OptiNET—Automatic Network Topology Optimization
The recent boom of artificial Neural Networks (NN) has shown that NN can provide viable solutions to a variety of problems. However, their complexity and the lack of efficient interpretation of NN architectures (commonly considered black box techniques) has adverse effects on the optimization of eac...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-08-01
|
| Series: | Information |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2078-2489/13/9/405 |
| _version_ | 1797486906856964096 |
|---|---|
| author | Andreas Maniatopoulos Paraskevi Alvanaki Nikolaos Mitianoudis |
| author_facet | Andreas Maniatopoulos Paraskevi Alvanaki Nikolaos Mitianoudis |
| author_sort | Andreas Maniatopoulos |
| collection | DOAJ |
| description | The recent boom of artificial Neural Networks (NN) has shown that NN can provide viable solutions to a variety of problems. However, their complexity and the lack of efficient interpretation of NN architectures (commonly considered black box techniques) has adverse effects on the optimization of each NN architecture. One cannot simply use a generic topology and have the best performance in every application field, since the network topology is commonly fine-tuned to the problem/dataset in question. In this paper, we introduce a novel method of computationally assessing the complexity of the dataset. The NN is treated as an information channel, and thus information theory is used to estimate the optimal number of neurons for each layer, reducing the memory and computational load, while achieving the same, if not greater, accuracy. Experiments using common datasets confirm the theoretical findings, and the derived algorithm seems to improve the performance of the original architecture. |
| first_indexed | 2024-03-09T23:41:01Z |
| format | Article |
| id | doaj.art-852a38c7c73a4291b140986503ffaebe |
| institution | Directory Open Access Journal |
| issn | 2078-2489 |
| language | English |
| last_indexed | 2024-03-09T23:41:01Z |
| publishDate | 2022-08-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Information |
| spelling | doaj.art-852a38c7c73a4291b140986503ffaebe2023-11-23T16:53:01ZengMDPI AGInformation2078-24892022-08-0113940510.3390/info13090405OptiNET—Automatic Network Topology OptimizationAndreas Maniatopoulos0Paraskevi Alvanaki1Nikolaos Mitianoudis2Electrical and Computer Engineering Department, Democritus University of Thrace, 69100 Komotini, GreeceElectrical and Computer Engineering Department, Democritus University of Thrace, 69100 Komotini, GreeceElectrical and Computer Engineering Department, Democritus University of Thrace, 69100 Komotini, GreeceThe recent boom of artificial Neural Networks (NN) has shown that NN can provide viable solutions to a variety of problems. However, their complexity and the lack of efficient interpretation of NN architectures (commonly considered black box techniques) has adverse effects on the optimization of each NN architecture. One cannot simply use a generic topology and have the best performance in every application field, since the network topology is commonly fine-tuned to the problem/dataset in question. In this paper, we introduce a novel method of computationally assessing the complexity of the dataset. The NN is treated as an information channel, and thus information theory is used to estimate the optimal number of neurons for each layer, reducing the memory and computational load, while achieving the same, if not greater, accuracy. Experiments using common datasets confirm the theoretical findings, and the derived algorithm seems to improve the performance of the original architecture.https://www.mdpi.com/2078-2489/13/9/405topology optimizationnetwork optimizationpruning |
| spellingShingle | Andreas Maniatopoulos Paraskevi Alvanaki Nikolaos Mitianoudis OptiNET—Automatic Network Topology Optimization Information topology optimization network optimization pruning |
| title | OptiNET—Automatic Network Topology Optimization |
| title_full | OptiNET—Automatic Network Topology Optimization |
| title_fullStr | OptiNET—Automatic Network Topology Optimization |
| title_full_unstemmed | OptiNET—Automatic Network Topology Optimization |
| title_short | OptiNET—Automatic Network Topology Optimization |
| title_sort | optinet automatic network topology optimization |
| topic | topology optimization network optimization pruning |
| url | https://www.mdpi.com/2078-2489/13/9/405 |
| work_keys_str_mv | AT andreasmaniatopoulos optinetautomaticnetworktopologyoptimization AT paraskevialvanaki optinetautomaticnetworktopologyoptimization AT nikolaosmitianoudis optinetautomaticnetworktopologyoptimization |