├── .gitignore ├── LICENSE ├── README.md └── sort_cite.md /.gitignore: -------------------------------------------------------------------------------- 1 | # Byte-compiled / optimized / DLL files 2 | __pycache__/ 3 | *.py[cod] 4 | *$py.class 5 | 6 | # C extensions 7 | *.so 8 | 9 | # Distribution / packaging 10 | .Python 11 | build/ 12 | develop-eggs/ 13 | dist/ 14 | downloads/ 15 | eggs/ 16 | .eggs/ 17 | lib/ 18 | lib64/ 19 | parts/ 20 | sdist/ 21 | var/ 22 | wheels/ 23 | pip-wheel-metadata/ 24 | share/python-wheels/ 25 | *.egg-info/ 26 | .installed.cfg 27 | *.egg 28 | MANIFEST 29 | 30 | # PyInstaller 31 | # Usually these files are written by a python script from a template 32 | # before PyInstaller builds the exe, so as to inject date/other infos into it. 33 | *.manifest 34 | *.spec 35 | 36 | # Installer logs 37 | pip-log.txt 38 | pip-delete-this-directory.txt 39 | 40 | # Unit test / coverage reports 41 | htmlcov/ 42 | .tox/ 43 | .nox/ 44 | .coverage 45 | .coverage.* 46 | .cache 47 | nosetests.xml 48 | coverage.xml 49 | *.cover 50 | *.py,cover 51 | .hypothesis/ 52 | .pytest_cache/ 53 | 54 | # Translations 55 | *.mo 56 | *.pot 57 | 58 | # Django stuff: 59 | *.log 60 | local_settings.py 61 | db.sqlite3 62 | db.sqlite3-journal 63 | 64 | # Flask stuff: 65 | instance/ 66 | .webassets-cache 67 | 68 | # Scrapy stuff: 69 | .scrapy 70 | 71 | # Sphinx documentation 72 | docs/_build/ 73 | 74 | # PyBuilder 75 | target/ 76 | 77 | # Jupyter Notebook 78 | .ipynb_checkpoints 79 | 80 | # IPython 81 | profile_default/ 82 | ipython_config.py 83 | 84 | # pyenv 85 | .python-version 86 | 87 | # pipenv 88 | # According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control. 89 | # However, in case of collaboration, if having platform-specific dependencies or dependencies 90 | # having no cross-platform support, pipenv may install dependencies that don't work, or not 91 | # install all needed dependencies. 92 | #Pipfile.lock 93 | 94 | # PEP 582; used by e.g. github.com/David-OConnor/pyflow 95 | __pypackages__/ 96 | 97 | # Celery stuff 98 | celerybeat-schedule 99 | celerybeat.pid 100 | 101 | # SageMath parsed files 102 | *.sage.py 103 | 104 | # Environments 105 | .env 106 | .venv 107 | env/ 108 | venv/ 109 | ENV/ 110 | env.bak/ 111 | venv.bak/ 112 | 113 | # Spyder project settings 114 | .spyderproject 115 | .spyproject 116 | 117 | # Rope project settings 118 | .ropeproject 119 | 120 | # mkdocs documentation 121 | /site 122 | 123 | # mypy 124 | .mypy_cache/ 125 | .dmypy.json 126 | dmypy.json 127 | 128 | # Pyre type checker 129 | .pyre/ 130 | -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- 1 | MIT License 2 | 3 | Copyright (c) 2019 oneTaken 4 | 5 | Permission is hereby granted, free of charge, to any person obtaining a copy 6 | of this software and associated documentation files (the "Software"), to deal 7 | in the Software without restriction, including without limitation the rights 8 | to use, copy, modify, merge, publish, distribute, sublicense, and/or sell 9 | copies of the Software, and to permit persons to whom the Software is 10 | furnished to do so, subject to the following conditions: 11 | 12 | The above copyright notice and this permission notice shall be included in all 13 | copies or substantial portions of the Software. 14 | 15 | THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 16 | IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 17 | FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE 18 | AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER 19 | LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, 20 | OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE 21 | SOFTWARE. 22 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | 2 | 3 | 4 | 5 | # awesome_deep_learning_interpretability 6 | 深度学习近年来关于模型解释性的相关论文。 7 | 8 | 按引用次数排序可见[引用排序](./sort_cite.md) 9 | 10 | 159篇论文pdf(有2篇需要上scihub找)上传到[腾讯微云](https://share.weiyun.com/5ddB0EQ)。 11 | 12 | 不定期更新。 13 | 14 | |Year|Publication|Paper|Citation|code| 15 | |:---:|:---:|:---:|:---:|:---:| 16 | |2020|CVPR|[Explaining Knowledge Distillation by Quantifying the Knowledge](https://arxiv.org/pdf/2003.03622.pdf)|81| 17 | |2020|CVPR|[High-frequency Component Helps Explain the Generalization of Convolutional Neural Networks](https://openaccess.thecvf.com/content_CVPR_2020/papers/Wang_High-Frequency_Component_Helps_Explain_the_Generalization_of_Convolutional_Neural_Networks_CVPR_2020_paper.pdf)|289| 18 | |2020|CVPRW|[Score-CAM: Score-Weighted Visual Explanations for Convolutional Neural Networks](https://openaccess.thecvf.com/content_CVPRW_2020/papers/w1/Wang_Score-CAM_Score-Weighted_Visual_Explanations_for_Convolutional_Neural_Networks_CVPRW_2020_paper.pdf)|414|[Pytorch](https://github.com/haofanwang/Score-CAM) 19 | |2020|ICLR|[Knowledge consistency between neural networks and beyond](https://arxiv.org/pdf/1908.01581.pdf)|28| 20 | |2020|ICLR|[Interpretable Complex-Valued Neural Networks for Privacy Protection](https://arxiv.org/pdf/1901.09546.pdf)|23| 21 | |2019|AI|[Explanation in artificial intelligence: Insights from the social sciences](https://arxiv.org/pdf/1706.07269.pdf)|3248| 22 | |2019|NMI|[Stop Explaining Black Box Machine Learning Models for High Stakes Decisions and Use Interpretable Models Instead](https://arxiv.org/pdf/1811.10154.pdf)|3505| 23 | |2019|NeurIPS|[Can you trust your model's uncertainty? Evaluating predictive uncertainty under dataset shift](https://papers.nips.cc/paper/9547-can-you-trust-your-models-uncertainty-evaluating-predictive-uncertainty-under-dataset-shift.pdf)|1052|-| 24 | |2019|NeurIPS|[This looks like that: deep learning for interpretable image recognition](http://papers.nips.cc/paper/9095-this-looks-like-that-deep-learning-for-interpretable-image-recognition.pdf)|665|[Pytorch](https://github.com/cfchen-duke/ProtoPNet)| 25 | |2019|NeurIPS|[A benchmark for interpretability methods in deep neural networks](https://papers.nips.cc/paper/9167-a-benchmark-for-interpretability-methods-in-deep-neural-networks.pdf)|413| 26 | |2019|NeurIPS|[Full-gradient representation for neural network visualization](http://papers.nips.cc/paper/8666-full-gradient-representation-for-neural-network-visualization.pdf)|155| 27 | |2019|NeurIPS|[On the (In) fidelity and Sensitivity of Explanations](https://papers.nips.cc/paper/9278-on-the-infidelity-and-sensitivity-of-explanations.pdf)|226| 28 | |2019|NeurIPS|[Towards Automatic Concept-based