Kick-off

On Tuesday December 8th 2026, we will kick-off the ELLIS Unit of Grenoble. To do so, we have put together a program with a set of wonderful speakers:

Key info

  • When: December 8th, 2026 – full day.
  • Where: Amphithéâtre Louis Néel, Bâtiment Ampère (GrenobleINP, UGA).
  • How to get there: Tram B or C, stop “Gabriel Faure – MUSE”, then follow the map below.
  • Registration is free, but mandatory, request your spot using this link.

List of Speakers

(Please, see the talks details below)

Programme (tentative)

09:00 — Opening
09:45 — Serge Belongie: Challenges in Fine-Grained Image Analysis
10:30 — Group Picture
10:45 — Coffee Break
11:15 — Mirco Ravanelli
12:00 — Dima Damen
12:45 — Lunch Break
14:15 — Laura Sevilla
15:00 — Hayley Hung
15:45 — Coffee Break + Poster Session
16:30 — Justin Carpentier
17:15 — Closing

Details of the talks

Serge Belongie: Challenges in Fine-Grained Image Analysis

Abstract: Fine-grained image analysis (FGIA) is a longstanding and fundamental problem in computer vision and pattern recognition, underpinning a diverse set of real-world applications. The task of FGIA concerns visual objects from subordinate categories, e.g., species of birds or models of cars. Capitalizing on advances in Large Vision-Language Models (LVLMs), in recent years we have witnessed remarkable progress on a variety of FGIA problems. In this talk we review representative examples in the context of recognition and retrieval alongside other key issues such as publicly available benchmark datasets, domain-specific applications, and connections with other modalities including text and audio. We conclude by highlighting several research directions and open problems.

Bio: Serge Belongie is a professor of Computer Science at the University of Copenhagen, where he also serves as the head of the Pioneer Centre for Artificial Intelligence (P1). Previously, he was a professor of Computer Science at Cornell University, an Associate Dean at Cornell Tech, a member of the Visiting Faculty program at Google, and a professor of Computer Science & Engineering at UC San Diego. His research interests include Computer Vision, Machine Learning, Augmented Reality, and Human-in-the-Loop Computing. He is also a co-founder of several companies including Digital Persona and Anchovi Labs. He is a recipient of the NSF CAREER Award, the Alfred P. Sloan Research Fellowship, the MIT Technology Review “Innovators Under 35” Award, the Stibitz-Wilson Award, the Helmholtz Prize, the Everingham Prize, and the Koenderink Prize for fundamental contributions to the Computer Vision community. He is a member of the Royal Danish Academy of Sciences and Letters and serves as President of the board of the European Laboratory for Learning and Intelligent Systems (ELLIS).

Mirco Ravanelli: Speech Representations in the Era of LLMs

Abstract: Large Language Models (LLMs) are transforming conversational AI, enabling systems that can listen, understand, and respond naturally through speech. In this talk, I will explore how speech representations have evolved and why they play a crucial role in modern AI systems. I will then highlight key challenges and present recent advances from our research toward more efficient, compact, and controllable speech representations for the next generation of conversational AI.

Bio: Mirco Ravanelli is an Associate Professor at Concordia University and an Associate Member of Mila – Quebec AI Institute. His research focuses on deep learning and Conversational AI, with particular emphasis on speech processing, representation learning, multimodality, and interpretability. He has authored over 100 publications in machine learning and speech processing venues and has received several international awards for his research. He founded and leads SpeechBrain, one of the most widely adopted open-source toolkits for speech and conversational AI, used by researchers and practitioners worldwide. His recent research explores speech tokenization, multimodal LLMs, and efficient foundation models for audio and speech.

