Let you know the principle of AI models from excellent works
This post is the collection of explainable AI techniques, tools, applications and reviews.
✅ Techniques
◾ Class Activation maps
◾ Anchors
◾ Prediction Difference Analysis
◾ Contextual Prediction Difference Analysis
◾ Shapley Value Sampling
◾ DeConvNet
◾ Gradient/Saliency Maps
◾ Gradient * input
◾ FullGrad
◾ GradCAM
◾ Guided GradCAM
◾ SmoothGrad
◾ VarGrad
◾ Integrated Gradients
◾ Expressive gradients
◾ Gradient SHAP
◾ Deep SHAP
◾ SHAP Interaction Index
◾ Kernel SHAP
◾ DeepLIFT SHAP
◾ DeepLIFT
◾ Excitation Backprop
◾ Guided backpropagation
◾ NeuronGuidedBackprop
◾ Tree Explainer
◾ GNNExplainer
◾ GNN-LRP
◾ Layer-wise Relevance Propagation
◾ Spectral Relevance Analysis
◾ DeepTaylor
◾ LayerConductance
◾ Local Rule-based Explanations
◾ LIME 📝Introduction
◾ STREAM
◾ RISE
◾ PatternNet
◾ Pattern Attribution
◾ Occlusion
◾ Meaningful Perturbation
◾ ExtremalPerturbation
◾ Internal Influence
◾ Representation Erasure
◾ FIDO
◾ NeuronConductance
◾ TotalConductance
◾ DeepDreams
◾ TCAV
◾ TCAV with RCV
◾ UBS
◾ SENN
✅ Tools & Applications
🔸 Heatmapping[web]
🔸 CNN-explainer[web]
🔸 Explainable AI Demos[web]
🔸 A Neural Network Playground[web]
🔸 Summit[web]
🔸 NeuralDivergence[web]
🔸 SCIN[web]
🔸 Neuroscope[software]
🔸 LUCID[library]
🔸 Keras-vis[library]
🔸 DeepExplain[library]
🔸 iNNvestigate[library]
🔸 TensorFlow Graph Visualizer[library]
🔸 tf-explain[library]
🔸 TorchRay[library]
🔸 Captum[library]
🔸 Explainable AI[commercial]
🔸 Explainable AI Platform[commercial]
🔸 exAID[commercial]
🔸 DASL[commercial]
✅ Reviews
🔹 Interpretable Machine Learning A Guide for Making Black Box Models Explainable.
Christoph Molnar
🔹 Towards Robust Interpretability with Self-Explaining Neural Networks, 2018
🔹 Towards Explainable Artificial Intelligence, 2019
🔹 On the Integration of Knowledge Graphs into Deep Learning Models for a More Comprehensible AI—Three Challenges for Future Research, 2020
🔹 Explainable Deep Learning: A Field Guide for the Uninitiated, 2020
🔹 A Survey on Explainable Artificial Intelligence (XAI): Towards Medical XAI, 2020
🔹 Explainable Artificial Intelligence (xAI) Approaches and Deep Meta-Learning Models, 2020
🔹 Explainable Deep Learning Models in Medical Image Analysis, 2020
🔹 Achievements and Challenges in Explaining Deep Learning based Computer-Aided Diagnosis Systems, 2020
🔹 Interpretability and Explainability: A Machine Learning Zoo Mini-tour, 2020
🔹 Explaining Deep Neural Networks and Beyond: A Review of Methods and Applications, 2021
🔹 Awesome machine learning interpretability, up to date
✅ Evaluations
◽ The (Un)reliability of Saliency Methods: input variant
◽ Sanity Checks for Saliency Maps: model and data
◽ Evaluating the Visualization of What a Deep Neural Network Has Learned: AOPC
◽ Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead, Cynthia Rudin