Topics in Optimization for Machine Learning

Seminar: Topics in Optimization for Machine Learning

Optimization lies at the heart of many machine learning algorithms. This seminar teaches an overview of modern mathematical optimization methods for machine learning and data science applications. In particular, we will discuss the theoretical basics of stochastic optimization, the scalability of algorithms to large datasets, and challenges in distributed optimization, such as federated machine learning and privacy aspects in optimization. The seminar is based on a mix of foundational papers and recent publications.

Previous attendance at the summer term lecture "Optimization for Machine Learning" or "Lectures on Modern Optimization Methods" is recommended but not required. A solid understanding of basic optimization principles is sufficient.

Organization

In this seminar, students will learn to present, discuss, and summarize papers in different areas of optimization for machine learning, and critically engage with peer work through reviews and discussion. Specifically, each student will get a topic assigned to them, consisting of two papers (a lead and follow-up paper). Each student will

• Write a short paper (max 6 pages) on the assigned topic, using the two provided papers as a starting point.
• Deliver a 20-minute presentation (plus 10 minutes for discussion).
• Write two short reviews (max 1 page each) of papers on other topics and prepare three discussion questions for each. The reviews will be shared with the group.

The final grade will depend on the presentation, the written deliverables, and the active participation in the discussions.

Important Dates

• Kick-off meeting TBA (to be held online, via zoom).
• The reviews (and questions) must be submitted during the semester.
• The presentations will be scheduled between December - February (grouped into clusters of talks)
• Hand-in of report: TBA.

Deliverables

• 2 short reviews (each contributes 10% of your final grade): Each review (max 1 page) should address:
  1. What is the problem addressed by the paper?
  2. What was done before, and how does the paper improve on previous work?
  3. What are the strengths and the limitations of the techniques in the paper
  4. What part of the paper was difficult to understand?
  5. What are possible improvements or extensions of the techniques in the paper?

  In addition to your review, you will have to submit 3 questions to ask the paper's presenter.

• Participation in discussion (20%): Contribute to the discussion during the seminar meetings.

• Presentation (40%): You will prepare and deliver a 20-minute presentation (followed by 10 minutes of questions/discussion) of the paper assigned to you. You will have the possibility to get feedback on your slides before the presentation.

• Seminar Paper: (20%) You will write a seminar paper on the topic that you have presented. It must not be longer than 6 pages, not counting references and appendices. Papers can be shorter, but in general, the provided page limit is a good indicator of how long a paper should be.