深度学习近年来关于模型解释性的相关论文。
按引用次数排序可见引用排序
159篇论文pdf(有2篇需要上scihub找)上传到腾讯微云。
不定期更新。
| Year | Publication | Paper | Citation | code |
|---|---|---|---|---|
| 2020 | ICLR | Knowledge Isomorphism between Neural Networks | 0 | |
| 2020 | ICLR | Interpretable Complex-Valued Neural Networks for Privacy Protection | 2 | |
| 2019 | AI | Explanation in artificial intelligence: Insights from the social sciences | 495 | |
| 2019 | NMI | Stop Explaining Black Box Machine Learning Models for High Stakes Decisions and Use Interpretable Models Instead | 142 | |
| 2019 | NeurIPS | This looks like that: deep learning for interpretable image recognition | 49 | Pytorch |
| 2019 | NeurIPS | A benchmark for interpretability methods in deep neural networks（同arxiv:1806.10758） | 6 | |
| 2019 | NeurIPS | Full-gradient representation for neural network visualization | 3 | |
| 2019 | NeurIPS | On the (In) fidelity and Sensitivity of Explanations | 3 | |
| 2019 | NeurIPS | Towards Automatic Concept-based Explanations | 6 | Tensorflow |
| 2019 | NeurIPS | CXPlain: Causal explanations for model interpretation under uncertainty | 3 | |
| 2019 | CVPR | Interpreting CNNs via Decision Trees | 65 | |
| 2019 | CVPR | From Recognition to Cognition: Visual Commonsense Reasoning | 63 | Pytorch |
| 2019 | CVPR | Attention branch network: Learning of attention mechanism for visual explanation | 22 | |
| 2019 | CVPR | Interpretable and fine-grained visual explanations for convolutional neural networks | 9 | |
| 2019 | CVPR | Learning to Explain with Complemental Examples | 9 | |
| 2019 | CVPR | Revealing Scenes by Inverting Structure from Motion Reconstructions | 8 | Tensorflow |
| 2019 | CVPR | Multimodal Explanations by Predicting Counterfactuality in Videos | 3 | |
| 2019 | CVPR | Visualizing the Resilience of Deep Convolutional Network Interpretations | 1 | |
| 2019 | ICCV | U-CAM: Visual Explanation using Uncertainty based Class Activation Maps | 9 | |
| 2019 | ICCV | Towards Interpretable Face Recognition | 7 | |
| 2019 | ICCV | Taking a HINT: Leveraging Explanations to Make Vision and Language Models More Grounded | 10 | |
| 2019 | ICCV | Understanding Deep Networks via Extremal Perturbations and Smooth Masks | 10 | Pytorch |
| 2019 | ICCV | Explaining Neural Networks Semantically and Quantitatively | 4 | |
| 2019 | ICLR | Hierarchical interpretations for neural network predictions | 22 | Pytorch |
| 2019 | ICLR | How Important Is a Neuron? | 10 | |
| 2019 | ICLR | Visual Explanation by Interpretation: Improving Visual Feedback Capabilities of Deep Neural Networks | 9 | |
| 2018 | ICML | Extracting Automata from Recurrent Neural Networks Using Queries and Counterexamples | 60 | Pytorch |
| 2019 | ICML | Towards A Deep and Unified Understanding of Deep Neural Models in NLP | 10 | Pytorch |
| 2019 | ICAIS | Interpreting black box predictions using fisher kernels | 14 | |
| 2019 | ACMFAT | Explaining explanations in AI | 78 | |
| 2019 | AAAI | Interpretation of neural networks is fragile | 87 | Tensorflow |
| 2019 | AAAI | Classifier-agnostic saliency map extraction | 7 | |
| 2019 | AAAI | Can You Explain That? Lucid Explanations Help Human-AI Collaborative Image Retrieval | 1 | |
| 2019 | AAAIW | Unsupervised Learning of Neural Networks to Explain Neural Networks | 9 | |
| 2019 | AAAIW | Network Transplanting | 4 | |
| 2019 | CSUR | A Survey of Methods for Explaining Black Box Models | 455 | |
| 2019 | JVCIR | Interpretable convolutional neural networks via feedforward design | 28 | Keras |
| 2019 | ExplainAI | The (Un)reliability of saliency methods(scihub) | 115 | |
| 2019 | ACL | Attention is not Explanation | 75 | |
| 2019 | arxiv | Attention Interpretability Across NLP Tasks | 5 | |
| 2019 | arxiv | Interpretable CNNs | 2 | |
| 2018 | ICLR | Towards better understanding of gradient-based attribution methods for deep neural networks | 151 | |
