deepsense.io webinar

Unblackboxing the black box webinar

This is what it is all about

• Class Activation Maps

• Saliency Maps

• Maximal Activation Image

• Quiver

• TSNE

• Neptune

• Tensorboard

• LIME

• Attention

Prerequisites

Install the requirements

pip install -r requirements.txt

Download Twitter Sentiment Dataset from http://thinknook.com/wp-content/uploads/2012/09/Sentiment-Analysis-Dataset.zip

or via

wget http://thinknook.com/wp-content/uploads/2012/09/Sentiment-Analysis-Dataset.zip

Download the pretrained glove model from https://nlp.stanford.edu/projects/glove/ and unpack it to some folder.

Specify the filepaths in read_between_the_tweets.py and train your twitter model by running:

neptune run read_between_the_tweets.py --config neptune_tweet.yaml

Specify the filepaths in see_my_face.py and train your facial recognition model by running:

neptune run see_my_face.py --config neptune_facereco.yaml

Change the filepaths in notebooks to the same ones you specified in read_between_the_tweets.py and see_my_face.py .

Resources and Acknowledgements

Thank you all for letting the world use this!

Keras.io https://keras.io/

Keras-vis https://raghakot.github.io/keras-vis/

Neptune http://neptune.deepsense.io/versions/latest/

Eli5 http://eli5.readthedocs.io/en/latest/overview.html

LIME https://homes.cs.washington.edu/~marcotcr/blog/lime/

Quiver https://github.com/keplr-io/quiver

Tensorflow https://www.tensorflow.org/

https://jacobgil.github.io/deeplearning/class-activation-maps

https://arxiv.org/abs/1312.6034

http://p.migdal.pl/2017/01/06/king-man-woman-queen-why.html

https://nlp.stanford.edu/projects/glove/

http://blog.echen.me/2017/05/30/exploring-lstms/

https://github.com/philipperemy/keras-attention-mechanism

https://medium.com/@plusepsilon/visualizations-of-recurrent-neural-networks-c18f07779d56

https://richliao.github.io/supervised/classification/2016/12/26/textclassifier-RNN/

https://arxiv.org/pdf/1409.0473.pdf

https://gist.github.com/wassname/5292f95000e409e239b9dc973295327a

https://gist.github.com/nigeljyng/37552fb4869a5e81338f82b338a304d3

This is us