explains Its Predictions
• Structured Disentangled Representations
• Beyond Sparsity: Tree Regularization of Deep Models for Interpretability
• An unexpected unity among methods for interpreting model predictions
• Biologically inspired protection of deep networks from adversarial attacks
• Interpretation of Prediction Models Using the Input Gradient
• Programs as Black-Box Explanations