Tools
Tools
Rai, Vivek Vindhyachal (2023) A novel AutoML library for pre-defined transfer learning workflows. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (1MB) | Preview |
Preview |
PDF (Configuration manual)
Download (1MB) | Preview |
Abstract
The field of artificial intelligence has experienced significant advancement due to the rapid growth of machine learning & pre-defined transfer learning techniques. As machine learning models become more complex there is a growing demand for automated and user-friendly model development. This research introduces an AutoML library that simplifies the utilization of pre-defined transfer learning workflows. The research includes the AutoML software which works with various pre-defined transfer learning models such as VGG19, ResNet-50, Imagenet, DenseNet-121 and EfficientNet-B7. These options are well-known for their versatility and outstanding results. We will show how users can bring their image classification datasets and use the our automl developed library to receive a comprehensive output. It includes performance measures like accuracy, precision, recall and F1 score for a wide range of transfer learning algorithms.To evaluate the new library two image classification use cases are presented with encouraging results. By addressing this research gap it is expected AutoML to play a greater role in simplifying and ensuring equal opportunities in machine learning empowering a wider audience and promoting further progress in the automated development of models.
Actions (login required)
 |
View Item |