| 2018 | ICLR | Learning how to explain neural networks: PatternNet and PatternAttribution | 111 | |
| 2018 | ICLR | On the importance of single directions for generalization | 102 | Pytorch |
| 2018 | ICLR | Detecting statistical interactions from neural network weights | 42 | Pytorch |
| 2018 | ICLR | Interpretable counting for visual question answering | 25 | Pytorch |
| 2018 | CVPR | Interpretable Convolutional Neural Networks | 190 | |
| 2018 | CVPR | Tell me where to look: Guided attention inference network | 98 | Chainer |
| 2018 | CVPR | Multimodal Explanations: Justifying Decisions and Pointing to the Evidence | 94 | Caffe |
| 2018 | CVPR | Transparency by design: Closing the gap between performance and interpretability in visual reasoning | 66 | Pytorch |
| 2018 | CVPR | Net2vec: Quantifying and explaining how concepts are encoded by filters in deep neural networks | 45 | |
| 2018 | CVPR | What have we learned from deep representations for action recognition? | 23 | |
| 2018 | CVPR | Learning to Act Properly: Predicting and Explaining Affordances from Images | 21 | |
| 2018 | CVPR | Teaching Categories to Human Learners with Visual Explanations | 17 | Pytorch |
| 2018 | CVPR | What do Deep Networks Like to See? | 12 | |
| 2018 | CVPR | Interpret Neural Networks by Identifying Critical Data Routing Paths | 10 | Tensorflow |
| 2018 | ECCV | Deep clustering for unsupervised learning of visual features | 245 | Pytorch |
| 2018 | ECCV | Explainable neural computation via stack neural module networks | 43 | Tensorflow |
| 2018 | ECCV | Grounding visual explanations | 34 | |
| 2018 | ECCV | Textual explanations for self-driving vehicles | 42 | |
| 2018 | ECCV | Interpretable basis decomposition for visual explanation | 32 | Pytorch |
| 2018 | ECCV | Convnets and imagenet beyond accuracy: Understanding mistakes and uncovering biases | 20 | |
| 2018 | ECCV | Vqa-e: Explaining, elaborating, and enhancing your answers for visual questions | 13 | |
| 2018 | ECCV | Choose Your Neuron: Incorporating Domain Knowledge through Neuron-Importance | 13 | Pytorch |
| 2018 | ECCV | Diverse feature visualizations reveal invariances in early layers of deep neural networks | 7 | Tensorflow |
| 2018 | ECCV | ExplainGAN: Model Explanation via Decision Boundary Crossing Transformations | 1 | |
| 2018 | ICML | Interpretability beyond feature attribution: Quantitative testing with concept activation vectors (tcav) | 148 | Tensorflow |
| 2018 | ICML | Learning to explain: An information-theoretic perspective on model interpretation | 88 | |
| 2018 | ACL | Did the Model Understand the Question? | 39 | Tensorflow |
| 2018 | FITEE | Visual interpretability for deep learning: a survey | 183 | |
| 2018 | NeurIPS | Sanity Checks for Saliency Maps | 165 | |
| 2018 | NeurIPS | Explanations based on the missing: Towards contrastive explanations with pertinent negatives | 47 | Tensorflow |
| 2018 | NeurIPS | Towards robust interpretability with self-explaining neural networks | 94 | Pytorch |
| 2018 | NeurIPS | Attacks meet interpretability: Attribute-steered detection of adversarial samples | 35 | |
| 2018 | NeurIPS Workshop | Interpretable Convolutional Filters with SincNet | 27 | |
| 2018 | NeurIPS | DeepPINK: reproducible feature selection in deep neural networks | 21 | Keras |
| 2018 | NeurIPS | Representer point selection for explaining deep neural networks | 15 | Tensorflow |
| 2018 | AAAI | Anchors: High-precision model-agnostic explanations | 269 | |
| 2018 | AAAI | Improving the adversarial robustness and interpretability of deep neural networks by regularizing their input gradients | 138 | Tensorflow |
| 2018 | AAAI | Deep learning for case-based reasoning through prototypes: A neural network that explains its predictions | 80 | Tensorflow |
| 2018 | AAAI | Interpreting CNN Knowledge via an Explanatory Graph | 67 | Matlab |
