Guest talk: Introduction to Industrial Applications of Federated Learning (FL): Opportunities and Challenges

Guest talk from Nokia, within the course CS-E4740 Federated Learning, on Monday 8 April at 16.15 in zoom. 

Introduction to Industrial Applications of Federated Learning (FL): Opportunities and Challenges

Abstract: In this talk, we will discuss some industrial applications of FL, with specific focus on computer vision. In the first part, we delve into the diverse array of applications where federated learning is revolutionizing processes, enhancing privacy, and optimizing models. The second part delves into the crucial aspect of model compression within federated learning, showcasing its significance in overcoming communication bottlenecks and optimizing resource utilization. In particular, we will discuss Neural Network Compression (NNC) as the first international Standard on compression of neural networks within FL for bandwidth saving.

Hamed R. Tavakoli, Doctor of Science (Tech.) from University of Oulu, is the Head of Visual AI Systems Research at Nokia. He conducts research in machine learning and contributes to the standardization of technologies relevant to artificial intelligence, including neural network compression. He is an experienced researcher with a background in developing deep learning algorithms at e.g. Aalto University in Finland.

Homayun Afrabandpey, Doctor of Science (Tech.) from Aalto University, is a senior researcher in the Visual AI Systems Research team at Nokia. His research interest lies in the broad field of machine learning with the focus being on probabilistic modeling, machine learning interpretability, and distributed learning. He is part of the team Nokia contributing to the standardization activities related to machine learning and data compression including Neural Network Compression (NNC) and Feature Coding for Machines (FCM). 

This guest talk is hosted by Associate Professor Alex Jung, Department of Computer Science.