Detailed schedule
Monday 26/08
Venue: Campus des Cordeliers, Sorbonne University (15 rue de l’école de médecine)
11:00 – 13:00
Walk-in, registration, lunch
13:00 – 13:30
Welcome address
13:30 – 15:00
Keynote address
Devis Tuia (EPFL) - Machine learning for Earth: monitoring the pulse of our Planet with sensor data, from your phone all the way to space
Abstract: We live an age full of data. In all areas of society, digital data is now abundant, but also unstructured and pretty much unexploited. Environmental science is no exception and the last years have seen an increase of use of digital sensing to observe an understand processes at work and their impacts of human activities. In this talk, I will present some recent research at the interface of machine learning and the geosciences, where satellite, drone and close sensing data were used to support mapping over land and sea, and even below the surface. I will then sketch a number of points of synergetic action necessary to strengthen such interface, a necessary step to jointly tackling the climate and biodiversity crisis.
15:00 – 17:00
Cocktail & socializing
Tuesday 27/08
Venue: Campus des Cordeliers, Sorbonne University (15 rue de l’école de médecine)
9:00 – 9:25
Walk-in, coffee
9:30 – 11:00
AI & Sustainability as an Industry Player
A Round table will discuss and endeavour to answer the following questions:
- How can AI be sustainable as a technique? What kind of AI should we design to ensure robustness, reliability, fairness and explainability? What kind of AI should we develop to avoid a too important energy consumption in terms of models, algorithms and data? Is frugality an asset that we should impose? Can we reduce the carbon footprint of AI?
- How can AI contribute to solve the biggest issues linked to climate change and resource limitation?
Speakers:
- Aurelie Boisbunon (Ericsson)
- Bruno Carron (Airbus)
- Vincent Feuillard (Renault)
- Patrick Perez (Kyutai)
- Piotr Sankowski (IDEAS NCBR)
11:00 – 11:30
Coffee break and move to industry fair in Réfectoire des Cordeliers11:30 – 13:30
Venue: Réfectoire des CordeliersIndustry fair + speed dating with industry partners (see schedule here)
13:30 – 15:00
Venue: Campus des CordeliersInnovation & Entrepreneurship session
Speakers:
- Arnaud de Moissac (DCBrain) – Gregor Von Dulong (Hasso-Plattner Institute) – How to validate your startup idea
- Rémi Bouzel (Qarnot Computing) – Thermodynamic challenges of HPC and AI
15:00 – 15:30
Coffee break and poster set up (Group 1: Last name’s starting letter: A - M)15:30 – 17:00
Poster session19:00
Gala Dinner at Salons de l’Hôtel des Arts et Métiers (9 bis, Avenue d’Iéna)Wednesday 28/08
Venue: Campus des Cordeliers, Sorbonne University (15 rue de l’école de médecine)
10:00 – 10:25
Walk-in, coffee
10:30 – 11:00
Claire Robin (ELLIS PhD student – Max-Planck Gesellschaft) – EarthNet: Bringing biogeoscience and machine learning together11:00 – 12:30
Keynote address
Rita Cucchiara (UNIMORE, ELLIS Unit Modena) - Learning, Unlearning, and Relearning
Abstract: The concept of "Learning, Unlearning, and Relearning" encapsulates a dynamic approach to knowledge acquisition and adaptation, crucial for both human cognition and artificial intelligence systems. Learning involves acquiring new information or skills, while unlearning refers to the process of discarding outdated or erroneous knowledge, and relearning entails reacquiring knowledge in a new or updated form. This process is not merely about forgetting but about restructuring understanding to accommodate new insights and perspectives. In machine learning and computer vision, unlearning presents unique challenges and opportunities. Machine unlearning refers to the ability to remove specific data from a model without degrading its overall performance. This capability is essential for addressing legal, ethical, and personalization issues. Techniques for unlearning in AI include the removal of data points, features, classes, or entire concepts from datasets. These methods aim to maintain model utility while ensuring the elimination of unwanted or harmful knowledge. Emerging methods in machine unlearning, such as weight filtering and low-rank class-wise unlearning, demonstrate promising results in maintaining model integrity while achieving desired unlearning outcomes. These techniques are evaluated through various metrics, including accuracy retention, activation distance, and Zero Retain Forgetting (ZRF) scores. Additionally, unlearning in multimodal spaces addresses the removal of sensitive or inappropriate content, enhancing the safety and ethical alignment of AI systems. The exploration of unlearning, especially in AI, paves the way for more robust, adaptable, and ethically sound models, capable of evolving in response to new data and societal norms. This ongoing research underscores the importance of unlearning and relearning as foundational components of both human and artificial intelligence development. During the seminar, some results carried out at UNIMORE within the EU project ELIAS will be presented and discussed.