| 2018 | AAAI | Examining CNN Representations with respect to Dataset Bias | 32 | |
| 2018 | WACV | Grad-cam++: Generalized gradient-based visual explanations for deep convolutional networks | 113 | |
| 2018 | IJCV | Top-down neural attention by excitation backprop | 285 | |
| 2018 | TPAMI | Interpreting deep visual representations via network dissection | 69 | |
| 2018 | DSP | Methods for interpreting and understanding deep neural networks(scihub) | 568 | |
| 2018 | Access | Peeking inside the black-box: A survey on Explainable Artificial Intelligence (XAI) | 228 | |
| 2018 | JAIR | Learning Explanatory Rules from Noisy Data | 111 | Tensorflow |
| 2018 | MIPRO | Explainable artificial intelligence: A survey | 82 | |
| 2018 | AIES | Detecting Bias in Black-Box Models Using Transparent Model Distillation | 28 | |
| 2018 | BMVC | Rise: Randomized input sampling for explanation of black-box models | 46 | |
| 2018 | arxiv | Manipulating and measuring model interpretability | 101 | |
| 2018 | arxiv | How convolutional neural network see the world-A survey of convolutional neural network visualization methods | 33 | |
| 2018 | arxiv | Revisiting the importance of individual units in cnns via ablation | 31 | |
| 2018 | arxiv | Computationally Efficient Measures of Internal Neuron Importance | 1 | |
| 2017 | ICML | Understanding Black-box Predictions via Influence Functions | 610 | Pytorch |
| 2017 | ICML | Axiomatic attribution for deep networks | 561 | Keras |
| 2017 | ICML | Learning Important Features Through Propagating Activation Differences | 489 | |
| 2017 | ICLR | Visualizing deep neural network decisions: Prediction difference analysis | 231 | Caffe |
| 2017 | ICLR | Exploring LOTS in Deep Neural Networks | 26 | |
| 2017 | NeurIPS | A Unified Approach to Interpreting Model Predictions | 894 | |
| 2017 | NeurIPS | Real time image saliency for black box classifiers | 128 | Pytorch |
| 2017 | NeurIPS | SVCCA: Singular Vector Canonical Correlation Analysis for Deep Learning Dynamics and Interpretability | 122 | |
| 2017 | CVPR | Mining Object Parts from CNNs via Active Question-Answering | 16 | |
| 2017 | CVPR | Network dissection: Quantifying interpretability of deep visual representations | 435 | |
| 2017 | CVPR | Improving Interpretability of Deep Neural Networks with Semantic Information | 52 | |
| 2017 | CVPR | MDNet: A Semantically and Visually Interpretable Medical Image Diagnosis Network | 100 | Torch |
| 2017 | CVPR | Interpretable 3d human action analysis with temporal convolutional networks | 133 | |
| 2017 | CVPR | Making the V in VQA matter: Elevating the role of image understanding in Visual Question Answering | 454 | |
| 2017 | CVPR | Knowing when to look: Adaptive attention via a visual sentinel for image captioning | 525 | Torch |
| 2017 | ICCV | Grad-cam: Visual explanations from deep networks via gradient-based localization | 1755 | Pytorch |
| 2017 | ICCV | Interpretable Explanations of Black Boxes by Meaningful Perturbation | 339 | Pytorch |
| 2017 | ICCV | Interpretable Learning for Self-Driving Cars by Visualizing Causal Attention | 96 | |
| 2017 | ICCV | Understanding and comparing deep neural networks for age and gender classification | 48 | |
| 2017 | ICCV | Learning to disambiguate by asking discriminative questions | 11 | |
| 2017 | IJCAI | Right for the right reasons: Training differentiable models by constraining their explanations | 118 | |
| 2017 | IJCAI | Understanding and improving convolutional neural networks via concatenated rectified linear units | 237 | Caffe |
| 2017 | AAAI | Growing Interpretable Part Graphs on ConvNets via Multi-Shot Learning | 28 | Matlab |
| 2017 | ACL | Visualizing and Understanding Neural Machine Translation | 63 | |
| 2017 | EMNLP | A causal framework for explaining the predictions of black-box sequence-to-sequence models | 68 | |
