Dr Jasmina Lazić (MathWorks) will be delivering two one-hour sessions on optimization and machine learning using MATLAB. This event is part of the Royal Society (UK Department of International Development Programme Grant) project “Harnessing unsteady phase-change heat exchange in high-performance concentrated solar power systems”.
10:00 – 11:00 – Best practices for Optimisation in MATLAB
In this section, we will discuss some best practices for setting up, solving and speeding up optimisation problems. You will learn about:
- How to set up optimisation problems in MATLAB?
- Choosing the best solver
- Local vs. global optimisation
- Accelerating your optimisation by:
- Using parallel computing
- Tuning solver options
- Leveraging symbolic computing
- MATLAB optimisation product enhancements in recent releases
11:00 – 12:00 – MATLAB for Image Processing with Machine Learning Applications
We demonstrate how MATLAB can accelerate algorithm development and exploration when working with image data and vision systems. We will cover the use of machine learning and deep learning methods for advanced Image Processing and Computer Vision applications. You will learn how to:
- Analyze and process (e.g. enhance, measure, segment) images
- Acquire live video from external cameras and deal with large collections of image files
- Categorize and retrieve images using advanced clustering and classification techniques
- Train models using large image datasets:
- Train deep neural networks from scratch
Using transfer learning to re-use trained deep networks for new tasks - Explore the tradeoffs between machine learning and deep learning
- Train deep neural networks from scratch
- Leverage multicore CPU and GPU computing for processing large images
Bio
Dr Jasmina Lazić is an application engineer at MathWorks, specialising in optimisation techniques and image processing in MATLAB. Before joining MathWorks in 2011, Jasmina held a number of academic positions focused on research in global optimisation and heuristic design. She has publications in the areas of mixed-integer programming and clustering. Jasmina has a Ph.D. in mathematics from Brunel University and a diploma degree in mathematics and computer science from the University of Belgrade in Serbia.