Programme > Keynote speakers

Christian Theobalt MaxPlanck Institut	Themis Palpanas  University of Paris	Diane Larlus Naver Labs Europe	Gilles Louppe University of Liège	Aurélien Bellet Inria Lille

WEDNESDAY 

Christian Theobalt (MaxPlanck Institut)

Title: Capturing the Real World in Motion: New ways to unite graphics, vision and machine learning

Abstract: In this presentation, I will talk about some of the recent work we did on new methods for reconstructing computer graphics models of real world scenes from sparse or even monocular video data. These methods are based on bringing together neural network-based and explicit model-based approaches. I will also talk about new neural rendering approaches that combine explicit model-based and neural network based concepts for image formation in new ways. They enable new means to synthesize highly realistic imagery and videos of real work scenes under user control.

Short bio: Christian Theobalt is The Scientific Director of the new Visual Computing and Artificial Intelligence Department at the Max-Planck-Institute for Informatics, Saarbrücken, Germany. He is also a Professor of Computer Science at Saarland University, Germany. From 2007 until 2009 he was a Visiting Assistant Professor in the Department of Computer Science at Stanford University. He received his MSc degree in Artificial Intelligence from the University of Edinburgh, his Diplom (MS) degree in Computer Science from Saarland University, and his PhD (Dr.-Ing.) from the Max-Planck-Institute for Informatics.In his research he looks at algorithmic problems that lie at the intersection of Computer Graphics, Computer Vision and Machine Learning, such as: static and dynamic 3D scene reconstruction, neural rendering and neural scene representations, marker-less motion and performance capture, virtual humans, virtual and augmented reality, computer animation, intrinsic video and inverse rendering, computational videography, machine learning for graphics and vision, new sensors for 3D acquisition, as well as image- and physically-based rendering. He is also interested in using reconstruction techniques for human computer interaction. For his work, he received several awards, including the Otto Hahn Medal of the Max-Planck Society in 2007, the EUROGRAPHICS Young Researcher Award in 2009, the German Pattern Recognition Award 2012, the Karl Heinz Beckurts Award in 2017, and the EUROGRAPHICS Outstanding Technical Contributions Award in 2020. He received two ERC grants, an ERC Starting Grant in 2013 and an ERC Consolidator Grant in 2017.

Themis Palpanas (University of Paris)

Title: Scalable High-Dimensional Vector Similarity Search: The New Kid on the Block

Abstract: There is an increasingly pressing need, by several applications in diverse domains, for developing techniques able to analyze very large collections of high-dimensional vectors. Examples of such applications come from scientific, manufacturing and social domains, where in several cases they need to apply machine learning techniques for knowledge extraction. It is not unusual for these applications to involve vector collections in the order of hundreds of millions to billions, which are often times not analyzed in their full detail due to their sheer size. In this talk, we describe examples of data sources that produce high-dimensional vectors, and focus on two popular types: data series and deep network embeddings. We discuss the solutions that have been independently developed and are used for each one of these types, and present the surprising results that emerge when we compare these solutions.

Short bio: Themis Palpanas is Senior Member of the French University Institute (IUF), a distinction that recognizes excellence across all academic disciplines, and professor of computer science at the University of Paris (France), where he is director of the Data Intelligence Institute of Paris (diiP), and director of the data management group, diNo. He received the BS degree from the National Technical University of Athens, Greece, and the MSc and PhD degrees from the University of Toronto, Canada. He has previously held positions at the University of California at Riverside, University of Trento, and at IBM T.J. Watson Research Center, and visited Microsoft Research, and the IBM Almaden Research Center. His interests include problems related to data science (big data analytics and machine learning applications). He is the author of 9 US patents (3 of which have been implemented in world-leading commercial data management products), and 2 French patents. He is the recipient of 3 Best Paper awards, and the IBM Shared University Research (SUR) Award. He is currently serving on the VLDB Endowment Board of Trustees, as an Associate Editor in the TKDE, and IDA journals, as well as on the Editorial Advisory Board of the IS journal, and the Editorial Board of the TLDKS Journal. He has served as Editor in Chief for the BDR Journal (that he drove to an impact factor of 3.578 and cite score of 8.6), as General Chair for VLDB 2013, Associate Editor for VLDB 2022, 2019 and 2017, Research PC Vice Chair for ICDE 2020, and Workshop Chair for EDBT 2016, ADBIS 2013, and ADBIS 2014, General Chair for the PDA@IOT International Workshop (in conjunction with VLDB 2014), and General Chair for the Event Processing Symposium 2009.