Explanations](http://papers.nips.cc/paper/9126-towards-automatic-concept-based-explanations.pdf)|342|[Tensorflow](https://github.com/amiratag/ACE)| 29 | |2019|NeurIPS|[CXPlain: Causal explanations for model interpretation under uncertainty](http://papers.nips.cc/paper/9211-cxplain-causal-explanations-for-model-interpretation-under-uncertainty.pdf)|133| 30 | |2019|CVPR|[Interpreting CNNs via Decision Trees](http://openaccess.thecvf.com/content_CVPR_2019/papers/Zhang_Interpreting_CNNs_via_Decision_Trees_CVPR_2019_paper.pdf)|293| 31 | |2019|CVPR|[From Recognition to Cognition: Visual Commonsense Reasoning](http://openaccess.thecvf.com/content_CVPR_2019/papers/Zellers_From_Recognition_to_Cognition_Visual_Commonsense_Reasoning_CVPR_2019_paper.pdf)|544|[Pytorch](https://github.com/rowanz/r2c)| 32 | |2019|CVPR|[Attention branch network: Learning of attention mechanism for visual explanation](http://openaccess.thecvf.com/content_CVPR_2019/papers/Fukui_Attention_Branch_Network_Learning_of_Attention_Mechanism_for_Visual_Explanation_CVPR_2019_paper.pdf)|371| 33 | |2019|CVPR|[Interpretable and fine-grained visual explanations for convolutional neural networks](http://openaccess.thecvf.com/content_CVPR_2019/papers/Wagner_Interpretable_and_Fine-Grained_Visual_Explanations_for_Convolutional_Neural_Networks_CVPR_2019_paper.pdf)|116| 34 | |2019|CVPR|[Learning to Explain with Complemental Examples](http://openaccess.thecvf.com/content_CVPR_2019/papers/Kanehira_Learning_to_Explain_With_Complemental_Examples_CVPR_2019_paper.pdf)|36| 35 | |2019|CVPR|[Revealing Scenes by Inverting Structure from Motion Reconstructions](http://openaccess.thecvf.com/content_CVPR_2019/papers/Pittaluga_Revealing_Scenes_by_Inverting_Structure_From_Motion_Reconstructions_CVPR_2019_paper.pdf)|84|[Tensorflow](https://github.com/francescopittaluga/invsfm)| 36 | |2019|CVPR|[Multimodal Explanations by Predicting Counterfactuality in Videos](http://openaccess.thecvf.com/content_CVPR_2019/papers/Kanehira_Multimodal_Explanations_by_Predicting_Counterfactuality_in_Videos_CVPR_2019_paper.pdf)|26| 37 | |2019|CVPR|[Visualizing the Resilience of Deep Convolutional Network Interpretations](http://openaccess.thecvf.com/content_CVPRW_2019/papers/Explainable%20AI/Vasu_Visualizing_the_Resilience_of_Deep_Convolutional_Network_Interpretations_CVPRW_2019_paper.pdf)|2| 38 | |2019|ICCV|[U-CAM: Visual Explanation using Uncertainty based Class Activation Maps](http://openaccess.thecvf.com/content_ICCV_2019/papers/Patro_U-CAM_Visual_Explanation_Using_Uncertainty_Based_Class_Activation_Maps_ICCV_2019_paper.pdf)|61| 39 | |2019|ICCV|[Towards Interpretable Face Recognition](https://arxiv.org/pdf/1805.00611.pdf)|66| 40 | |2019|ICCV|[Taking a HINT: Leveraging Explanations to Make Vision and Language Models More Grounded](http://openaccess.thecvf.com/content_ICCV_2019/papers/Selvaraju_Taking_a_HINT_Leveraging_Explanations_to_Make_Vision_and_Language_ICCV_2019_paper.pdf)|163| 41 | |2019|ICCV|[Understanding Deep Networks via Extremal Perturbations and Smooth Masks](http://openaccess.thecvf.com/content_ICCV_2019/papers/Fong_Understanding_Deep_Networks_via_Extremal_Perturbations_and_Smooth_Masks_ICCV_2019_paper.pdf)|276|[Pytorch](https://github.com/facebookresearch/TorchRay)| 42 | |2019|ICCV|[Explaining Neural Networks Semantically and Quantitatively](http://openaccess.thecvf.com/content_ICCV_2019/papers/Chen_Explaining_Neural_Networks_Semantically_and_Quantitatively_ICCV_2019_paper.pdf)|49| 43 | |2019|ICLR|[Hierarchical interpretations for neural network predictions](https://arxiv.org/pdf/1806.05337.pdf)|111|[Pytorch](https://github.com/csinva/hierarchical-dnn-interpretations)| 44 | |2019|ICLR|[How Important Is a Neuron?](https://arxiv.org/pdf/1805.12233.pdf)|101| 45 | |2019|ICLR|[Visual Explanation by Interpretation: Improving Visual Feedback Capabilities of Deep Neural Networks](https://arxiv.org/pdf/1712.06302.pdf)|56| 46 | |2018|ICML|[Extracting Automata from Recurrent Neural Networks Using Queries and Counterexamples](https://arxiv.org/pdf/1711.09576.pdf)|169|[Pytorch](https://github.com/tech-srl/lstar_extraction)| 47 | |2019|ICML|[Towards A Deep and Unified Understanding of Deep Neural Models in NLP](http://proceedings.mlr.press/v97/guan19a/guan19a.pdf)|80|[Pytorch](https://github.com/icml2019paper2428/Towards-A-Deep-and-Unified-Understanding-of-Deep-Neural-Models-in-NLP)| 48 | |2019|ICAIS|[Interpreting black box predictions using fisher kernels](https://arxiv.org/pdf/1810.10118.pdf)|80| 49 | |2019|ACMFAT|[Explaining explanations in AI](https://s3.amazonaws.com/academia.edu.documents/57692790/Mittelstadt__Russell_and_Wachter_-_2019_-_Explaining_Explanations_in_AI.pdf?response-content-disposition=inline%3B%20filename%3DExplaining_Explanations_in_AI.pdf&X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=ASIATUSBJ6BAJW2TMFXG%2F20200528%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Date=20200528T052420Z&X-Amz-Expires=3600&X-Amz-SignedHeaders=host&X-Amz-Security-Token=IQoJb3JpZ2luX2VjEHUaCXVzLWVhc3QtMSJIMEYCIQDCCKV%2FpUmJZHn03yzTquQ%2FNMtaXW%2FC63WPmQd%2FhImmYAIhAMelsFwqb9IfV4W2xlfL%2FHk4qeovouLdYbXKf%2B1%2FMwvyKr0DCM7%2F%2F%2F%2F%2F%2F%2F%2F%2F%2FwEQABoMMjUwMzE4ODExMjAwIgytA%2BM6OWOGN4XLrlUqkQN2f8ywZT0AEUzKdbVDyGvZN%2B1repdgXrfgT2rAJiGacTK8IRCoyECvRgcgS%2BWJWYpjS7CjoL%2BlTm1c%2BWDWdo%2FYnVM0U6shk9OQivK089W064ZR64AQCCkBDutI3vYhP%2BOJ8AtEUDE%2B7W5EWVQ4zeUDG4ryxzdomFnrHpzA5fp05qWrOmPS0vd%2FFabC%2FPKXO34bpfgyRzz3PHrIsUC2%2BPB0EAo7CPKS0Ux%2FlxmiIOYOIj5u1ZKoP8NVLgOfueQe7%2F%2F3VJUnUXSAIsAThszDTnbi0AJEjvNvUHjm8E%2F7zqBApJ6YVd39NkKl8%2BTE7MRwKuITAOIq8jsyta%2FcmIY5igpHpVCkYcG395rHfScDu3CODXIAcKRLX%2F7brNz%2FRHuGhddK3Q2XuGTjQaeLTEYTmTj2e7VDDmEOt%2BpxvXx7UaImPakzpVZ1Ks6APy1JHupKgBhM6JJkeFprlK62e4sf09wqwxk9KsJSot3TMLVwM63yGr7VmXdg61ETsg0D%2BO1DOnnMprsFhEkb%2Bt%2FpCVafebolsjCN%2Frz2BTrqAZiqy6Obte6J%2BeHJ5bzB1sy1oF%2Fi7ueF56nd1C9ObB%2FXLx930j8wqmakO%2FnoaUiYM6gHh1jZbl8cCeLr8Xu0YSGecpe1J5HECU0A5%2Fq68zoBDfyY6UGNZJ%2B87Br6crqpfaHFkP5g4zXvuN2%2F0fp6S9m2iuSRBr%2B%2Bh2Z1rXmvb3Vequ2qgqeJBS2nHOX8pLp2LhJsVMqdl218jeQDsjYnbxJKq86peVGr66Cuv7TmNiimVl0c0dPr1jgjr25N9hvMnpX83n2Xa%2Fz%2BHUmaYfwFLrD0YLkUWaS2Khcpm0%2BwvrcYsQEyOmYkVG8x5Q%3D%3D&X-Amz-Signature=4fcca52f4ae92746068ea2164846aca05c2bb44e04c1330947ba70f75e676171)|558| 50 | |2019|AAAI|[Interpretation of neural networks is fragile](https://machine-learning-and-security.github.io/papers/mlsec17_paper_18.pdf)|597|[Tensorflow](https://github.com/amiratag/InterpretationFragility)| 51 | |2019|AAAI|[Classifier-agnostic saliency map extraction](https://arxiv.org/pdf/1805.08249.pdf)|23| 