Enregistré dans:
Détails bibliographiques
Auteurs principaux: Nouri, Mahbod, Rotermund, David, Garcia-Ortiz, Alberto, Pawelzik, Klaus R.
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2503.20398
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866909554040635392
author Nouri, Mahbod
Rotermund, David
Garcia-Ortiz, Alberto
Pawelzik, Klaus R.
author_facet Nouri, Mahbod
Rotermund, David
Garcia-Ortiz, Alberto
Pawelzik, Klaus R.
contents The brain uses positive signals as a means of signaling. Forward interactions in the early visual cortex are also positive, realized by excitatory synapses. Only local interactions also include inhibition. Non-negative matrix factorization (NMF) captures the biological constraint of positive long-range interactions and can be implemented with stochastic spikes. While NMF can serve as an abstract formalization of early neural processing in the visual system, the performance of deep convolutional networks with NMF modules does not match that of CNNs of similar size. However, when the local NMF modules are each followed by a module that mixes the NMF's positive activities, the performances on the benchmark data exceed that of vanilla deep convolutional networks of similar size. This setting can be considered a biologically more plausible emulation of the processing in cortical (hyper-)columns with the potential to improve the performance of deep networks.
format Preprint
id arxiv_https___arxiv_org_abs_2503_20398
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Including local feature interactions in deep non-negative matrix factorization networks improves performance
Nouri, Mahbod
Rotermund, David
Garcia-Ortiz, Alberto
Pawelzik, Klaus R.
Machine Learning
Artificial Intelligence
The brain uses positive signals as a means of signaling. Forward interactions in the early visual cortex are also positive, realized by excitatory synapses. Only local interactions also include inhibition. Non-negative matrix factorization (NMF) captures the biological constraint of positive long-range interactions and can be implemented with stochastic spikes. While NMF can serve as an abstract formalization of early neural processing in the visual system, the performance of deep convolutional networks with NMF modules does not match that of CNNs of similar size. However, when the local NMF modules are each followed by a module that mixes the NMF's positive activities, the performances on the benchmark data exceed that of vanilla deep convolutional networks of similar size. This setting can be considered a biologically more plausible emulation of the processing in cortical (hyper-)columns with the potential to improve the performance of deep networks.
title Including local feature interactions in deep non-negative matrix factorization networks improves performance
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2503.20398