Показати скорочений опис матеріалу
Why do we need a Post-Train Adaptive neural network?
dc.contributor.author | Khabarlak, Kostiantyn | |
dc.date.accessioned | 2024-07-08T11:45:14Z | |
dc.date.available | 2024-07-08T11:45:14Z | |
dc.date.issued | 2024 | |
dc.identifier.citation | Khabarlak K. Why do we need a Post-Train Adaptive neural network? / Kostiantyn Khabarlak // Інформаційні технології: теорія і практика : тези доповідей I (VII) міжнародної науково-практичної конференції здобувачів вищої освіти і молодих учених (Дніпро 20-22 березня 2024) – Дніпро : Свідлер А.Л., 2024. – С. 23-25. | uk_UA |
dc.identifier.uri | http://ir.nmu.org.ua/handle/123456789/167282 | |
dc.description.abstract | Neural networks have shown to be effective in many areas. Convolutional neural networks solve computer vision problems, such as classification, detection and segmentation depending on task at human level or better. Recurrent and transformer-based neural networks are actively used for natural language understanding. More recently neural networks have shown high quality in generative tasks in both vision and language domains. All of it results in an increased usage of neural networks. Some of the applications require offline data processing due to privacy requirements, lack of Internet access or high server load which is better to be distributed among user devices to reduce server maintenance cost. | uk_UA |
dc.language.iso | en | uk_UA |
dc.subject | neural networks | uk_UA |
dc.subject | нейронні мережі | uk_UA |
dc.subject | server | uk_UA |
dc.subject | сервер | uk_UA |
dc.title | Why do we need a Post-Train Adaptive neural network? | uk_UA |
dc.type | Article | uk_UA |
dc.identifier.udk | 004.93 | uk_UA |