Zuse Research Seminar
What? The Zuse Research Seminar at the Zuse Institute Berlin serves as an interdisciplinary forum for researchers in the field of Applied Mathematics and Computer Science.
How? Talks are plenary-style and last 45 minutes plus time for questions. They should provide both an accessible overview over the research field as well as some broadly understandable insights into the speaker's own research and how it relates to the field. Active discussion and questions are encouraged.
Who? Talks are predominantly given by researchers from the institute with the aim to showcase their department's work in an accessible way. We also invite a select number of external speakers each semester to present their research. The seminar is open to the public and we encourage everyone interested to attend.
When? The seminar includes about five talks each semester during lecture time and talks are usually scheduled in the morning from 11:30 to 12:30. Coffee is available in the lobby half an hour before the talk.
Where? The venue is the institute's large lecture hall.
Backpropagation is the major workhorse for many machine learning algorithms. In this presentation, we will examine the theory behind backpropagation as provided by the technique of algorithmic differentiation. Subsequently, we will discuss how this classic derivative information can be used for nonsmooth optimization. Examples from reail will illustrate the application of the proposed nonsmooth optimization algorithm.
Please contact Tim Conrad and Christoph Spiegel by email for any organisational questions.