Menu

Lab Course Automated Machine Learning

Note: You do not need to register for the fall school if you have registered for this course. Your participation in the fall school will be free of charge. You can register for the course via HISinOne.

Background

Applying machine learning (ML) and in particular deep learning (DL) in practice is a challenging task and requires a lot of expertise. Among other things, the success of ML/DL applications depends on many design decisions, including an appropriate preprocessing of the data, choosing a well-performing machine learning algorithm and tuning its hyperparameters, giving rise to a complex pipeline. Unfortunately, even experts need days, weeks or even months to find well-performing pipelines and can still make mistakes when optimizing their pipelines. The lab course will start with a fall school featuring lectures on hot topics in AutoML such as "automating data science", "automated reinforcement learning" or "neural architecture search", as well as present tutorials on various topics in AutoML. After this fall school, students are tasked with implementing their own AutoML system for a particular task or problem domain.

Requirements

We require that you have heard a lecture on

  • Machine Learning, and/or
  • Deep Learning

We strongly recommend that you have taken the AutoML lecture.

Organization

This lab course officially starts one week before the start of the lecture period. During that week, students take part in the AutoML fall school, to be held in Freiburg, to hear from world leading AutoML experts about current hot topics in the field. After this week, we will meet to present potential projects from which the students are free to select which one they want to tackle. We expect that students work in groups of up to three to implement their AutoML system. During the semester the students will meet with a supervisor to discuss potential issues they are facing when implementing their system. At the end of the semester all groups will present their work during a poster presentation.

Grading

The grades are determined based on the quality of the project part.

Important Dates

  • 10.10.22 - 13.10.22: AutoML Fall School
  • 19.10.22 14:00 - 16:00: Introduction to the topics In: HS 00 006 (G.-Köhler-Allee 082)
  • TBD: Poster Session