Incremental training of CNNs for user customization
| Autor: | Kiyoung Choi, Hyemi Min, Sukjin Kim, Mansureh S. Moghaddam, Barend Harris, Soonhoi Ha, Duseok Kang, Inpyo Bae, Euiseok Kim, Bernhard Egger, Hansu Cho |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Caffè
business.industry Computer science Real-time computing Work in process Machine learning computer.software_genre Convolutional neural network Personalization 03 medical and health sciences 0302 clinical medicine Factor (programming language) 030212 general & internal medicine Artificial intelligence Inference engine Transfer of learning business computer MNIST database computer.programming_language |
| Zdroj: | CASES |
| DOI: | 10.1145/3125501.3125519 |
| Popis: | This paper presents a convolutional neural network architecture that supports transfer learning for user customization. The architecture consists of a large basic inference engine and a small augmenting engine. Initially, both engines are trained using a large dataset. Only the augmenting engine is tuned to the user-specific dataset. To preserve the accuracy for the original dataset, the novel concept of quality factor is proposed. The final network is evaluated with the Caffe framework, and our own implementation on a coarse-grained reconfigurable array (CGRA) processor. Experiments with MNIST, NIST'19, and our user-specific datasets show the effectiveness of the proposed approach and the potential of CGRAs as DNN processors. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|