Deep Networks Visualisation and Interpretation

  1. [ ] The (Un)reliability of saliency methods \cite{kindermans2017reliability}
  2. [ ] DeepXplore: Automated Whitebox Testing of Deep Learning Systems \cite{pei2017deepxplore}
  3. [ ] Feature Visualisation \cite{olah2017feature}  (new blog from google with fancy images)
  4. [ ] Visualizing the loss landscape of neural nets\cite{li2017visualizing}
-----------------------------------------------------------------------------------------------------------------------------------------------------

Notes

  1. [Bernhard] An online book about interpretable machine learning - mentioned on 07/12/2017: https://christophm.github.io/interpretable-ml-book/

-----------------------------------------------------------------------------------------------------------------------------------------------------

Previous Papers

2018

  1. [x] Deep Adversarial Networks for Biomedical Image Segmentation Utilizing Unannotated Images \cite{zhang2017deep}

2017

  1. [x] CNN Fixations: An unraveling approach to visualize the discriminative image regions \cite{Mopuri2017}
  2. [x] Visualizing Deep Neural Network Decisions: Prediction Difference Analysis \cite{Zintgraf2017}
  3. [x] Soft Proposal Networks for Weakly Supervised Object Localization \cite{Zhu2017}
  4. [x] Self-supervised Learning for Spinal MRIs \cite{Jamaludin_2017}
  5. [x] Visual Feature Attribution using Wasserstein GANs \cite{1711.08998}
  6. [x] Non-local Neural Networks \cite{1711.07971} - from facebookAI - Achieves best results in video classification, object segmentation and pose estimation - Check Reddit discussion
  7. [x] Distilling a Neural Network Into a Soft Decision Tree \cite{1711.09784} - describes a way of using a trained neural net to create a type of soft decision tree that generalizes better than one learned directly from the training data - Reddit

-----------------------------------------------------------------------------------------------------------------------------------------------------