IMNN

• Tool name , objectives, feature: IMNN (Information Maximization Neural Network)

The IMNN is a python library based on tensorflow that aim at building neural network that compute data summaries that are optimized for a given extraction of information on some pre-defined parameters. The basic example of such summary is the variance of a Gaussian distribution.

• contact (person witin ADE, ie. that can help, not necessarily author): Guilhem Lavaux
• author(s): Thomas Charnock, Benjamin Wandelt, Guilhem Lavaux.
• publication(s), refs:

Charnock, Lavaux & Wandelt (2019)

• main url:

https://github.com/tomcharnock/IMNN

• documentation: https://github.com/tomcharnock/IMNN
• type: Library
• language: Python/Tensorflow
• parallelism: GPU
• ressources required: Workstation / Mesocenter
• availability:
• last update: may 22nd, 2020