52 | |2019|AAAI|[Can You Explain That? Lucid Explanations Help Human-AI Collaborative Image Retrieval](https://arxiv.org/pdf/1904.03285.pdf)|11| 53 | |2019|AAAIW|[Unsupervised Learning of Neural Networks to Explain Neural Networks](https://arxiv.org/pdf/1805.07468.pdf)|28| 54 | |2019|AAAIW|[Network Transplanting](https://arxiv.org/pdf/1804.10272.pdf)|4| 55 | |2019|CSUR|[A Survey of Methods for Explaining Black Box Models](https://kdd.isti.cnr.it/sites/kdd.isti.cnr.it/files/csur2018survey.pdf)|3088| 56 | |2019|JVCIR|[Interpretable convolutional neural networks via feedforward design](https://arxiv.org/pdf/1810.02786)|134|[Keras](https://github.com/davidsonic/Interpretable_CNNs_via_Feedforward_Design)| 57 | |2019|ExplainAI|[The (Un)reliability of saliency methods](https://arxiv.org/pdf/1711.00867.pdf)|515| 58 | |2019|ACL|[Attention is not Explanation](https://arxiv.org/pdf/1902.10186.pdf)|920| 59 | |2019|EMNLP|[Attention is not not Explanation](https://arxiv.org/pdf/1908.04626.pdf)|667| 60 | |2019|arxiv|[Attention Interpretability Across NLP Tasks](https://arxiv.org/pdf/1909.11218.pdf)|129| 61 | |2019|arxiv|[Interpretable CNNs](https://arxiv.org/pdf/1901.02413.pdf)|2| 62 | |2018|ICLR|[Towards better understanding of gradient-based attribution methods for deep neural networks](https://arxiv.org/pdf/1711.06104.pdf)|775| 63 | |2018|ICLR|[Learning how to explain neural networks: PatternNet and PatternAttribution](https://arxiv.org/pdf/1705.05598.pdf)|342| 64 | |2018|ICLR|[On the importance of single directions for generalization](https://arxiv.org/pdf/1803.06959.pdf)|282|[Pytorch](https://github.com/1Konny/class_selectivity_index)| 65 | |2018|ICLR|[Detecting statistical interactions from neural network weights](https://arxiv.org/pdf/1705.04977.pdf)|148|[Pytorch](https://github.com/mtsang/neural-interaction-detection)| 66 | |2018|ICLR|[Interpretable counting for visual question answering](https://arxiv.org/pdf/1712.08697.pdf)|55|[Pytorch](https://github.com/sanyam5/irlc-vqa-counting)| 67 | |2018|CVPR|[Interpretable Convolutional Neural Networks](http://openaccess.thecvf.com/content_cvpr_2018/papers/Zhang_Interpretable_Convolutional_Neural_CVPR_2018_paper.pdf)|677| 68 | |2018|CVPR|[Tell me where to look: Guided attention inference network](http://openaccess.thecvf.com/content_cvpr_2018/papers/Li_Tell_Me_Where_CVPR_2018_paper.pdf)|454|[Chainer](https://github.com/alokwhitewolf/Guided-Attention-Inference-Network)| 69 | |2018|CVPR|[Multimodal Explanations: Justifying Decisions and Pointing to the Evidence](http://openaccess.thecvf.com/content_cvpr_2018/papers/Park_Multimodal_Explanations_Justifying_CVPR_2018_paper.pdf)|349|[Caffe](https://github.com/Seth-Park/MultimodalExplanations)| 70 | |2018|CVPR|[Transparency by design: Closing the gap between performance and interpretability in visual reasoning](http://openaccess.thecvf.com/content_cvpr_2018/papers/Mascharka_Transparency_by_Design_CVPR_2018_paper.pdf)|180|[Pytorch](https://github.com/davidmascharka/tbd-nets)| 71 | |2018|CVPR|[Net2vec: Quantifying and explaining how concepts are encoded by filters in deep neural networks](http://openaccess.thecvf.com/content_cvpr_2018/papers/Fong_Net2Vec_Quantifying_and_CVPR_2018_paper.pdf)|186| 72 | |2018|CVPR|[What have we learned from deep representations for action recognition?](http://openaccess.thecvf.com/content_cvpr_2018/papers/Feichtenhofer_What_Have_We_CVPR_2018_paper.pdf)|52| 73 | |2018|CVPR|[Learning to Act Properly: Predicting and Explaining Affordances from Images](http://openaccess.thecvf.com/content_cvpr_2018/papers/Chuang_Learning_to_Act_CVPR_2018_paper.pdf)|57| 74 | |2018|CVPR|[Teaching Categories to Human Learners with Visual Explanations](http://openaccess.thecvf.com/content_cvpr_2018/papers/Aodha_Teaching_Categories_to_CVPR_2018_paper.pdf)|64|[Pytorch](https://github.com/macaodha/explain_teach)| 75 | |2018|CVPR|[What do deep networks like to see?](http://openaccess.thecvf.com/content_cvpr_2018/papers/Palacio_What_Do_Deep_CVPR_2018_paper.pdf)|36| 76 | |2018|CVPR|[Interpret Neural Networks by Identifying Critical Data Routing Paths](http://openaccess.thecvf.com/content_cvpr_2018/papers/Wang_Interpret_Neural_Networks_CVPR_2018_paper.pdf)|73|[Tensorflow](https://github.com/lidongyue12138/CriticalPathPruning)| 77 | |2018|ECCV|[Deep clustering for unsupervised learning of visual features](http://openaccess.thecvf.com/content_ECCV_2018/papers/Mathilde_Caron_Deep_Clustering_for_ECCV_2018_paper.pdf)|2056|[Pytorch](https://github.com/asanakoy/deep_clustering)| 78 | |2018|ECCV|[Explainable neural computation via stack neural module networks](http://openaccess.thecvf.com/content_ECCV_2018/papers/Ronghang_Hu_Explainable_Neural_Computation_ECCV_2018_paper.pdf)|164|[Tensorflow](https://github.com/ronghanghu/snmn)| 79 | |2018|ECCV|[Grounding visual explanations](http://openaccess.thecvf.com/content_ECCV_2018/papers/Lisa_Anne_Hendricks_Grounding_Visual_Explanations_ECCV_2018_paper.pdf)|184| 80 | |2018|ECCV|[Textual explanations for self-driving vehicles](http://openaccess.thecvf.com/content_ECCV_2018/papers/Jinkyu_Kim_Textual_Explanations_for_ECCV_2018_paper.pdf)|196| 81 | |2018|ECCV|[Interpretable basis decomposition for visual explanation](http://openaccess.thecvf.com/content_ECCV_2018/papers/Antonio_Torralba_Interpretable_Basis_Decomposition_ECCV_2018_paper.pdf)|228|[Pytorch](https://github.com/CSAILVision/IBD)| 82 | |2018|ECCV|[Convnets and imagenet beyond accuracy: Understanding mistakes and uncovering biases](http://openaccess.thecvf.com/content_ECCV_2018/papers/Pierre_Stock_ConvNets_and_ImageNet_ECCV_2018_paper.pdf)|147| 83 | |2018|ECCV|[Vqa-e: Explaining, elaborating, and enhancing your answers for visual questions](http://openaccess.thecvf.com/content_ECCV_2018/papers/Qing_Li_VQA-E_Explaining_Elaborating_ECCV_2018_paper.pdf)|71| 84 | |2018|ECCV|[Choose Your Neuron: Incorporating Domain Knowledge through Neuron-Importance](http://openaccess.thecvf.com/content_ECCV_2018/papers/Ramprasaath_Ramasamy_Selvaraju_Choose_Your_Neuron_ECCV_2018_paper.pdf)|41|[Pytorch](https://github.com/ramprs/neuron-importance-zsl)| 85 | |2018|ECCV|[Diverse feature visualizations reveal invariances in early layers of deep neural networks](http://openaccess.thecvf.com/content_ECCV_2018/papers/Santiago_Cadena_Diverse_feature_visualizations_ECCV_2018_paper.pdf)|23|[Tensorflow](https://github.com/sacadena/diverse_feature_vis)| 86 | |2018|ECCV|[ExplainGAN: Model Explanation via Decision