| 2017 | CVPR Workshop | Looking under the hood: Deep neural network visualization to interpret whole-slide image analysis outcomes for colorectal polyps | 16 | |
| 2017 | survey | Interpretability of deep learning models: a survey of results | 63 | |
| 2017 | arxiv | SmoothGrad: removing noise by adding noise | 265 | |
| 2017 | arxiv | Interpretable & explorable approximations of black box models | 91 | |
| 2017 | arxiv | Distilling a neural network into a soft decision tree | 153 | Pytorch |
| 2017 | arxiv | Towards interpretable deep neural networks by leveraging adversarial examples | 52 | |
| 2017 | arxiv | Explainable artificial intelligence: Understanding, visualizing and interpreting deep learning models | 294 | |
| 2017 | arxiv | Contextual Explanation Networks | 32 | Pytorch |
| 2017 | arxiv | Challenges for transparency | 74 | |
| 2017 | ACMSOPP | Deepxplore: Automated whitebox testing of deep learning systems | 364 | |
| 2017 | CEURW | What does explainable AI really mean? A new conceptualization of perspectives | 83 | |
| 2017 | TVCG | ActiVis: Visual Exploration of Industry-Scale Deep Neural Network Models | 131 | |
| 2016 | NeurIPS | Synthesizing the preferred inputs for neurons in neural networks via deep generator networks | 277 | Caffe |
| 2016 | NeurIPS | Understanding the effective receptive field in deep convolutional neural networks | 358 | |
| 2016 | CVPR | Inverting Visual Representations with Convolutional Networks | 295 | |
| 2016 | CVPR | Visualizing and Understanding Deep Texture Representations | 89 | |
| 2016 | CVPR | Analyzing Classifiers: Fisher Vectors and Deep Neural Networks | 94 | |
| 2016 | ECCV | Generating Visual Explanations | 256 | Caffe |
| 2016 | ECCV | Design of kernels in convolutional neural networks for image classification | 13 | |
| 2016 | ICML | Understanding and improving convolutional neural networks via concatenated rectified linear units | 237 | |
| 2016 | ICML | Visualizing and comparing AlexNet and VGG using deconvolutional layers | 31 | |
| 2016 | EMNLP | Rationalizing Neural Predictions | 276 | Pytorch |
| 2016 | IJCV | Visualizing deep convolutional neural networks using natural pre-images | 243 | Matlab |
| 2016 | IJCV | Visualizing Object Detection Features | 23 | Caffe |
| 2016 | KDD | Why should i trust you?: Explaining the predictions of any classifier | 2766 | |
| 2016 | TVCG | Visualizing the hidden activity of artificial neural networks | 143 | |
| 2016 | TVCG | Towards better analysis of deep convolutional neural networks | 210 | |
| 2016 | NAACL | Visualizing and understanding neural models in nlp | 292 | Torch |
| 2016 | arxiv | Understanding neural networks through representation erasure | 150 | |
| 2016 | arxiv | Grad-CAM: Why did you say that? | 106 | |
| 2016 | arxiv | Investigating the influence of noise and distractors on the interpretation of neural networks | 29 | |
| 2016 | arxiv | Attentive Explanations: Justifying Decisions and Pointing to the Evidence | 46 | |
| 2016 | arxiv | The Mythos of Model Interpretability | 1108 | |
| 2016 | arxiv | Multifaceted feature visualization: Uncovering the different types of features learned by each neuron in deep neural networks | 140 | |
| 2015 | ICLR | Striving for Simplicity: The All Convolutional Net | 1998 | Pytorch |
| 2015 | CVPR | Understanding deep image representations by inverting them | 1009 | Matlab |
| 2015 | ICCV | Understanding deep features with computer-generated imagery | 100 | Caffe |
| 2015 | ICML Workshop | Understanding Neural Networks Through Deep Visualization | 1072 | Tensorflow |
| 2015 | AAS | Interpretable classifiers using rules and Bayesian analysis: Building a better stroke prediction model | 331 | |
| 2014 | ECCV | Visualizing and Understanding Convolutional Networks | 8856 | Pytorch |
| 2014 | ICLR | Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps | 2277 | Pytorch |
- 论文talk