Adrian Galdran 

Thursday, 15th of June 2023 

Room and time TBA

Title: A Tutorial on Model Calibration for Neural Nets 

Abstract: A calibrated machine learning model produces probabilistic predictions that are well-aligned with real probabilities: it tends to be more certain when it is correct. Unfortunately, the unique characteristics of modern neural networks, e.g. over-parametrization or iterative training dynamics, can often result in overfitting the training data and generating over-confident predictions. The goal of this talk is to introduce machine learning practitioners to the main ideas and methods of modern model calibration: its formal definition, how we can measure it, and how we can improve it. This talk is part of a tutorial (accepted to MICCAI 2023) on uncertainty quantification for medical image analysis, and it will also have a short hands-on code session: we will train a skin lesion classifier in pytorch, measure its calibration, and learn a post-processing transform to improve it. About the speaker: Adrian Galdran is a Marie Skłodowska-Curie Research Fellow at Universitat Pompeu Fabra in Barcelona jointly with the Australian Institute of Machine Learning in Adelaide, leading a project on Uncertainty Quantification for medical imaging. Prior to this, he worked as lecturer/post-doctoral researcher in biomedical image analysis at Bournemouth University in the UK, ETS Montréal in Canada and INESC-TEC Porto in Portugal.

Biography: Adrian Galdran, UPF & Australian Institute for Machine Learning Date

Schahram Dustdar 

Friday 24th of February 2023, 14:30 am 52.123 

Title: Edge Intelligence - Research Opportunities in the Distributed Computing  Continuum 

Abstract: As humans, Internet of Things, software services and AI continue to  become the entangled fabric of distributed systems, systems engineers and researchers  are facing novel challenges. In this talk, we analyze the role of AI in the context of IoT, Edge, and Cloud in the co-evolution of distributed systems for the new decade. We  identify challenges and discuss a roadmap that these new distributed systems have to  address in order to bring intelligence to the edge. We take a closer look at how a cyber physical fabric will be complemented by AI operationalization to enable seamless end to-end distributed systems. 

Biography: Schahram Dustdar is a Full Professor of Computer Science at the  Vienna Technical University, heading the Research Division of Distributed Systems at the TU Wien, Austria. He holds several honorary positions: University of California (USC) Los Angeles; Monash University in Melbourne, Shanghai University, Macquarie University in Sydney, University Pompeu Fabra, Barcelona, Spain.  

From Dec 2016 until Jan 2017 he was a Visiting Professor at the University of Sevilla,  Spain and from January until June 2017 he was a Visiting Professor at UC Berkeley, USA.  

From 1999 – 2007 he worked as the co founder and chief scientist of Caramba Labs Software AG in Vienna (acquired by ProjectNetWorld AG), a venture capital co-funded software company focused on  software for collaborative processes in teams. He is co-founder of edorer.com (an  EdTech company based in the US) and co-founder and chief scientist of Sinoaus.net,  a Nanjing, China based R&D organization focusing on IoT and Edge Intelligence. 

He is founding co-Editor-in-Chief of ACM Transactions on Internet of Things (ACM  TIoT) as well as Editor-in-Chief of Computing (Springer). He is an Associate Editor of  IEEE Transactions on Services Computing, IEEE Transactions on Cloud Computing,  ACM Computing Surveys, ACM Transactions on the Web, and ACM Transactions on  Internet Technology, as well as on the editorial board of IEEE Internet Computing and  IEEEComputer. 

Dustdar is recipient of multiple awards: IEEE TCSVC Outstanding Leadership Award  (2018), IEEE TCSC Award for Excellence in Scalable Computing (2019), ACM  Distinguished Scientist (2009), ACM Distinguished Speaker (2021), IBM Faculty  Award (2012). He is an elected member of the Academia Europaea: The Academy of  Europe, as well as an IEEE Fellow (2016) and an Asia-Pacific Artificial Intelligence  Association (AAIA) Fellow (2021) and the AAIA president (2021).