Reproducible Research PrimerΒΆ

@author: Vincent Marois, Tomasz Kornuta

One of the main reasons for the existence of this framework is to enable reproducible research, i.e. to enable other researchers to run experiments designed by us and reproduce our results.

This section gathers information about our published work and describes how to reproduce the associated experiments.


We have tested the compositional generalization of the MAC and S-MAC models using CLEVR and CoGenT datasets. Our work [1] has been accepted and presented during the ViGIL Workshop @ NeurIPS 2018.

See Compositional generalization of the MAC and S-MAC models

[1] Marois, Jayram, Albouy, Kornuta, Bouhadjar, Ozcan (2018). On transfer learning using a MAC model variant. arxiv/1811.06529

Preliminary results of experiments with COG dataset.

See Testing working memory using Cog dataset and model