This Looks Like That: Deep Learning for Interpretable Image Recognition

Abstract

When we are faced with challenging image classification tasks, we often explain our reasoning by dissecting the image, and pointing out prototypical aspects of one class or another. The mounting evidence for each of the classes helps us make our final decision. In this work, we introduce a deep network architecture -- prototypical part network (ProtoPNet), that reasons in a similar way: the network dissects the image by finding prototypical parts, and combines evidence from the prototypes to make a final classification. The model thus reasons in a way that is qualitatively similar to the way ornithologists, physicians, and others would explain to people on how to solve challenging image classification tasks. The network uses only image-level labels for training without any annotations for parts of images. We demonstrate our method on the CUB-200-2011 dataset and the Stanford Cars dataset. Our experiments show that ProtoPNet can achieve comparable accuracy with its analogous non-interpretable counterpart, and when several ProtoPNets are combined into a larger network, it can achieve an accuracy that is on par with some of the best-performing deep models. Moreover, ProtoPNet provides a level of interpretability that is absent in other interpretable deep models.

Chaofan Chen and Oscar Li contributed equally to this work. This work has been accepted for spotlight presentation (top 3% of papers) at NeurIPS 2019

References in corpus (16)

• Adam: A Method for Stochastic Optimization
• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• Show, Attend and Tell: Neural Image Caption Generation with Visual Attention
• Axiomatic Attribution for Deep Networks
• Methods for Interpreting and Understanding Deep Neural Networks
• Striving for Simplicity: The All Convolutional Net
• Pixel Recurrent Neural Networks
• Understanding Neural Networks Through Deep Visualization
• SmoothGrad: removing noise by adding noise
• Bird Species Categorization Using Pose Normalized Deep Convolutional Nets
• Deep k-Nearest Neighbors: Towards Confident, Interpretable and Robust Deep Learning
• ABC-CNN: An Attention Based Convolutional Neural Network for Visual Question Answering
• Prototype selection for interpretable classification
• Deep Learning for Case-Based Reasoning through Prototypes: A Neural Network that Explains Its Predictions
• Fully Convolutional Attention Networks for Fine-Grained Recognition
• Prototypal Analysis and Prototypal Regression

Cited by in corpus (65)

• Deep neural network models for computational histopathology: A survey
• Explainable Artificial Intelligence (XAI) 2.0: A Manifesto of Open Challenges and Interdisciplinary Research Directions
• Concept Whitening for Interpretable Image Recognition
• Crop Yield Prediction Integrating Genotype and Weather Variables Using Deep Learning
• From Attribution Maps to Human-Understandable Explanations through Concept Relevance Propagation
• One Model to Synthesize Them All: Multi-contrast Multi-scale Transformer for Missing Data Imputation
• Explainable Diabetic Retinopathy Detection and Retinal Image Generation
• Augmenting Interpretable Models with LLMs during Training
• On Interpretability of Artificial Neural Networks: A Survey
• This Looks Like That, Because ... Explaining Prototypes for Interpretable Image Recognition
• Privacy-preserving Federated Adversarial Domain Adaption over Feature Groups for Interpretability
• Improving Clinician Performance in Classification of EEG Patterns on the Ictal-Interictal-Injury Continuum using Interpretable Machine Learning
• Model-Based Counterfactual Synthesizer for Interpretation
• Counterfactual Explanation Based on Gradual Construction for Deep Networks
• Scientific Inference With Interpretable Machine Learning: Analyzing Models to Learn About Real-World Phenomena
• Improving deep learning with prior knowledge and cognitive models: A survey on enhancing explainability, adversarial robustness and zero-shot learning
• HYDRA: Hypergradient Data Relevance Analysis for Interpreting Deep Neural Networks
• IDP-PGFE: An Interpretable Disruption Predictor based on Physics-Guided Feature Extraction
• Privacy Meets Explainability: A Comprehensive Impact Benchmark
• Knowledge Distillation to Ensemble Global and Interpretable Prototype-Based Mammogram Classification Models
• Explaining Deep Classification of Time-Series Data with Learned Prototypes
• Explainable Differential Privacy-Hyperdimensional Computing for Balancing Privacy and Transparency in Additive Manufacturing Monitoring
• Predicting Events in MOBA Games: Prediction, Attribution, and Evaluation
• Synthetic Benchmarks for Scientific Research in Explainable Machine Learning
• ProtoAttend: Attention-Based Prototypical Learning
• On the Interpretability of Part-Prototype Based Classifiers: A Human Centric Analysis
• Deeply Explain CNN via Hierarchical Decomposition
• Explaining Explanations to Society
• Interpretable Medical Image Classification using Prototype Learning and Privileged Information
• FUTURE-AI: Guiding Principles and Consensus Recommendations for Trustworthy Artificial Intelligence in Medical Imaging
• Hybrid Predictive Model: When an Interpretable Model Collaborates with a Black-box Model
• Spatio-Temporal Perturbations for Video Attribution
• Trustworthy Convolutional Neural Networks: A Gradient Penalized-based Approach
• SUBPLEX: Towards a Better Understanding of Black Box Model Explanations at the Subpopulation Level
• Enhancing Deep Neural Network Saliency Visualizations with Gradual Extrapolation
• Human-Centered Evaluation of XAI Methods
• E Pluribus Unum Interpretable Convolutional Neural Networks
• Adversarial Attacks and Defenses: An Interpretation Perspective
• XProtoNet: Diagnosis in Chest Radiography with Global and Local Explanations
• Interpreting Neural Networks Using Flip Points
• AI Enabling Technologies: A Survey
• Distilling Ensemble of Explanations for Weakly-Supervised Pre-Training of Image Segmentation Models
• Towards Ground Truth Explainability on Tabular Data
• Optimising for Interpretability: Convolutional Dynamic Alignment Networks
• This Reads Like That: Deep Learning for Interpretable Natural Language Processing
• Analyzing Effects of Mixed Sample Data Augmentation on Model Interpretability
• Towards Human-Interpretable Prototypes for Visual Assessment of Image Classification Models
• Interpretable and Trustworthy Deepfake Detection via Dynamic Prototypes
• Analyzing the Interpretability Robustness of Self-Explaining Models
• Graph Neural Networks Including Sparse Interpretability
• Explainers in the Wild: Making Surrogate Explainers Robust to Distortions through Perception
• Local Concept Embeddings for Analysis of Concept Distributions in Vision DNN Feature Spaces
• DP-Net: Learning Discriminative Parts for image recognition
• Advancing Post Hoc Case Based Explanation with Feature Highlighting
• Interactive Mars Image Content-Based Search with Interpretable Machine Learning
• Birds look like cars: Adversarial analysis of intrinsically interpretable deep learning
• Tensorization of neural networks for improved privacy and interpretability
• Visual Probing: Cognitive Framework for Explaining Self-Supervised Image Representations
• SONG: Self-Organizing Neural Graphs
• Inductive Granger Causal Modeling for Multivariate Time Series
• Focus Longer to See Better:Recursively Refined Attention for Fine-Grained Image Classification
• A Framework to Learn with Interpretation
• Deep learning for bioimage analysis
• How Faithful are Self-Explainable GNNs?
• Explainable Deep Modeling of Tabular Data using TableGraphNet