Boundary Crossing Transformations](http://openaccess.thecvf.com/content_ECCV_2018/papers/Nathan_Silberman_ExplainGAN_Model_Explanation_ECCV_2018_paper.pdf)|36| 87 | |2018|ICML|[Interpretability beyond feature attribution: Quantitative testing with concept activation vectors](https://arxiv.org/pdf/1711.11279.pdf)|1130|[Tensorflow](https://github.com/fursovia/tcav_nlp)| 88 | |2018|ICML|[Learning to explain: An information-theoretic perspective on model interpretation](https://arxiv.org/pdf/1802.07814.pdf)|421| 89 | |2018|ACL|[Did the Model Understand the Question?](https://arxiv.org/pdf/1805.05492.pdf)|171|[Tensorflow](https://github.com/pramodkaushik/acl18_results)| 90 | |2018|FITEE|[Visual interpretability for deep learning: a survey](https://arxiv.org/pdf/1802.00614)|731| 91 | |2018|NeurIPS|[Sanity Checks for Saliency Maps](http://papers.nips.cc/paper/8160-sanity-checks-for-saliency-maps.pdf)|1353| 92 | |2018|NeurIPS|[Explanations based on the missing: Towards contrastive explanations with pertinent negatives](http://papers.nips.cc/paper/7340-explanations-based-on-the-missing-towards-contrastive-explanations-with-pertinent-negatives.pdf)|443|[Tensorflow](https://github.com/IBM/Contrastive-Explanation-Method)| 93 | |2018|NeurIPS|[Towards robust interpretability with self-explaining neural networks](http://papers.nips.cc/paper/8003-towards-robust-interpretability-with-self-explaining-neural-networks.pdf)|648|[Pytorch](https://github.com/raj-shah/senn)| 94 | |2018|NeurIPS|[Attacks meet interpretability: Attribute-steered detection of adversarial samples](https://papers.nips.cc/paper/7998-attacks-meet-interpretability-attribute-steered-detection-of-adversarial-samples.pdf)|142| 95 | |2018|NeurIPS|[DeepPINK: reproducible feature selection in deep neural networks](https://papers.nips.cc/paper/8085-deeppink-reproducible-feature-selection-in-deep-neural-networks.pdf)|125|[Keras](https://github.com/younglululu/DeepPINK)| 96 | |2018|NeurIPS|[Representer point selection for explaining deep neural networks](https://papers.nips.cc/paper/8141-representer-point-selection-for-explaining-deep-neural-networks.pdf)|182|[Tensorflow](https://github.com/chihkuanyeh/Representer_Point_Selection)| 97 | |2018|NeurIPS Workshop|[Interpretable convolutional filters with sincNet](https://arxiv.org/pdf/1811.09725)|97| 98 | |2018|AAAI|[Anchors: High-precision model-agnostic explanations](https://dm-gatech.github.io/CS8803-Fall2018-DML-Papers/anchors.pdf)|1517| 99 | |2018|AAAI|[Improving the adversarial robustness and interpretability of deep neural networks by regularizing their input gradients](https://asross.github.io/publications/RossDoshiVelez2018.pdf)|537|[Tensorflow](https://github.com/dtak/adversarial-robustness-public)| 100 | |2018|AAAI|[Deep learning for case-based reasoning through prototypes: A neural network that explains its predictions](https://arxiv.org/pdf/1710.04806.pdf)|396|[Tensorflow](https://github.com/OscarcarLi/PrototypeDL)| 101 | |2018|AAAI|[Interpreting CNN Knowledge via an Explanatory Graph](https://arxiv.org/pdf/1708.01785.pdf)|199|[Matlab](https://github.com/zqs1022/explanatoryGraph)| 102 | |2018|AAAI|[Examining CNN Representations with respect to Dataset Bias](http://www.stat.ucla.edu/~sczhu/papers/Conf_2018/AAAI_2018_DNN_Learning_Bias.pdf)|88| 103 | |2018|WACV|[Grad-cam++: Generalized gradient-based visual explanations for deep convolutional networks](https://www.researchgate.net/profile/Aditya_Chattopadhyay2/publication/320727679_Grad-CAM_Generalized_Gradient-based_Visual_Explanations_for_Deep_Convolutional_Networks/links/5a3aa2e5a6fdcc3889bd04cb/Grad-CAM-Generalized-Gradient-based-Visual-Explanations-for-Deep-Convolutional-Networks.pdf)|1459| 104 | |2018|IJCV|[Top-down neural attention by excitation backprop](https://arxiv.org/pdf/1608.00507)|778| 105 | |2018|TPAMI|[Interpreting deep visual representations via network dissection](https://arxiv.org/pdf/1711.05611)|252| 106 | |2018|DSP|[Methods for interpreting and understanding deep neural networks](http://iphome.hhi.de/samek/pdf/MonDSP18.pdf)|2046| 107 | |2018|Access|[Peeking inside the black-box: A survey on Explainable Artificial Intelligence (XAI)](https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=8466590)|3110| 108 | |2018|JAIR|[Learning Explanatory Rules from Noisy Data](https://www.ijcai.org/Proceedings/2018/0792.pdf)|440|[Tensorflow](https://github.com/ai-systems/DILP-Core)| 109 | |2018|MIPRO|[Explainable artificial intelligence: A survey](https://www.researchgate.net/profile/Mario_Brcic/publication/325398586_Explainable_Artificial_Intelligence_A_Survey/links/5b0bec90a6fdcc8c2534d673/Explainable-Artificial-Intelligence-A-Survey.pdf)|794| 110 | |2018|BMVC|[Rise: Randomized input sampling for explanation of black-box models](https://arxiv.org/pdf/1806.07421.pdf)|657| 111 | |2018|arxiv|[Distill-and-Compare: Auditing Black-Box Models Using Transparent Model Distillation](https://arxiv.org/pdf/1710.06169.pdf)|194| 112 | |2018|arxiv|[Manipulating and measuring model interpretability](https://arxiv.org/pdf/1802.07810.pdf)|496| 113 | |2018|arxiv|[How convolutional neural network see the world-A survey of convolutional neural network visualization methods](https://arxiv.org/pdf/1804.11191.pdf)|211| 114 | |2018|arxiv|[Revisiting the importance of individual units in cnns via ablation](https://arxiv.org/pdf/1806.02891.pdf)|93| 115 | |2018|arxiv|[Computationally Efficient Measures of Internal Neuron Importance](https://arxiv.org/pdf/1807.09946.pdf)|10| 116 | |2017|ICML|[Understanding Black-box Predictions via Influence Functions](https://dm-gatech.github.io/CS8803-Fall2018-DML-Papers/influence-functions.pdf)|2062|[Pytorch](https://github.com/nimarb/pytorch_influence_functions)| 117 | |2017|ICML|[Axiomatic attribution for deep networks](https://mit6874.github.io/assets/misc/sundararajan.pdf)|3654|[Keras](https://github.com/hiranumn/IntegratedGradients)| 118 | |2017|ICML|[Learning Important Features Through Propagating Activation Differences](https://mit6874.github.io/assets/misc/shrikumar.pdf)|2835| 119 | |2017|ICLR|[Visualizing deep neural network decisions: Prediction difference analysis](https://arxiv.org/pdf/1702.04595.pdf)|674|[Caffe](https://github.com/lmzintgraf/DeepVis-PredDiff)| 120 | |2017|ICLR|[Exploring LOTS in Deep Neural Networks](https://openreview.net/pdf?id=SkCILwqex)|34| 121 | |2017|NeurIPS|[A Unified Approach to Interpreting Model Predictions](http://papers.nips.cc/paper/7062-a-unified-approach-to-interpreting-model-predictions.pdf)|11511| 