Dima Damen: Modelling in an Ego-Sensed World

Abstract: Forecasting the rise of wearable devices equipped with audio-visual feeds and onboard spatial sensors, this talk will present novel approaches for generative and perception world modelling from egocentric videos. The talk argues for new ways to foresee egocentric videos as partial observations of a dynamic 3D world, where objects are out of sight but not out of mind and activities take place in a physically-grounded space. All projects details are at: https://dimadamen.github.io/index.html#Projects 

Bio: Dima Damen is a Professor of Computer Vision at the University of Bristol and Senior Research Scientist at Google DeepMind. Her research contributes to first-person (egocentric) vision, video understanding, multimodal machine learning, with focused research on human-object interactions, fine-grained action recognition, audio-visual learning, and integrating video understanding with 3D/4D spatial reasoning to model dynamic physical environments. She has also contributed to novel research questions including learning from streaming video, learning from a single-life experience, mono-to-3D, video object segmentation, assessing action completion, domain adaptation, skill/expertise determination from video sequences, discovering task-relevant objects, dual-domain and dual-time learning as well as multi-modal fusion using vision, audio and language. She is the project lead for EPIC-KITCHENS, the seminal dataset in egocentric vision, with accompanying open challenges and follow-up works: EPIC-Sounds, VISOR and EPIC Fields, as well as the recent HD-EPIC. She is part of the large-scale consortium effort Ego4D and Ego-Exo4D. She is an ELLIS Fellow and was EPSRC Early Career Fellow (2020-2026), associate editor (AE) of IJCV (2020-2026), program chair for ICCV 2021 and Associate Editor-in-Chief (AEIC) of IEEE TPAMI (2023-2025). She is frequently a Senior Area Chair and Area Chair in major conferences and was selected as Outstanding SAC in ECCV 2024 and Outstanding Reviewer in CVPR2021, CVPR2020, ICCV2017, CVPR2013 and CVPR2012. Dima received her PhD from the University of Leeds (2009), joined the University of Bristol as a Postdoctoral Researcher (2010-2012), Assistant Professor (2013-2018), Associate Professor (2018-2021) and was appointed as chair in August 2021. She supervises 3 postdoctoral researchers, 8 PhD students and 1 Visiting PhD student. At the University of Bristol, Dima leads the Machine Learning and Computer Vision (MaVi) lab. At Google DeepMind, Dima is part of the Vision team, led by Andrew Zisserman, focusing on video understanding research. Her latest contribution is to the CVPR 2026 paper: Unique Lives, Shared World: Learning from Single-Life Videos

Hayley Hung: On being human in the age of Artificial (Superhuman?) Intelligence

Abstract: Modern developments machine learning are changing what it means to be human. Fundamental to being is the need for social connection, a sense of belonging, and agency. In this talk, I argue that an underexplored role of ML and AI that is to investigate how to support humans to become super empathetic, super emotionally intelligent, super team workers, super productive, etc. In other words, what can be done to develop systems that can help us not only remember what it means to be human but to also overcome our weaknesses and accelerate our strengths as emotional and social beings? In this talk, I discuss how in my own journey to address these goals, I reach into fundamental questions about multimodal multisensor learning, modelling and evaluation of subjective phenomena, how learning and perception might change or not as we trade off privacy for ubiquity, what happens when neither human nor machine has access to the ground truth, and whether it is even always desirable to assume a single truth. I conclude the talk by challenging whether ML research practices have evolved appropriately and offer an alternate perspective on a predominantly bold-number driven field.

Bio: Hayley Hung is an Associate Professor at Delft University of Technology. She leads the Human Oriented Machine Intelligence Unit which consists of three labs focused on Artificial Social Perception, Digital Human Generation, Cognition and Affective Modelling. In 2023, she was awarded on ERC Consolidator Grant on modelling the multiplicity of apparent intent. Her research contributions have also been recognised by an invited talk at the ACM Multimedia 2016 Rising Star Session. She co-founded the emotional and social signals area at ACM Multimedia and has contributed various organizational roles including being Program Chair in 2019. She is an Associate Editor in Chief of the IEEE Transactions on Affective Computing. She is an ELLIS Fellow and member of the ELLIS Unit Delft Management Board.

Map to reach the venue