AlexNet

AlexNet (Aliknit) nisqaqa huk k'uyukuq ankucha llika (CNN) nisqap sutinmi, Alex Krizhevsky sutiyuq runap rurasqan, Ilya Sutskeverwan Geoffrey Hintonwan kuska rurasqa, payqa Krizhevskyp Ph.D. yuyaychaq.^[1]^[2]

AlexNet 30 ñiqin tarpuy killapi 2012 watapi ImageNet hatun rikuy riqsiy sasachakuypi atipanakurqan ^[3] Llikaqa 15,3% pantayta 5 puntapi ayparqan, aswan 10,8 pachakmanta aswan pisi subcampeón nisqamanta. Ñawpaq qillqasqapa ñawpaq ruwayninqa karqan, modelop ukhunqa ancha allinmi karqan hatun ruwayninpaq, chaymi computacionalmente ancha chaninniyuq karqan, ichaqa ruway atikuqmi karqan qillqa thatkichiy hukkay (GPU) nisqakuna yachachiypi llamk'achisqankurayku.^[2]

Ñawpaq kawsay kananpi

AlexNet mana ñawpaq utqaylla GPU-rurachiychu karqan huk CNN kaqmanta huk siq'i riqsiy atipanakuypi atipananpaq. Huk CNN GPU nisqamanta K. Chellapilla et al. (2006) 4 kuti aswan utqaylla karqa huk kaqlla implementacionmanta CPU kaqpi.^[4] Huk ukhu CNN Dan Cireșan et al. (2011) IDSIA kaqpi 60 kuti aswan utqayllaña karqa ^[5] chaymanta ñawpaq kaqkunamanta aswan allinta ruwarqa agosto killapi 2011 watapi ^[6] 15 ñiqin qhulla puquy killapi 2011 watapi 10 ñiqin tarpuy killapi 2012 watapikama, paykunap CNN nisqa tawa siq'i atipanakuykunapi mana pisichu atiparqan.^[7] ^[8] Hinallataqmi achka siq'ikuna waqaychana wasikunapaq qillqakunapi aswan allin ruwaypipas anchata allincharqaku.^[9]

Llika ruway

AlexNet nisqapiqa pusaq qatakunam karqa; ñawpaq pichqaqa capas k'uyukuqkuna nisqakunam karqa, wakinqa max-pooling nisqa capas nisqakunam qatipasqa, kimsa qipa kaqtaq capas completamente conectadas nisqa. Llikaqa, qhipa qatamanta aswan, iskay copiaman rakisqa, sapa juk GPU kaqpi purichisqa. ^[2] Tukuy estructurata kayhinata qillqachwan $(CNN\to RN\to MP)^{2}\to (CNN^{3}\to MP)\to (FC\to DO)^{2}\to Linear\to softmax$ maypi

CNN = capa convolucional (ReLU activación nisqawan) .
RN = normalización de respuesta local nisqa
MP = maxpooling nisqa
FC = hunt'asqa tinkisqa qata (ReLU llamk'achiyninwan) .
Lineal = hunt'asqa tinkisqa qata (mana llamk'achisqa) .
DO = saqisqa yachay

Kayqa mana saturador ReLU activación ruwayta llamk’achirqa, kaytaq rikuchirqa allinchasqa entrenamiento ruwayta tanh chanta sigmoide kaqmanta. ^[2]

Influencia

AlexNet nisqaqa computadora rikuypi aswan atiyniyuq qillqakunamanta hukninmi, aswan achka qillqasqakunata CNN nisqakunata, GPU nisqakunata llamk'achispa ukhu yachayta utqaylla purichinapaq kallpachasqanrayku.^[10] 2022 wata tukukuykama, AlexNet qillqasqaqa 100.000 masnin kutitam citasqa karqan Google Scholar nisqanman hina.

Pukyukuna

<div class="reflist {{#if:

   | columns {{#iferror: 
     | references-column-width

↑ Gershgorn, Dave (26 July 2017). The data that transformed AI research—and possibly the world
↑ ^2,0 ^2,1 ^2,2 ^2,3 Krizhevsky, Alex (2017-05-24). ImageNet classification with deep convolutional neural networks. https://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf. Pukyumanta willaypi pantasqa: Etiqueta <ref> no válida; el nombre «:0» está definido varias veces con contenidos diferentes
↑ ImageNet Large Scale Visual Recognition Competition 2012 (ILSVRC2012)
↑ Kumar Chellapilla; Sid Puri; Patrice Simard (2006). "High Performance Convolutional Neural Networks for Document Processing". In Lorette, Guy (ed.). Tenth International Workshop on Frontiers in Handwriting Recognition. Suvisoft.
↑ Flexible, High Performance Convolutional Neural Networks for Image Classification. http://www.idsia.ch/~juergen/ijcai2011.pdf. Retrieved 17 November 2013.
↑ IJCNN 2011 Competition result table (en-US) (2010)
↑ History of computer vision contests won by deep CNNs on GPU (en-US) (17 March 2017)
↑ Deep Learning. 2015. http://www.scholarpedia.org/article/Deep_Learning.
↑ Cireșan, Dan; Meier, Ueli; Schmidhuber, Jürgen (June 2012). Multi-column deep neural networks for image classification. 2012 IEEE Conference on Computer Vision and Pattern Recognition. New York, NY: Institute of Electrical and Electronics Engineers (IEEE). pp. 3642–3649. arXiv:1202.2745. CiteSeerX 10.1.1.300.3283. doi:10.1109/CVPR.2012.6248110. ISBN 978-1-4673-1226-4. OCLC 812295155. S2CID 2161592.
↑ Deshpande, Adit. The 9 Deep Learning Papers You Need To Know About (Understanding CNNs Part 3)

[:1-1] Gershgorn, Dave (26 July 2017). The data that transformed AI research—and possibly the world

[:0-2] 2,0 ^2,1 ^2,2 ^2,3 Krizhevsky, Alex (2017-05-24). ImageNet classification with deep convolutional neural networks. https://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf. Pukyumanta willaypi pantasqa: Etiqueta <ref> no válida; el nombre «:0» está definido varias veces con contenidos diferentes

[:2-3] ImageNet Large Scale Visual Recognition Competition 2012 (ILSVRC2012)

[4] Kumar Chellapilla; Sid Puri; Patrice Simard (2006). "High Performance Convolutional Neural Networks for Document Processing". In Lorette, Guy (ed.). Tenth International Workshop on Frontiers in Handwriting Recognition. Suvisoft.

[flexible-5] Flexible, High Performance Convolutional Neural Networks for Image Classification. http://www.idsia.ch/~juergen/ijcai2011.pdf. Retrieved 17 November 2013.

[6] IJCNN 2011 Competition result table (en-US) (2010)

[7] History of computer vision contests won by deep CNNs on GPU (en-US) (17 March 2017)

[schdeepscholar-8] Deep Learning. 2015. http://www.scholarpedia.org/article/Deep_Learning.

[mcdns-9] Cireșan, Dan; Meier, Ueli; Schmidhuber, Jürgen (June 2012). Multi-column deep neural networks for image classification. 2012 IEEE Conference on Computer Vision and Pattern Recognition. New York, NY: Institute of Electrical and Electronics Engineers (IEEE). pp. 3642–3649. arXiv:1202.2745. CiteSeerX 10.1.1.300.3283. doi:10.1109/CVPR.2012.6248110. ISBN 978-1-4673-1226-4. OCLC 812295155. S2CID 2161592.

[10] Deshpande, Adit. The 9 Deep Learning Papers You Need To Know About (Understanding CNNs Part 3)

[1]

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]