Low-cost Retrain
- 0 Collaborators
Some supervised learning tasks require continually re-training models. In this project, I explore different alternatives to minimize the re-training costs with the purpose of deploying deep learning model retraining on resource-constrained platforms. ...learn more
Project status: Concept
Mobile, Internet of Things, Artificial Intelligence
Intel Technologies
DevCloud,
oneAPI,
AI DevCloud / Xeon,
Intel Python
Overview / Usage
Edge training is the main application of this project, but it can also be applied to other resource-constrained scenarios. For example, in computer architecture prediction-based modules. Researchers in computer architecture have proposed the use of neural networks to tackle prediction and optimization problems at the hardware level. One of the major limitations of using NN at this level is the latency and power consumption of such predictors.