Date: Monday to Friday, 22 to 26 August, 2022.
Time: 10 a.m. to 3:30 p.m. (with a lunch break from 12 to 1:30).
Location: room 511 (5th floor), inside the Geographic Information Center.
Instructor: , Associate Professor of Statistics, Department of Mathematics and Statistics.
Overview
This 5-day workshop is an introductory, short course on deep learning. It is designed for anyone who wishes to obtain basic knowledge of machine learning and explore deep learning from scratch. The workshop provides you with a practical introduction to basic and selected topics on deep learning through the pytorch framework in Python.
The outcomes of this workshop include knowledge of:
- The building blocks of neural networks
- Introduction to pytorch
- Classification and regression with neural networks
- Fundamentals and workflows of machine learning
- Deep learning for time series (and text)
听
Prerequisites
- Basic Python programming experience. Familiarity with NumPy library would be helpful, however it is not required.
- High school-level mathematics.
听
Optional materials: Deep Learning with Python, Second Edition, by Francois Chollet (Textbook)
Lesson plan
Day 1
- Data representations for neural networks: tensors, vectors, matrices.
- Tensor operations
- Gradient-based optimization
Day 2
- Jupyter notebooks and Colaboratory
- First steps with Keras and Tensor flow
- Understanding core Keras APIs
Day 3
- Getting started with a binary classification example: movie reviews
- A multiclass classification example: newswires
- A regression example: predicting house prices
Day 4
- Evaluating machine learning models
- Improving model fit and generalization
- Model development and deployment
Day 5
- Different kinds of time series tasks
- A temperature-forecasting example
- Getting started with recurrent neural networks
听
Registration
Registration for the workshop Introduction to deep learning is closed.
By registering, you commit to attend and participate actively. In the event that you need to cancel your registration, please notify us as soon as possible so that we can offer your space to another participant. To cancel registration, fill in this form.
