[Re] Neural Network Model of Memory Retrieval

ReScience C, 2021

Carlos de la Torre-Ortiz and Aurélien Nioche

We replicate a connectionist model of associative memory and correct the original manuscript


  • Recanatesi et al. present a model of memory retrieval based on a Hopfield model for associative learning, with network dynamics that reflect associations between items due to semantic similarities.


  • In this work, we proceed to replicate the model as presented by Recanatesi et al. During our replication efforts, we discover several errors in parameters and collaborate with the original work authors to provide a successful replication and correct the original article.


Publication

Carlos de la Torre-Ortiz and Aurélien Nioche

[Re] Neural Network Model of Memory Retrieval

ReScience C

@article{delatorreortiz2021reneural,
author = {de la Torre-Ortiz, Carlos and Nioche, Aurélien},
title = {{[Re] Neural Network Model of Memory Retrieval}},
journal = {ReScience C},
year = {2021},
month = jan,
volume = {6},
number = {3},
pages = {{#8}},
doi = {10.5281/zenodo.4461767},
url = {https://zenodo.org/record/4461767/files/article.pdf},
code_url = {https://github.com/c-torre/replication-recanatesi-2015},
code_doi = {},
code_swh = {swh:1:dir:0b541aeb4707ecedfbbcdd85adfd0100d748cc03},
data_url = {},
data_doi = {},
review_url = {https://github.com/ReScience/submissions/issues/10},
type = {Replication},
language = {Python},
domain = {Computational Neuroscience},
keywords = {attractor neural networks, recall, oscillations, memory, neural representations, python}
}