This workshop gives a practical introduction to deep learning, convolutional and recurrent neural networks, GPU computing, and tools to train and apply deep neural networks for natural language processing, images, and other applications. It consists of lectures and hands-on exercises. Keras (https://keras.io/) and PyTorch (https://pytorch.org/) will be used in the exercise sessions.
After the workshop the participants should have the skills and knowledge needed to begin applying deep learning for different tasks and utilizing the GPU resources available at PDC for training and deploying their own neural networks.
The participants are assumed to have working knowledge of Python and suitable background in data analysis, machine learning, or a related field. Previous experience in deep learning is not required, but the fundamentals of machine learning are not covered on this course.
Agenda
Time |
Day 1 / March 20, 2019 |
09:00-10:30 |
Lecture: Introduction to deep learning |
10:30-11:00 |
Exercises: Introduction to Notebooks, Keras fundamentals |
11:00-12:00 |
Lecture: Image data, multi-layer percepton networks, convolutional neural networks |
12:00-13:00 |
Lunch |
13:00-14:00 |
Exercises: Image classification with MLPs, CNNs |
14:00-15:00 |
Lecture: Text data, embeddings, neural NLP, recurrent neural network |
15:00 - 16:00 |
Exercises: Text sentiment classification with CNNs, RNNs |
Time |
Day 2 / March 21, 2019 |
09:00-10:00 |
Lecture: GPUs, batch jobs, using PDC-GPU |
10:00-12:00 |
Exercises: Image classification |
12:00-13:00 |
Lunch |
13:00-14:00 |
Exercises: Text categorization and labelling |
14:00-15:00 |
Lecture: Cloud, GPU utilization, multiple GPUs |
15:00-16:00 |
Exercises: Using multiple GPUs |
Pre-requisites:
1. Own laptop;
2. Valid account at PDC, KTH. After getting an acceptance confirmation you have to apply for a PDC account on Beskow supercomputer. The instructions how to apply for the account will be sent by the confirmation e-mail.
Lecturers: Markus Koskela & Mats Sjöberg (CSC-Finland)
Language: English
Price: Free of charge