We organise a Research Seminar on Neural Information Processing Systems. If you wish to take part then please register for the seminar and attend the first meeting in Sp.A.01.08 at 8.45 on February 7th 2017.

The goal of this seminar is to investigate how human-like intelligence can be implemented in synthetic brains. You will work on group or individual projects to implement brain-inspired computing machinery to solve problems related to this topic. Projects can be theoretical or applied in nature. The goal is to work from the inception of an idea towards a publishable research paper which addresses your research question.

The research seminar will take one semester. You will meet once a week with the instructor(s) and your fellow students to discuss background literature and ongoing projects. During the first weeks you will orient yourself on a possible project. Next, you will work individually or as a group on this project. Students are expected to communicate regularly and share know-how.

The final product will be a working and well-documented implementation which solves an outstanding problem in computational neuroscience or artificial intelligence as well as a NIPS style paper (see nips.cc) written under supervision of the coordinator, who will provide feedback during the process.

Your grade will depend on active participation and sophistication of your work. Your final product (paper + source code) will be evaluated in terms of novelty, quality and scientific rigour. Soft criteria are the ability to communicate with the instructor(s) and peers, problem-solving abilities, and academic attitude.

Background material

MOOCS

https://www.udacity.com/course/deep-learning–ud730
http://cs231n.stanford.edu

Deep learning frameworks
http://docs.chainer.org/en/stable/tutorial/basic.html
https://www.tensorflow.org
http://keras.io
http://deeplearning.net/software/theano/
http://www.vlfeat.org/matconvnet/
http://caffe.berkeleyvision.org

Useful Python packages
Pandas
Beautiful Soup
NLTK
Gensim
Scikit-learn
Pillow
Librosa
Seaborn

Reading material

http://www.deeplearningbook.org
http://deeplearning.net
http://neuralnetworksanddeeplearning.com
http://colah.github.io
http://karpathy.github.io

Conferences
NIPS
ICLR
Cosyne

Other useful links
https://www.microsoft.com/cognitive-services/en-us/apis
http://dlib.net
Kaggle