Tutorial on Transfer Learning for Deep Learning
A hands-on tutorial on Transfer Learning for Deep Neural Networks will be carried out during IWANN 2019. The tutorial will include a practical session with connection to the Barcelona Supercomputing Center.
Organizer:
Dario Garcia-Gasulla
High Performance Artificial Intelligence Group
Departament of Computer Science
Barcelona Supercomputing Center
Abstract:
Training deep neural networks from scratch is not easy. Finding a good model for a given problem requires of huge amounts of data, lots of computational power, and a team of DL experts dedicated to the task for weeks. Since we cannot dedicate these resources for every single problem that may be appropriate for deep learning, the community has been actively looking for easier and faster solutions, mostly focused on the reuse of pre-trained deep learning models. This is the main goal of the transfer learning field, which seeks to exploit models designed and trained for a problem A to solve a potentially unrelated problem B. In this tutorial we will introduce the two main approaches to transfer learning, fine tuning and feature extraction, detailing the benefits and handicaps of each one. We will provide hands-on experience on running both types of transfer learning, while working on the CTE-POWER9 cluster hosted at BSC, which includes state-of-the-art Volta GPU racks.
Scheduling:
- One hour theory on fine tunning, transfer leaning and feature extraction.
- One hour practice with the CTE-POWER9 Cluster of the Barcelona Supercomputing Center.
Best Paper and Best Special Session Prizes
As a part of IWANN 2019 scientific programme, two prizes will be awarded:
– Best contribution.
– Best special session.
Each award will include a cash prize granted by Springer, publisher of the proceedings within its LNCS series. The support by Springer is gratefully acknowledged.
The concession will be decided by IWANN 2019 chairs, based upon scientific merit and relevance.
Workshop
| Artificial Intelligence in Nanophotonics | Dr. Nikolay Zheludev Dr. Cesare Soci |
Confirmed Special Sessions
| # | Title | Organizers |
| SS01 | Artificial Neural Network for biomedical image processing |
Dr. Yu-Dong Zhang |
| SS02 | Deep learning models in healthcare and biomedicine | Dr. Leonardo Franco Dr. Ruxandra Stoean Dr. Francisco Veredas |
| SS03 | Deep learning beyond convolution | Dr. Miguel Atencia |
| SS04 | Machine Learning in Vision and Robotics | Dr. José García-Rodríguez Dr. Enrique Domínguez Dr. Ramón Moreno |
| SS05 | Data-driven Intelligent Transportation Systems | Dr. Ignacio J. Turías Domínguez Dr. David Elizondo Dr. Francisco Ortega Zamorano |
| SS06 | Software Testing and Intelligent Systems | Dr. Juan Boubeta Dr. Pablo C. Cañizares Dr. Gregorio Díaz |
| SS07 | Deep Learning and Natural Language Processing | Dr. Leonor Becerra-Bonache Dr. M. Dolores Jiménez-López Dr. Benoit Favre |
| SS08 | Random-Weights Neural Networks | Dr. Claudio Gallicchio |
| SS09 | New and future tendencies in Brain-Computer Interface systems | Dr. Ricardo Ron Dr. Ivan Volosyak |
| SS10 | Human Activity Recognition | Dr.-Ing. habil. Matthias Pätzold |
| SS11 | Computational Intelligence Methods for Time Series | Dr. Héctor Pomares |
| SS12 | Advanced Methods for Personalized/Precision Medicine | Dr. Luis Javier Herrera Dr. Fernando Rojas |
| SS13 | Exploring document information to improve neural summarization models | Dr. Luigi Di Caro |
| SS15 | Machine learning in weather observation and forecasting | Dr. Juan Luis Navarro-Mesa Dr. Antonio Ravelo-García Dr. Carmen Paz Suárez Araujo |