Invited speakers

Dražen Brščić

Kyoto University, Japan

Towards Social Robots Working in Public

The talk will discuss the challenges of bringing robots, and in particular socially interactive robots, into human everyday environments. The presenter will introduce some of the past works that he and his colleagues have been doing, focusing on the observing and modeling of human behavior and the social interaction between robots and humans in public spaces.

Biography

Dražen Brščić received his Dr. Eng. degree in electrical engineering from The University of Tokyo, Japan in 2008. He spent 2 years as postdoctoral researcher at TU Munich, Germany and 5 years at the ATR institute in Kyoto, Japan. From 2017 to 2019 he was at the University of Rijeka, Croatia as Assistant Professor. He is currently an Associate Professor at Kyoto University, Japan.

Mario Lučić

Google Research (Brain team), Switzerland

Web: Mario Lucic

Learning General Visual Representations at Scale

Learning general visual representations useful for many downstream tasks is a key challenge in Computer Vision. For many years, Convolutional Neural Networks trained on the ImageNet dataset were the de-facto standard. Recently, with the introduction of the Vision Transformer and the MLP-Mixer, we have observed a strong shift towards models with fewer inductive biases, which, when trained at scale, may outperform the convolutional counterparts. In this talk we will discuss these neural architectures and their impact on downstream applications through the lens of accuracy, robustness, and model calibration.

Biography

Mario Lučić is a Staff Research Scientist at Google Research (Brain team) where he is pursuing fundamental challenges in machine learning and artificial intelligence. He received his Ph.D. in Computer Science from ETH Zurich (2017), a M.Sc. degree (cum laude) in Computer Science from Politecnico di Milano (Italy), and a B.Sc. degree in Computing from University of Zagreb (Croatia). During his PhD he was supported by an IBM Fellowship and was an Associated Fellow at the Max Planck ETH Center for Learning Systems. Before starting his doctoral studies, he was working at IBM Research on ML models for predictive maintenance. His research on machine learning and artificial intelligence has received awards at several premier conferences, most notably the best paper award at the International Conference on Machine Learning and the best student paper award at the International Conference on Artificial Intelligence and Statistics. Mario Lucic is serving as Area Chair for NeurIPS and ICLR conferences.

Amit Mishra

University of Cape Town, South Africa

Web: Amit Mishra

Bio-inspired methods for sensing and instrumentation

In this talk, the author shall discuss some of the recent projects he has been working on where the innovations have been inspired by the biosphere. In this, he shall present a new paradigm of sensing (enabled by machine learning) called application specific instrumentation. He shall, further, discuss another bio-inspired sensing system innovated by him called Communication based Sensing. Lastly, he will present some of his recent work on brain-inspired AI architecture.

Biography

Prof. Amit Kumar Mishra is currently a full professor in the Department of Electrical Engineering at the University of Cape Town (ranked at 157 by THE in 2021). Before Cape Town, Amit has worked in Australia and India. He did his PhD at the University of Edinburgh. His areas of interest include radar system design and applied machine learning. He has successfully supervised nine PhD students so far and holds five patents.

Vlaho Petrović

Carl von Ossietzky University of Oldenburg, Germany

Web: Vlaho Petrovic

Experimental approach to development and validation of wind farm control

Wind energy has been growing rapidly in recent decades, and it is currently the main alternative energy source. In order to reduce construction, maintenance and commissioning costs, wind energy production is typically organised in wind farms rather than single isolated wind turbines, leading to wake interactions among different turbines, which can lead to power losses and increased structural loads. Traditionally, these interactions have been neglected by wind farm controllers, and each wind turbine has been optimizing its own performance. However, as it has been shown in recent years, a coordinated wind turbine operation can significantly improve the wind farm performance, resulting in a lower cost of energy and better grid integration. This talk will present the main control objectives in a wind farm and possible wake control strategies. The current research gaps and open challenges regarding modelling, estimation and control of wind farms will be explained. The main focus will be on the importance of the experimental approach and our recent experimental activities, including a big turbulent wind tunnel with scaled wind turbine models and field experiments.

Biography

Vlaho Petrović received his PhD degree in Electrical Engineering from University of Zagreb, Faculty of Electrical Engineering and Computing. Currently, he is a senior researcher at ForWind, University of Oldenburg, leading a team with a focus on wind turbine and wind farm control. His research interests include experimental work, optimal and predictive control, and system identification.

Google Sites

Report abuse