122 | |2017|NeurIPS|[Real time image saliency for black box classifiers](https://papers.nips.cc/paper/7272-real-time-image-saliency-for-black-box-classifiers.pdf)|483|[Pytorch](https://github.com/karanchahal/SaliencyMapper)| 123 | |2017|NeurIPS|[SVCCA: Singular Vector Canonical Correlation Analysis for Deep Learning Dynamics and Interpretability](http://papers.nips.cc/paper/7188-svcca-singular-vector-canonical-correlation-analysis-for-deep-learning-dynamics-and-interpretability.pdf)|473| 124 | |2017|CVPR|[Mining Object Parts from CNNs via Active Question-Answering](http://openaccess.thecvf.com/content_cvpr_2017/papers/Zhang_Mining_Object_Parts_CVPR_2017_paper.pdf)|29| 125 | |2017|CVPR|[Network dissection: Quantifying interpretability of deep visual representations](http://openaccess.thecvf.com/content_cvpr_2017/papers/Bau_Network_Dissection_Quantifying_CVPR_2017_paper.pdf)|1254| 126 | |2017|CVPR|[Improving Interpretability of Deep Neural Networks with Semantic Information](http://openaccess.thecvf.com/content_cvpr_2017/papers/Dong_Improving_Interpretability_of_CVPR_2017_paper.pdf)|118| 127 | |2017|CVPR|[MDNet: A Semantically and Visually Interpretable Medical Image Diagnosis Network](http://openaccess.thecvf.com/content_cvpr_2017/papers/Zhang_MDNet_A_Semantically_CVPR_2017_paper.pdf)|307|[Torch](https://github.com/zizhaozhang/mdnet-cvpr2017)| 128 | |2017|CVPR|[Making the V in VQA matter: Elevating the role of image understanding in Visual Question Answering](http://openaccess.thecvf.com/content_cvpr_2017/papers/Goyal_Making_the_v_CVPR_2017_paper.pdf)|1686| 129 | |2017|CVPR|[Knowing when to look: Adaptive attention via a visual sentinel for image captioning](http://openaccess.thecvf.com/content_cvpr_2017/papers/Lu_Knowing_When_to_CVPR_2017_paper.pdf)|1392|[Torch](https://github.com/jiasenlu/AdaptiveAttention)| 130 | |2017|CVPRW|[Interpretable 3d human action analysis with temporal convolutional networks](http://openaccess.thecvf.com/content_cvpr_2017_workshops/w20/papers/Kim_Interpretable_3D_Human_CVPR_2017_paper.pdf)|539| 131 | |2017|ICCV|[Grad-cam: Visual explanations from deep networks via gradient-based localization](http://openaccess.thecvf.com/content_ICCV_2017/papers/Selvaraju_Grad-CAM_Visual_Explanations_ICCV_2017_paper.pdf)|13006|[Pytorch](https://github.com/leftthomas/GradCAM)| 132 | |2017|ICCV|[Interpretable Explanations of Black Boxes by Meaningful Perturbation](http://openaccess.thecvf.com/content_ICCV_2017/papers/Fong_Interpretable_Explanations_of_ICCV_2017_paper.pdf)|1293|[Pytorch](https://github.com/jacobgil/pytorch-explain-black-box)| 133 | |2017|ICCV|[Interpretable Learning for Self-Driving Cars by Visualizing Causal Attention](http://openaccess.thecvf.com/content_ICCV_2017/papers/Kim_Interpretable_Learning_for_ICCV_2017_paper.pdf)|323| 134 | |2017|ICCV|[Understanding and comparing deep neural networks for age and gender classification](http://openaccess.thecvf.com/content_ICCV_2017_workshops/papers/w23/Lapuschkin_Understanding_and_Comparing_ICCV_2017_paper.pdf)|130| 135 | |2017|ICCV|[Learning to disambiguate by asking discriminative questions](http://openaccess.thecvf.com/content_ICCV_2017/papers/Li_Learning_to_Disambiguate_ICCV_2017_paper.pdf)|26| 136 | |2017|IJCAI|[Right for the right reasons: Training differentiable models by constraining their explanations](https://arxiv.org/pdf/1703.03717.pdf)|429| 137 | |2017|IJCAI|[Understanding and improving convolutional neural networks via concatenated rectified linear units](http://www.jmlr.org/proceedings/papers/v48/shang16.pdf)|510|[Caffe](https://github.com/chakkritte/CReLU)| 138 | |2017|AAAI|[Growing Interpretable Part Graphs on ConvNets via Multi-Shot Learning](https://arxiv.org/pdf/1611.04246.pdf)|67|[Matlab](https://github.com/zqs1022/partGraphForCNN)| 139 | |2017|ACL|[Visualizing and Understanding Neural Machine Translation](https://www.aclweb.org/anthology/P17-1106.pdf)|179| 140 | |2017|EMNLP|[A causal framework for explaining the predictions of black-box sequence-to-sequence models](https://arxiv.org/pdf/1707.01943.pdf)|192| 141 | |2017|CVPR Workshop|[Looking under the hood: Deep neural network visualization to interpret whole-slide image analysis outcomes for colorectal polyps](http://openaccess.thecvf.com/content_cvpr_2017_workshops/w8/papers/Korbar_Looking_Under_the_CVPR_2017_paper.pdf)|47| 142 | |2017|survey|[Interpretability of deep learning models: a survey of results](https://discovery.ucl.ac.uk/id/eprint/10059575/1/Chakraborty_Interpretability%20of%20deep%20learning%20models.pdf)|345| 143 | |2017|arxiv|[SmoothGrad: removing noise by adding noise](https://arxiv.org/pdf/1706.03825.pdf)|1479| 144 | |2017|arxiv|[Interpretable & explorable approximations of black box models](https://arxiv.org/pdf/1707.01154.pdf)|259| 145 | |2017|arxiv|[Distilling a neural network into a soft decision tree](https://arxiv.org/pdf/1711.09784.pdf)|520|[Pytorch](https://github.com/kimhc6028/soft-decision-tree)| 146 | |2017|arxiv|[Towards interpretable deep neural networks by leveraging adversarial examples](https://arxiv.org/pdf/1708.05493.pdf)|111| 147 | |2017|arxiv|[Explainable artificial intelligence: Understanding, visualizing and interpreting deep learning models](https://arxiv.org/pdf/1708.08296.pdf)|1279| 148 | |2017|arxiv|[Contextual Explanation Networks](https://arxiv.org/pdf/1705.10301.pdf)|77|[Pytorch](https://github.com/alshedivat/cen)| 149 | |2017|arxiv|[Challenges for transparency](https://arxiv.org/pdf/1708.01870.pdf)|142| 150 | |2017|ACMSOPP|[Deepxplore: Automated whitebox testing of deep learning systems](https://machine-learning-and-security.github.io/papers/mlsec17_paper_1.pdf)|1144| 151 | |2017|CEURW|[What does explainable AI really mean? A new conceptualization of perspectives](https://arxiv.org/pdf/1710.00794.pdf)|518| 152 | |2017|TVCG|[ActiVis: Visual Exploration of Industry-Scale Deep Neural Network Models](https://arxiv.org/pdf/1704.01942.pdf)|346| 153 | |2016|NeurIPS|[Synthesizing the preferred inputs for neurons in neural networks via deep generator networks](http://papers.nips.cc/paper/6519-synthesizing-the-preferred-inputs-for-neurons-in-neural-networks-via-deep-generator-networks.pdf)|659|[Caffe](https://github.com/Evolving-AI-Lab/synthesizing)| 154 | |2016|NeurIPS|[Understanding the effective receptive field in deep convolutional neural networks](https://papers.nips.cc/paper/6203-understanding-the-effective-receptive-field-in-deep-convolutional-neural-networks.pdf)|1356| 