THURSDAY

Diane Larlus (Naver Labs Europe)

Title: Lifelong visual representation learning

Abstract: Computer vision has found its way towards an increasingly large number of applications. One reason for this success is the development of large and powerful deep learning architectures which produce visual features that are generic enough to be applied directly to - or be the starting point of - a large variety of target tasks. Yet, training such generic architectures requires large-scale data, extensive human annotations, and heavy computational resources. In order to reduce the training cost of transferable descriptors, recent approaches have looked at ways to allow for noisy, fewer, or even no annotations to perform such pretraining. The first part of this presentation will cover our recent contributions in this direction. Second, we consider the deployment of a model which is sequentially exposed to different visual domains and incrementally adapts through model updates. Most standard learning approaches lead to fragile models which are prone to drift when such updates are performed, a problem known as the ‘catastrophic forgetting’ issue. In this part, we will discuss a strategy based on meta-learning and domain randomization designed to mitigate catastrophic forgetting.

Short bio: Diane Larlus is a principal research scientist at Naver Labs Europe and leads a Chair on Lifelong representation learning within the MIAI research institute of Grenoble. Her research mainly focuses on learning representations with weak supervision, continual learning, and visual search. As a PhD student, she worked at INRIA Grenoble, France. After a postdoctoral experience at TU Darmstadt, Germany, she joined the European research center of Xerox. She now works at NAVER LABS Europe.

Gilles Louppe (University of Liège)

Title: The frontier of simulation-based inference

Abstract: Many domains of science have developed complex simulations to describe phenomena of interest. While these simulations provide high-fidelity models, they are poorly suited for inference and lead to challenging inverse problems. In this talk, we will review the rapidly developing field of simulation-based inference and identify the forces giving additional momentum to the field. We will describe how the frontier is expanding so that a broad audience can appreciate the profound influence these developments may have on science.

Short bio: Gilles Louppe an associate professor in artificial intelligence and deep learning at the University of Liège (Belgium). Previously, he held positions as a research fellow at CERN and as a postdoctoral associate at New York University with the Physics Department and the Center for Data Science. His research is at the intersection of deep learning, approximate inference and the physical sciences. Together with his team and collaborators, he has been developing a new generation of simulation-based inference algorithms based on deep learning, with several applications in particle physics, astrophysics, astronomy or gravitational wave science.

FRIDAY

Aurélien Bellet (Inria Lille)

Title: Introduction to Federated Learning

Abstract: Federated learning (FL) is a novel machine learning paradigm where several participants collaboratively train a model while keeping their data decentralized. FL embodies the principles of focused data collection and minimization and can mitigate many of the systemic privacy risks and costs resulting from traditional, centralized machine learning and data science approaches. In this talk, I will provide an introduction to FL and some of its classic algorithms. I will then highlight some important questions related to data heterogeneity and privacy protection, and describe some recent contributions that aim to address them.

Short bio: Aurélien Bellet is a researcher at Inria (France). He obtained his Ph.D. from the University of Saint-Etienne (France) in 2012 and was a postdoctoral researcher at the University of Southern California (USA) and at Télécom Paris (France). His current research focuses on the design of privacy-preserving and federated machine learning algorithms. Aurélien has served as area chair for ICML (since 2019), NeurIPS (since 2020) and AISTATS (since 2022). He co-organized several international workshops on machine learning and privacy at NIPS/NeurIPS 2016, 2018 and 2020 and at CCS 2021. He also co-organizes FLOW, an online seminar on federated learning with 1000+ registered attendees.