Delete search term

Header

Quick navigation

Main navigation

CAI Colloquium

We invite AI experts to present and discuss topics relevant to the scope of the CAI. The colloquium takes place on Wednesdays from 11:00-12:00 at the Winterthur Campus of ZHAW.

Future Events

Date Time&Place Speaker Title Abstract
Wed, 01.03. 11:00-12:00, TN O1.46 and Zoom Prof. Dr. Thomas Martinetz (Univ. Luebeck, D) Large deep neural networks do learn with small data sets Traditional machine learning wisdom was, that it needs more training data points than there are parameters in a network to be able to learn a given task. Learning theory based on VC-dimension or Rademacher Complexity provides an extended and deeper framework for this "wisdom". Modern deep neural networks have millions of weights, so one needs extremely large training data sets. That's the common narrative. But is it really true? In practice, these large neural networks are often trained with much less data than one would expect to be necessary. We show in experiments that even a few hundred data points can be sufficient for millions of weights, and provide a mathematical framework to understand this surprising phenomenon.
Wed, 10.05. 11:00-12:00, t.b.a. Prof. Dr. Jan Dirk Wegner (UZH) Large-scale analysis of geospatial data with machine learning Worldwide analyses and estimates of vegetation parameters such as biomass or vegetation height are essential for modelling climate change and biodiversity. Traditional allometric approaches usually have to be adapted for specific ecosystems and regions. It is therefore very difficult to carry out homogeneous, global modelling with high spatial and temporal resolution and, at the same time, good accuracy. Data-driven approaches, especially modern deep learning methods, promise great potential here. In this talk, new research results on the large-scale determination of vegetation parameters will be presented.

Past Events

Date Speaker Title Slides
2022      
Wed, 23.11. Prof. Dr. Benjamin Grewe (ETHZ/UZH) Why auto-encoding is not enough
Wed, 19.10. Frank Wittmann & Meret Weiser (ZHAW) Algorithms and ML in Social Work
Wed, 20.7. 2nd Panel Discussion (Ch.v.d. Malsburg, R. Douglas, Y. Sandamirskaya, B. Grewe, T. Stadelmann, R. Chavarriaga) Pathways beyond present AI (Part 2): Artificial Intelligence: Game Over? News
Wed, 27.04. 1st Panel Discussion Pathways beyond present AI News
Wed, 30.3. Dr. Lucas Beyer (Google Brain) Transformers as general vision backbones
Wed, 2.3. Sebastian Welter (IKEA) Scaling AI in Enterprises PDF
Wed, 16.2. Prof. Dr. Volker Dellwo (UZH) Can speakers make themselves more recognisable?: Voice dynamics and its influence on voice recognition PDF
2021      
Wed, 17.11. Prof. Dr. Christoph von der Malsburg How can kids learn so much faster than transformers? Recording PDF
Wed, 8.9. Prof. Nicolaj Stache (HS Heilbronn) From Simulation to Reality using Reinforcement Learning PDF
Wed, 25.8. Prof. Marco Gori (U Siena) Learning to See by Motion Invariance PDF