After completion of this course the student is able to;
- describe and compare basic machine learning methods and techniques, associated complexity, such as discrimination, support vector machines, neural networks, deep learning;
- select and implement machine learning methods and apply it to a real life problem.
Intended Learning Outcomes
The required level of physics during and after the programme is determined nationally and internationally. In view of the objectives of the programme, education is aimed at acquiring:
- Thorough knowledge of the basic theories in the domain of physics and mathematics;
- Thorough knowledge of one or more sub-areas in the physics domain;
- Knowledge of physics technology, including skills for designing and using measurement instruments and experimental techniques;
- Orientation into the application areas of Applied Physics;
- Insight in the interrelation between sciences and the relationship between sciences, and the resulting responsibilities;
- Skills such as being able to independently acquire knowledge; being able to creatively and systematically contribute to resolving problems in the field; being able to cooperate with colleagues and non-colleagues; and communicational, social and organizational skills.
|
 |
|
This course can be done for 3 or 5 EC. Registration for both variants is via this course. During the course, you indicate to the teacher how many EC you would like to participate for.
Machine Learning (ML) and Artificial Intelligence (AI) are fast expanding topics in mathematics (e.g., statistics, control), computer science, robotics, engineering and science in general. In the field of computational physics and chemistry ML is on the rise to learn simple models based on data produced by numerically solving complex equations on supercomputers.
This course will give you an overview on ML (and a little AI) and will let you learn the concepts in a ‘hands-on’ manner.
(1) We start with
- 'programming in Python'. Do not worry if you have never programmed in Python: you will get Jupyter notebooks from level 0 up to the level that is required. Of course, you should have some affinity with programming;
- a little 'probability theory'. Algorithms can be analyzed, but methods in ML and AI look pretty ad hoc, especially when it comes to neural networks (deep learning). Statisticians claim that ML is in fact statistics.
(2) The first few weeks we treat the basic ML topics on supervised, unsupervised and reinforcement learning. This leads to methods for classification, clustering, linear and nonlinear regression, searching (so a little AI as well);
(3) In the second half of the course you will do a project: you can choose 'whatever' you like. This can be;
- building a neural network for some application,
- getting a deeper understanding of why certain algorithms (do not) work,
- joining in to a Kaggle competition problem,
- making a world champion (well, a very good) backgammon player,
- etc.
In any case, before you start on your project you should prepare a short project proposal (max. 1 A4) in which you write what you want to do, how you are going to do it and why you think this is the right strategy.
Assessment
- This course can be done for 3EC and 5EC. The extra 2EC will lie in the size of the project.
- The homework related to part (1) can be made in pairs or alone.
- The final project for the 3EC course is an individual assignment.
- The final project for the 5EC course will be made in small groups (~2-4, depending on the total number of students).
- The assignments in the first half should be handed in (completed Jupyter notebooks) with a description of what has been, and what has not been achieved.
- The end product of the project is a report and a piece of software. In all cases the end product will be reviewed by the other students trough a presentation and demonstration of the product. The learning outcome for you all will be high if the projects lead to different solutions/approaches.
- You, students, have a large freedom in choosing what to do. The more enthusiasm the better!
In the first meeting of the new semester we will discuss all the topics above: we construct the final plan of this year's course
|
|
|