An adaptive locally connected neuron model: Focusing neuron
Yükleniyor...
Dosyalar
Tarih
2021-01-02
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Elsevier B.V.
Erişim Hakkı
info:eu-repo/semantics/closedAccess
Özet
This paper presents a new artificial neuron model capable of learning its receptive field in the topological domain of inputs. The experiments include tests of focusing neuron networks of one or two hidden layers on synthetic and well-known image recognition data sets. The results demonstrated that the focusing neurons can move their receptive fields towards more informative inputs. In the simple two-hidden layer networks, the focusing layers outperformed the dense layers in the classification of the 2D spatial data sets. Moreover, the focusing networks performed better than the dense networks even when 70% of the weights were pruned. The tests on convolutional networks revealed that using focusing layers instead of dense layers for the classification of convolutional features may work better in some data sets.
Açıklama
Anahtar Kelimeler
Adaptive locally connected neuron, Adaptive receptive field, Attention, Focusing neuron, Pruning, Convolution, Convolutional neural networks, Focusing, Image recognition, Network layers, Neurons, Topology, Artificial neurons, Convolutional networks, Dense network, Hidden layer networks, Hidden layers, Neuron networks, Receptive fields, Spatial datasets, Classification (of information), Article, Attention, Learning, Nerve cell, Receptive field
Kaynak
Neurocomputing
WoS Q Değeri
Q2
Scopus Q Değeri
Q1
Cilt
419
Sayı
Künye
Tek, F. B. (2021). An adaptive locally connected neuron model: Focusing neuron. Neurocomputing, 419, 306-321. doi:10.1016/j.neucom.2020.08.008