155 | |2016|CVPR|[Inverting Visual Representations with Convolutional Networks](https://www.cv-foundation.org/openaccess/content_cvpr_2016/papers/Dosovitskiy_Inverting_Visual_Representations_CVPR_2016_paper.pdf)|626| 156 | |2016|CVPR|[Visualizing and Understanding Deep Texture Representations](http://openaccess.thecvf.com/content_cvpr_2016/papers/Lin_Visualizing_and_Understanding_CVPR_2016_paper.pdf)|147| 157 | |2016|CVPR|[Analyzing Classifiers: Fisher Vectors and Deep Neural Networks](https://www.cv-foundation.org/openaccess/content_cvpr_2016/papers/Bach_Analyzing_Classifiers_Fisher_CVPR_2016_paper.pdf)|191| 158 | |2016|ECCV|[Generating Visual Explanations](https://arxiv.org/pdf/1603.08507)|613|[Caffe](https://github.com/LisaAnne/ECCV2016)| 159 | |2016|ECCV|[Design of kernels in convolutional neural networks for image classification](https://arxiv.org/pdf/1511.09231.pdf)|24| 160 | |2016|ICML|[Understanding and improving convolutional neural networks via concatenated rectified linear units](http://www.jmlr.org/proceedings/papers/v48/shang16.pdf)|510| 161 | |2016|ICML|[Visualizing and comparing AlexNet and VGG using deconvolutional layers](https://icmlviz.github.io/icmlviz2016/assets/papers/4.pdf)|126| 162 | |2016|EMNLP|[Rationalizing Neural Predictions](https://arxiv.org/pdf/1606.04155)|738|[Pytorch](https://github.com/zhaopku/Rationale-Torch)| 163 | |2016|IJCV|[Visualizing deep convolutional neural networks using natural pre-images](https://arxiv.org/pdf/1512.02017)|508|[Matlab](https://github.com/aravindhm/nnpreimage)| 164 | |2016|IJCV|[Visualizing Object Detection Features](https://arxiv.org/pdf/1502.05461.pdf)|38|[Caffe](https://github.com/cvondrick/ihog)| 165 | |2016|KDD|[Why should i trust you?: Explaining the predictions of any classifier](https://chu-data-lab.github.io/CS8803Fall2018/CS8803-Fall2018-DML-Papers/lime.pdf)|11742| 166 | |2016|TVCG|[Visualizing the hidden activity of artificial neural networks](https://www.researchgate.net/profile/Samuel_Fadel/publication/306049229_Visualizing_the_Hidden_Activity_of_Artificial_Neural_Networks/links/5b13ffa7aca2723d9980083c/Visualizing-the-Hidden-Activity-of-Artificial-Neural-Networks.pdf)|309| 167 | |2016|TVCG|[Towards better analysis of deep convolutional neural networks](https://arxiv.org/pdf/1604.07043.pdf)|474| 168 | |2016|NAACL|[Visualizing and understanding neural models in nlp](https://arxiv.org/pdf/1506.01066)|650|[Torch](https://github.com/jiweil/Visualizing-and-Understanding-Neural-Models-in-NLP)| 169 | |2016|arxiv|[Understanding neural networks through representation erasure](https://arxiv.org/pdf/1612.08220.pdf))|492| 170 | |2016|arxiv|[Grad-CAM: Why did you say that?](https://arxiv.org/pdf/1611.07450.pdf)|398| 171 | |2016|arxiv|[Investigating the influence of noise and distractors on the interpretation of neural networks](https://arxiv.org/pdf/1611.07270.pdf)|108| 172 | |2016|arxiv|[Attentive Explanations: Justifying Decisions and Pointing to the Evidence](https://arxiv.org/pdf/1612.04757)|88| 173 | |2016|arxiv|[The Mythos of Model Interpretability](http://www.zacklipton.com/media/papers/mythos_model_interpretability_lipton2016.pdf)|3786| 174 | |2016|arxiv|[Multifaceted feature visualization: Uncovering the different types of features learned by each neuron in deep neural networks](https://arxiv.org/pdf/1602.03616)|317| 175 | |2015|ICLR|[Striving for Simplicity: The All Convolutional Net](https://arxiv.org/pdf/1412.6806.pdf)|4645|[Pytorch](https://github.com/StefOe/all-conv-pytorch)| 176 | |2015|CVPR|[Understanding deep image representations by inverting them](https://www.cv-foundation.org/openaccess/content_cvpr_2015/papers/Mahendran_Understanding_Deep_Image_2015_CVPR_paper.pdf)|1942|[Matlab](https://github.com/aravindhm/deep-goggle)| 177 | |2015|ICCV|[Understanding deep features with computer-generated imagery](http://openaccess.thecvf.com/content_iccv_2015/papers/Aubry_Understanding_Deep_Features_ICCV_2015_paper.pdf)|156|[Caffe](https://github.com/mathieuaubry/features_analysis)| 178 | |2015|ICML Workshop|[Understanding Neural Networks Through Deep Visualization](https://arxiv.org/pdf/1506.06579.pdf)|2038|[Tensorflow](https://github.com/jiye-ML/Visualizing-and-Understanding-Convolutional-Networks)| 179 | |2015|AAS|[Interpretable classifiers using rules and Bayesian analysis: Building a better stroke prediction model](https://projecteuclid.org/download/pdfview_1/euclid.aoas/1446488742)|749| 180 | |2014|ECCV|[Visualizing and Understanding Convolutional Networks](https://arxiv.org/pdf/1311.2901.pdf)|18604|[Pytorch](https://github.com/huybery/VisualizingCNN)| 181 | |2014|ICLR|[Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps](https://arxiv.org/pdf/1312.6034.pdf)|6142|[Pytorch](https://github.com/huanghao-code/VisCNN_ICLR_2014_Saliency)| 182 | |2013|ICCV|[Hoggles: Visualizing object detection features](https://www.cv-foundation.org/openaccess/content_iccv_2013/papers/Vondrick_HOGgles_Visualizing_Object_2013_ICCV_paper.pdf)|352| 183 | 184 | + [ ] 论文talk 185 | -------------------------------------------------------------------------------- /sort_cite.md: -------------------------------------------------------------------------------- 1 | |Year|Publication|Paper|Citation| 2 | |:---:|:---:|:---:|:---:| 3 | |2014|ECCV|Visualizing and Understanding Convolutional Networks|8009| 4 | |2016|KDD|Why should i trust you?: Explaining the predictions of any classifier|2255| 5 | |2014|ICLR|Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps|2014| 6 | |2015|ICLR|Striving for Simplicity: The All Convolutional Net|1762| 7 | |2017|ICCV|Grad-cam: Visual explanations from deep networks via gradient-based localization|1333| 8 | |2015|ICMLW|Understanding Neural Networks Through Deep Visualization|974| 9 | |2016|arxiv|The Mythos of Model Interpretability|951| 10 | |2015|CVPR|Understanding deep image representations by inverting them|929| 11 | |2017|NIPS|A Unified Approach to Interpreting Model Predictions|591| 12 | |2017|ICML|Understanding Black-box Predictions via Influence Functions|517| 13 | |2018|DSP|Methods for interpreting and understanding deep neural networks(scihub)|469| 14 | |2017|CVPR|Knowing when to look: Adaptive attention via a visual sentinel for image captioning|458| 15 | |2017|ICML|Axiomatic attribution for deep networks|448| 16 | |2017|CVPR|Making the V in VQA matter: Elevating the role of image understanding in Visual Question Answering|393| 17 | |2017|ICML|Learning Important Features Through Propagating Activation Differences|383| 18 | |2019|AI|Explanation in artificial intelligence: Insights from the social sciences|380| 19 | |2017|CVPR|Network dissection: Quantifying interpretability of deep visual representations|373| 20 | |2019|CSUR|A Survey of Methods for Explaining Black Box Models|344| 21 | |2016|NIPS|Understanding the effective receptive field in deep convolutional neural networks|310| 22 | |2015|AAS|Interpretable classifiers using rules and Bayesian analysis: Building a better stroke prediction model|304| 23 | |2017|ACMSOPP|Deepxplore: Automated whitebox testing of deep learning systems|302| 24 | |2017|ICCV|Interpretable Explanations of Black Boxes by Meaningful Perturbation|284| 25 | |2016|NAACL|Visualizing and understanding neural models in nlp|269| 26 | |2016|CVPR|Inverting Visual Representations with Convolutional Networks|266| 27 | |2018|IJCV|Top-down neural attention by excitation backprop|256| 28 | |2016|NIPS|Synthesizing the preferred inputs for neurons in neural networks via deep generator networks|251| 29 | |2016|EMNLP|Rationalizing Neural Predictions|247| 30 | |2016|ECCV|Generating Visual Explanations|224| 31 | |2016|ICML|Understanding and improving convolutional neural networks via concatenated rectified linear units|216| 32 | |2016|IJCV|Visualizing deep convolutional neural networks using natural pre-images|216| 33 | |2017|ICLR|Visualizing deep neural network decisions: Prediction difference analysis|212| 34 | |2017|arxiv|SmoothGrad: removing noise by adding noise|212| 35 | |2017|arxiv|Explainable artificial intelligence: Understanding, visualizing and interpreting deep learning models|210| 36 | |2018|AAAI|Anchors: High-precision model-agnostic explanations|200| 37 | |2016|TVCG|Towards better analysis of deep convolutional neural networks|184| 38 | |2018|ECCV|Deep clustering for unsupervised learning of visual features|167| 39 | |2018|CVPR|Interpretable Convolutional Neural Networks|154| 40 | |2018|FITEE|Visual interpretability for deep learning: a survey|140| 41 | |2016|arxiv|Understanding neural networks through representation erasure|137| 42 | |2018|Access|Peeking inside the black-box: A survey on Explainable Artificial Intelligence (XAI)|131| 43 | |2016|arxiv|Multifaceted feature visualization: Uncovering the different types of features learned by each neuron in deep neural networks|130| 44 | |2017|arxiv|Distilling a neural network into a soft decision tree|126| 45 | |2018|ICLR|Towards better understanding of gradient-based attribution methods for deep neural networks|123| 46 | |2018|NIPS|Sanity Checks for Saliency Maps|122| 47 | |2016|TVCG|Visualizing the hidden activity of artificial neural networks|122| 48 | |2017|TVCG|ActiVis: Visual Exploration of Industry-Scale Deep Neural Network Models|113| 49 | |2018|AAAI|Improving the adversarial robustness and interpretability of deep neural networks by regularizing their input gradients|112| 50 | |2017|NIPS|Real time image saliency for black box classifiers|111| 51 | |2018|ICML|Interpretability beyond feature attribution: Quantitative testing with concept activation vectors (tcav)|110| 52 | |2017|CVPR|Interpretable 3d human action analysis with temporal convolutional networks|106| 53 | |2017|IJCAI|Right for the right reasons: Training differentiable models by constraining their explanations|102| 54 | |2017|NIPS|SVCCA: Singular Vector Canonical Correlation Analysis for Deep Learning Dynamics and Interpretability|97| 55 | |2019|ExplainAI|The (Un)reliability of saliency methods(scihub)|95| 56 | |2015|ICCV|Understanding deep features with computer-generated imagery|94| 57 | |2018|ICLR|Learning how to explain neural networks: PatternNet and PatternAttribution|90| 58 | |2018|JAIR|Learning Explanatory Rules from Noisy Data|90| 59 | |2016|arxiv|Grad-CAM: Why did you say that?|87| 60 | |2017|CVPR|MDNet: A Semantically and Visually Interpretable Medical Image Diagnosis Network|86| 61 | |2018|WACV|Grad-cam++: Generalized gradient-based visual explanations for deep convolutional networks|85| 62 | |2016|CVPR|Visualizing and Understanding Deep Texture Representations|83| 63 | |2016|CVPR|Analyzing Classifiers: Fisher Vectors and Deep Neural Networks|82| 64 | |2018|ICLR|On the importance of single directions for generalization|81| 65 | |2018|CVPR|Tell me where to look: Guided attention inference network|81| 66 | |2017|ICCV|Interpretable Learning for Self-Driving Cars by Visualizing Causal Attention|80| 67 | |2018|CVPR|Multimodal Explanations: Justifying Decisions and Pointing to the Evidence|78| 68 | |2018|arxiv|Manipulating and measuring model interpretability|73| 69 | |2018|ICML|Learning to explain: An information-theoretic perspective on model interpretation|72| 70 | |2017|arxiv|Challenges for transparency|69| 71 | |2017|arxiv|Interpretable & explorable approximations of black box models|68| 72 | |2018|AAAI|Deep learning for case-based reasoning through prototypes: A neural network that explains its predictions|67| 73 | |2017|EMNLP|A causal framework for explaining the predictions of black-box sequence-to-sequence models|64| 74 | |2017|CEURW|What does explainable AI really mean? A new conceptualization of perspectives|64| 75 | |2019|AAAI|Interpretation of neural networks is fragile|63| 76 | |2019|ACL|Attention is not Explanation|57| 77 | |2018|TPAMI|Interpreting deep visual representations via network dissection|56| 78 | |2017|ACL|Visualizing and Understanding Neural Machine Translation|56| 79 | |2019|NMI|Stop Explaining Black Box Machine Learning Models for High Stakes Decisions and Use Interpretable Models Instead|54| 80 | |2019|ACMFAT|Explaining explanations in AI|54| 81 | |2018|CVPR|Transparency by design: Closing the gap between performance and interpretability in visual reasoning|54| 82 | |2018|AAAI|Interpreting CNN Knowledge via an Explanatory Graph|54| 83 | |2018|MIPRO|Explainable artificial intelligence: A survey|54| 84 | |2019|CVPR|Interpreting CNNs via Decision Trees|49| 85 | |2017|survey|Interpretability of deep learning models: a survey of results|49| 86 | |2018|ICML|Extracting Automata from Recurrent Neural Networks Using Queries and Counterexamples|47| 87 | |2019|CVPR|From Recognition to Cognition: Visual Commonsense Reasoning|44| 88 | |2017|arxiv|Towards interpretable deep neural networks by leveraging adversarial examples|44| 89 | |2017|CVPR|Improving Interpretability of Deep Neural Networks with Semantic Information|43| 90 | |2016|arxiv|Attentive Explanations: Justifying Decisions and Pointing to the Evidence|41| 91 | |2018|ECCV|Explainable neural computation via stack neural module networks|40| 92 | |2018|CVPR|Net2vec: Quantifying and explaining how concepts are encoded by filters in deep neural networks|39| 93 | |2017|ICCV|Understanding and comparing deep neural networks for age and gender classification|39| 94 | |2018|ECCV|Grounding visual explanations|38| 95 | |2019|NIPS|This looks like that: deep learning for interpretable image recognition|35| 96 | |2018|NIPS|Explanations based on the missing: Towards contrastive explanations with pertinent negatives|35| 97 | |2017|IJCAI|Understanding and improving convolutional neural networks via concatenated rectified linear units|35| 98 | |2018|ACL|Did the Model Understand the Question?