Cumulative inhibition in neural networks

Trond Arild Tjøstheim; Christian Balkenius

doi:10.1007/s10339-018-0888-z

Cumulative inhibition in neural networks

Trond Arild Tjøstheim, Christian Balkenius

Forskningsoutput: Tidskriftsbidrag › Artikel i vetenskaplig tidskrift › Peer review

4 Citeringar (SciVal)

Sammanfattning

We show how a multi-resolution network can model the development of acuity and coarse-to-fine processing in the mammalian visual cortex. The network adapts to input statistics in an unsupervised manner, and learns a coarse-to-fine representation by using cumulative inhibition of nodes within a network layer. We show that a system of such layers can represent input by hierarchically composing larger parts from smaller components. It can also model aspects of top-down processes, such as image regeneration.

Originalspråk	engelska
Sidor (från-till)	87-102
Tidskrift	Cognitive Processing
Volym	20
Utgåva	1
Tidigt onlinedatum	2018 nov. 3
DOI	https://doi.org/10.1007/s10339-018-0888-z
Status	Published - 2019

Ämnesklassifikation (UKÄ)

Jämförande språkvetenskap och lingvistik

Tillgång till dokument

10.1007/s10339-018-0888-z

1 Avslutade
1 Aktiva

CogPhi: Cognitive Philosophy Research Group (CogPhi)
Stephens, A., Tjöstheim, T. A., Roszko, M., Olsson, E. J., Anikin, A. & Arthur Schwaninger, T. P. U. O. Z.
2018/11/01 → …
Projekt: Nätverk
eSSENCE@LU 4:1 - Method development for analysis and modelling of large scale electrophysiological recordings using deep artificial neural networks
Balkenius, C., Petersson, P., Åström, K., Tjöstheim, T. A., Johansson, B. & Sjöbom, J.
2017/07/01 → 2020/06/30
Projekt: Forskning

Lund University Cognitive Robotics Lab
Birger Johansson (Manager) & Christian Balkenius (Manager)
Kognitionsvetenskap
Infrastruktur

Citera det här

@article{93c5ca41ce5b4ea78a8ed82533013fc1,

title = "Cumulative inhibition in neural networks",

abstract = "We show how a multi-resolution network can model the development of acuity and coarse-to-fine processing in the mammalian visual cortex. The network adapts to input statistics in an unsupervised manner, and learns a coarse-to-fine representation by using cumulative inhibition of nodes within a network layer. We show that a system of such layers can represent input by hierarchically composing larger parts from smaller components. It can also model aspects of top-down processes, such as image regeneration.",

keywords = "Cumulative inhibition, Multi-resolution, Coarse-to-fine processing, Unsupervised learning, Acuity, Cortical microcolumn, Visual cortex",

author = "Tj{\o}stheim, {Trond Arild} and Christian Balkenius",

year = "2019",

doi = "10.1007/s10339-018-0888-z",

language = "English",

volume = "20",

pages = "87--102",

journal = "Cognitive Processing",

issn = "1612-4782",

publisher = "Springer",

number = "1",

}

TY - JOUR

T1 - Cumulative inhibition in neural networks

AU - Tjøstheim, Trond Arild

AU - Balkenius, Christian

PY - 2019

Y1 - 2019

N2 - We show how a multi-resolution network can model the development of acuity and coarse-to-fine processing in the mammalian visual cortex. The network adapts to input statistics in an unsupervised manner, and learns a coarse-to-fine representation by using cumulative inhibition of nodes within a network layer. We show that a system of such layers can represent input by hierarchically composing larger parts from smaller components. It can also model aspects of top-down processes, such as image regeneration.

AB - We show how a multi-resolution network can model the development of acuity and coarse-to-fine processing in the mammalian visual cortex. The network adapts to input statistics in an unsupervised manner, and learns a coarse-to-fine representation by using cumulative inhibition of nodes within a network layer. We show that a system of such layers can represent input by hierarchically composing larger parts from smaller components. It can also model aspects of top-down processes, such as image regeneration.

KW - Cumulative inhibition

KW - Multi-resolution

KW - Coarse-to-fine processing

KW - Unsupervised learning

KW - Acuity

KW - Cortical microcolumn

KW - Visual cortex

U2 - 10.1007/s10339-018-0888-z

DO - 10.1007/s10339-018-0888-z

M3 - Article

C2 - 30392141

VL - 20

SP - 87

EP - 102

JO - Cognitive Processing

JF - Cognitive Processing

SN - 1612-4782

IS - 1

ER -

Cumulative inhibition in neural networks

Sammanfattning

Ämnesklassifikation (UKÄ)

Tillgång till dokument

Fingeravtryck

Projekt

CogPhi: Cognitive Philosophy Research Group (CogPhi)

eSSENCE@LU 4:1 - Method development for analysis and modelling of large scale electrophysiological recordings using deep artificial neural networks

Utrustning

Lund University Cognitive Robotics Lab

Citera det här