DNC-SemesterProject

Semester Project: exploring Differentiable Neural Computers and applications

Authors: Marco Angius and Giorgio Avalle

The goals of the project are:

• explore the architecture of DNCs
• define a base code for running simulations
• possible create a pytorch version of the code
• try to implement the graph tasks
• map the graph task to an application

The project is based on the article of DeepMind and the code available here.

Running Notes

For the networkx library use the 1.8.1 version.