Benjamin Ellenberger

icon company IDSC.io icon location Bern, Switzerland icon location Data Scientist @ IDSC Insel University Hospital, PhD in the Comp. Neuro. Group of Walter Senn at Uni Bern, personally on Google @bulletphysics @brax and Deep-RL

Results 146 comments of Benjamin Ellenberger

Make it easy to check out and compile

It could at least help to write a wiki article on everything that is needed and how you set it up.

Implement epochs to have different evaluation modes

evaluation modes should be evolution/learning track.

Calculate the Transfer Entropy of the Neural Network

And show it in the neural graph if possible

Add the feature cutting operator

Add the feature cutting mode/operator (Togglable: Feature must be expressed)

Look at M5 partitioning algorithm

Also look at the algorithm that was used in the Deepmind paper to flatten the space to 2D.

LSTM networks

A huge bunch of deep learning links: - http://www.slideshare.net/mobile/jmugan/deep-learningforartificialintelligence - http://memkite.com/deep-learning-bibliography/ - https://m.youtube.com/watch?v=jna8tAd87yw - http://nikhilbuduma.com/2015/01/11/a-deep-dive-into-recurrent-neural-networks/ - http://people.idsia.ch/~juergen/rnnbook.html - http://journal.frontiersin.org/article/10.3389/fncom.2015.00032/abstract - https://courses.cs.ut.ee/2015/neuro/spring/Main/Lectures - https://courses.cs.ut.ee/2015/neuro/spring/Main/Links - http://www.nature.com/nature/journal/v518/n7540/full/nature14236.html - http://neuralnetworksanddeeplearning.com/