Statistical Machine Learning (SML) PhD course (9+3hp)
During the course you are encouraged to carry out a project (1-3 students in each team), although it is not mandatory. The main purpose of these projects is to provide the opportunity to learn more about concepts relevant to the course and for successfully doing this you will get payed 3hp. Some of the projects usually produce fruitful ideas/starting points for new research directions.
The theme for the 2016 edition of the course is high-energy particle physics (HEP). This means that the preferred theme for a project falls within the area of HEP. However, you can probably convince me to be allowed to do a project with another application. Besides the obvious requirement that the topic of the project must fit the course, there is one more requirement, which is that all projects must be centered around a data set or some sort.
A solid project would be to try to reproduce some of the results from one of the papers written based on the open datasets, see for example these papers from the 2014 NIPS workshop. Even better would of course be to try out an idea of your own.
16 March Project proposals are due. Send the pdf via e-mail to Thomas Schön.
18 March Project proposal presentations, 10.15-12 in room ITC/2344. Send your presentation as a pdf-file to Marcus Björck (marcus.bjork_at_it.uu.se) no later that 09.00 on 18 March.
18 April Final reports are due
21 April Final project presentations
Brief description of the steps mentioned above.
Project Proposal: The report should explain the project background, the data set, the idea and how the work will be carried out. Upper page limit: 2. For the presentation, each team has 5-10 minutes for presentation and discussion.
Final Report: Final project report, clearly explains what has been accomplished within the project. The data set used should also be properly introduced. Upper page limit: 5.
Final Presentation: We will use the presentation form used at most conferences, that is each team has 15 min. for presentation and 5 min. for discussion. These time limits are strict.