Enhancing the accuracies by performing pooling decisions adjacent to the output layer
Abstract Learning classification tasks of $$({2}^{n}\times {2}^{n})$$ ( 2 n × 2 n ) inputs typically consist of $$\le n(2\times 2$$ ≤ n ( 2 × 2 ) max-pooling (MP) operators along the entire feedforward deep architecture. Here we show, using the CIFAR-10 database, that pooling decisions adjacent to t...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Nature Portfolio
2023-08-01
|
Series: | Scientific Reports |
Online Access: | https://doi.org/10.1038/s41598-023-40566-y |
_version_ | 1797452965199478784 |
---|---|
author | Yuval Meir Yarden Tzach Ronit D. Gross Ofek Tevet Roni Vardi Ido Kanter |
author_facet | Yuval Meir Yarden Tzach Ronit D. Gross Ofek Tevet Roni Vardi Ido Kanter |
author_sort | Yuval Meir |
collection | DOAJ |
description | Abstract Learning classification tasks of $$({2}^{n}\times {2}^{n})$$ ( 2 n × 2 n ) inputs typically consist of $$\le n(2\times 2$$ ≤ n ( 2 × 2 ) max-pooling (MP) operators along the entire feedforward deep architecture. Here we show, using the CIFAR-10 database, that pooling decisions adjacent to the last convolutional layer significantly enhance accuracies. In particular, average accuracies of the advanced-VGG with $$m$$ m layers (A-VGGm) architectures are 0.936, 0.940, 0.954, 0.955, and 0.955 for m = 6, 8, 14, 13, and 16, respectively. The results indicate A-VGG8’s accuracy is superior to VGG16’s, and that the accuracies of A-VGG13 and A-VGG16 are equal, and comparable to that of Wide-ResNet16. In addition, replacing the three fully connected (FC) layers with one FC layer, A-VGG6 and A-VGG14, or with several linear activation FC layers, yielded similar accuracies. These significantly enhanced accuracies stem from training the most influential input–output routes, in comparison to the inferior routes selected following multiple MP decisions along the deep architecture. In addition, accuracies are sensitive to the order of the non-commutative MP and average pooling operators adjacent to the output layer, varying the number and location of training routes. The results call for the reexamination of previously proposed deep architectures and their accuracies by utilizing the proposed pooling strategy adjacent to the output layer. |
first_indexed | 2024-03-09T15:16:36Z |
format | Article |
id | doaj.art-71571b4bdb5b459399ffc7f088d13ea4 |
institution | Directory Open Access Journal |
issn | 2045-2322 |
language | English |
last_indexed | 2024-03-09T15:16:36Z |
publishDate | 2023-08-01 |
publisher | Nature Portfolio |
record_format | Article |
series | Scientific Reports |
spelling | doaj.art-71571b4bdb5b459399ffc7f088d13ea42023-11-26T13:04:43ZengNature PortfolioScientific Reports2045-23222023-08-011311810.1038/s41598-023-40566-yEnhancing the accuracies by performing pooling decisions adjacent to the output layerYuval Meir0Yarden Tzach1Ronit D. Gross2Ofek Tevet3Roni Vardi4Ido Kanter5Department of Physics, Bar-Ilan UniversityDepartment of Physics, Bar-Ilan UniversityDepartment of Physics, Bar-Ilan UniversityDepartment of Physics, Bar-Ilan UniversityGonda Interdisciplinary Brain Research Center, Bar-Ilan UniversityDepartment of Physics, Bar-Ilan UniversityAbstract Learning classification tasks of $$({2}^{n}\times {2}^{n})$$ ( 2 n × 2 n ) inputs typically consist of $$\le n(2\times 2$$ ≤ n ( 2 × 2 ) max-pooling (MP) operators along the entire feedforward deep architecture. Here we show, using the CIFAR-10 database, that pooling decisions adjacent to the last convolutional layer significantly enhance accuracies. In particular, average accuracies of the advanced-VGG with $$m$$ m layers (A-VGGm) architectures are 0.936, 0.940, 0.954, 0.955, and 0.955 for m = 6, 8, 14, 13, and 16, respectively. The results indicate A-VGG8’s accuracy is superior to VGG16’s, and that the accuracies of A-VGG13 and A-VGG16 are equal, and comparable to that of Wide-ResNet16. In addition, replacing the three fully connected (FC) layers with one FC layer, A-VGG6 and A-VGG14, or with several linear activation FC layers, yielded similar accuracies. These significantly enhanced accuracies stem from training the most influential input–output routes, in comparison to the inferior routes selected following multiple MP decisions along the deep architecture. In addition, accuracies are sensitive to the order of the non-commutative MP and average pooling operators adjacent to the output layer, varying the number and location of training routes. The results call for the reexamination of previously proposed deep architectures and their accuracies by utilizing the proposed pooling strategy adjacent to the output layer.https://doi.org/10.1038/s41598-023-40566-y |
spellingShingle | Yuval Meir Yarden Tzach Ronit D. Gross Ofek Tevet Roni Vardi Ido Kanter Enhancing the accuracies by performing pooling decisions adjacent to the output layer Scientific Reports |
title | Enhancing the accuracies by performing pooling decisions adjacent to the output layer |
title_full | Enhancing the accuracies by performing pooling decisions adjacent to the output layer |
title_fullStr | Enhancing the accuracies by performing pooling decisions adjacent to the output layer |
title_full_unstemmed | Enhancing the accuracies by performing pooling decisions adjacent to the output layer |
title_short | Enhancing the accuracies by performing pooling decisions adjacent to the output layer |
title_sort | enhancing the accuracies by performing pooling decisions adjacent to the output layer |
url | https://doi.org/10.1038/s41598-023-40566-y |
work_keys_str_mv | AT yuvalmeir enhancingtheaccuraciesbyperformingpoolingdecisionsadjacenttotheoutputlayer AT yardentzach enhancingtheaccuraciesbyperformingpoolingdecisionsadjacenttotheoutputlayer AT ronitdgross enhancingtheaccuraciesbyperformingpoolingdecisionsadjacenttotheoutputlayer AT ofektevet enhancingtheaccuraciesbyperformingpoolingdecisionsadjacenttotheoutputlayer AT ronivardi enhancingtheaccuraciesbyperformingpoolingdecisionsadjacenttotheoutputlayer AT idokanter enhancingtheaccuraciesbyperformingpoolingdecisionsadjacenttotheoutputlayer |