|34| 99 | |2018|ICLR|Detecting statistical interactions from neural network weights|30| 100 | |2018|ECCV|Textual explanations for self-driving vehicles|30| 101 | |2018|BMVC|Rise: Randomized input sampling for explanation of black-box models|30| 102 | |2017|arxiv|Contextual Explanation Networks|28| 103 | |2016|ICML|Visualizing and comparing AlexNet and VGG using deconvolutional layers|28| 104 | |2018|NIPS|Towards robust interpretability with self-explaining neural networks|27| 105 | |2018|AIES|Detecting Bias in Black-Box Models Using Transparent Model Distillation|27| 106 | |2018|arxiv|How convolutional neural network see the world-A survey of convolutional neural network visualization methods|27| 107 | |2018|ECCV|Interpretable basis decomposition for visual explanation|26| 108 | |2018|NIPS|Attacks meet interpretability: Attribute-steered detection of adversarial samples|26| 109 | |2017|ICLR|Exploring LOTS in Deep Neural Networks|26| 110 | |2017|AAAI|Growing Interpretable Part Graphs on ConvNets via Multi-Shot Learning|26| 111 | |2018|arxiv|Revisiting the importance of individual units in cnns via ablation|25| 112 | |2018|AAAI|Examining CNN Representations with respect to Dataset Bias|24| 113 | |2016|arxiv|Investigating the influence of noise and distractors on the interpretation of neural networks|24| 114 | |2016|IJCV|Visualizing Object Detection Features|22| 115 | |2018|ICLR|Interpretable counting for visual question answering|21| 116 | |2018|CVPR|What have we learned from deep representations for action recognition?|20| 117 | |2018|CVPR|Learning to Act Properly: Predicting and Explaining Affordances from Images|17| 118 | |2018|ECCV|Convnets and imagenet beyond accuracy: Understanding mistakes and uncovering biases|17| 119 | |2018|NIPS Workshop|Interpretable Convolutional Filters with SincNet|17| 120 | |2019|JVCIR|Interpretable convolutional neural networks via feedforward design|16| 121 | |2019|ICLR|Hierarchical interpretations for neural network predictions|15| 122 | |2018|NIPS|DeepPINK: reproducible feature selection in deep neural networks|15| 123 | |2017|CVPR|Mining Object Parts from CNNs via Active Question-Answering|15| 124 | |2019|CVPR|Attention branch network: Learning of attention mechanism for visual explanation|14| 125 | |2017|CVPRW|Looking under the hood: Deep neural network visualization to interpret whole-slide image analysis outcomes for colorectal polyps|14| 126 | |2018|CVPR|Teaching Categories to Human Learners with Visual Explanations|13| 127 | |2018|ECCV|Vqa-e: Explaining, elaborating, and enhancing your answers for visual questions|12| 128 | |2018|NIPS|Representer point selection for explaining deep neural networks|11| 129 | |2016|ECCV|Design of kernels in convolutional neural networks for image classification|11| 130 | |2019|ICLR|How Important Is a Neuron?|10| 131 | |2017|ICCV|Learning to disambiguate by asking discriminative questions|10| 132 | |2019|AAAIW|Unsupervised Learning of Neural Networks to Explain Neural Networks|9| 133 | |2018|CVPR|What do Deep Networks Like to See?|9| 134 | |2019|CVPR|Interpretable and fine-grained visual explanations for convolutional neural networks|8| 135 | |2018|ECCV|Choose Your Neuron: Incorporating Domain Knowledge through Neuron-Importance|8| 136 | |2019|ICLR|Visual Explanation by Interpretation: Improving Visual Feedback Capabilities of Deep Neural Networks|7| 137 | |2019|ICAIS|Interpreting black box predictions using fisher kernels|7| 138 | |2019|CVPR|Learning to Explain with Complemental Examples|6| 139 | |2019|ICCV|U-CAM: Visual Explanation using Uncertainty based Class Activation Maps|6| 140 | |2019|ICCV|Towards Interpretable Face Recognition|6| 141 | |2019|CVPR|Revealing Scenes by Inverting Structure from Motion Reconstructions|5| 142 | |2019|ICCV|Taking a HINT: Leveraging Explanations to Make Vision and Language Models More Grounded|5| 143 | |2018|CVPR|Interpret Neural Networks by Identifying Critical Data Routing Paths|5| 144 | |2018|ECCV|Diverse feature visualizations reveal invariances in early layers of deep neural networks|5| 145 | |2019|ICML|Towards A Deep and Unified Understanding of Deep Neural Models in NLP|4| 146 | |2019|AAAI|Classifier-agnostic saliency map extraction|4| 147 | |2019|AAAIW|Network Transplanting|4| 148 | |2019|arxiv|Attention Interpretability Across NLP Tasks|4| 149 | |2019|NIPS|A benchmark for interpretability methods in deep neural networks（同arxiv:1806.10758）|3| 150 | |2019|arxiv|Interpretable CNNs|3| 151 | |2019|NIPS|Full-gradient representation for neural network visualization|2| 152 | |2019|NIPS|On the (In) fidelity and Sensitivity of Explanations|2| 153 | |2019|ICCV|Understanding Deep Networks via Extremal Perturbations and Smooth Masks|2| 154 | |2019|NIPS|Towards Automatic Concept-based Explanations|1| 155 | |2019|NIPS|CXPlain: Causal explanations for model interpretation under uncertainty|1| 156 | |2019|CVPR|Multimodal Explanations by Predicting Counterfactuality in Videos|1| 157 | |2019|CVPR|Visualizing the Resilience of Deep Convolutional Network Interpretations|1| 158 | |2019|ICCV|Explaining Neural Networks Semantically and Quantitatively|1| 159 | |2018|arxiv|Computationally Efficient Measures of Internal Neuron Importance|1| 160 | |2020|ICLR|Knowledge Isomorphism between Neural Networks|0| 161 | |2020|ICLR|Interpretable Complex-Valued Neural Networks for Privacy Protection|0| 162 | |2019|AAAI|Can You Explain That? Lucid Explanations Help Human-AI Collaborative Image Retrieval|0| 163 | |2018|ECCV|ExplainGAN: Model Explanation via Decision Boundary Crossing Transformations|0| 164 | --------------------------------------------------------------------------------