12:30 – 14:00
Lunch14:00 – 15:30
Keynote address
David Rolnick (McGill University and Mila – Quebec AI Institute) - Tackling climate change with machine learning: An opportunity for application-driven innovation
Abstract: Machine learning is increasingly being called upon to help address climate change, from processing satellite imagery to modeling Earth systems. Such settings represent an important frontier for machine learning innovation, where traditional paradigms of large, general-purpose datasets and models often fall short. In this talk, we show how an application-driven paradigm for algorithm design can respond to problem-specific goals and incorporate relevant domain knowledge. We introduce novel techniques that leverage the structure of the problem (such as physical constraints and multi-modal self-supervision) to improve accuracy and usability across applications, including monitoring land use with remote sensing, designing chemical catalysts for the energy transition, and downscaling climate data.
15:30 – 16:00
Coffee break16:00 – 17:30
Keynote address
Caroline Therwath-Chavier (The Allyance) - AI²: why Artificial Intelligence should be All Inclusive
Abstract: In today’s rapidly evolving technological landscape, the transformative potential of Artificial Intelligence (AI) is undeniable. However, as AI continues to revolutionize various industries and aspects of our lives, questions of inclusivity and equity have come to the forefront for practitioners. This presentation delves into the imperative of making AI all-inclusive, ensuring that its benefits are accessible to all segments of society. We will explore the ethical, social, and economic implications of AI exclusivity, emphasizing the need for diverse representation in AI development and deployment. Through case studies and real-world examples, we will highlight the tangible benefits of inclusive AI, from enhancing decision-making processes to fostering innovation and creativity. Moreover, the presentation will address the challenges and barriers to achieving inclusivity in AI, including biases in algorithms, lack of diversity in AI teams, and digital divides. By identifying these obstacles, we can formulate strategies to overcome them and create a more equitable AI ecosystem. Ultimately, AI² advocates for a future where Artificial Intelligence is not only powerful and innovative but also inclusive and accessible to everyone. By embracing diversity and promoting inclusion, we can harness the full potential of AI to address pressing societal challenges and build a better world for all.
Thursday 29/08
Venue: Campus des Cordeliers, Sorbonne University (15 rue de l’école de médecine)
9:00 – 9:25
Walk-in, coffee
9:30 – 11:00
Keynote address
Gaël Varoquaux (INRIA Saclay) - AI from tabular data to healthcare and society
Abstract: I will talk about some lessons learned trying to develop AI for tabular data, central to all organizations, with the goal of being useful for healthcare and society. While much of the excitement is on machine learning, much of the work happens before and after. For tabular data, the bread and butter of every organization, extensive manual transformation or "data wrangling" is required before machine learning. I will discuss how we progressively rethought this process, eventually laying the ground of foundation models for tabular data. I will then discuss some considerations about machine learning in healthcare, and a broader picture on AI developments in society.
11:00 – 11:30
Coffee break11:30 – 12:30
Bethany Cagnol - Elevate your Scientific Communication: Key Principles for International Outreach
Abstract: Effective international scientific outreach hinges on understanding and engaging diverse audiences. This talk explores key strategies for success, including identifying target demographics and crafting clear, compelling messages that focus on the "What" and "Why" of the content while, in certain contexts, avoiding the "How." Emphasizing the psychological concept of the Illusion of Transparency, we highlight the challenges of harnessing clarity. Additionally, the cross-disciplinary advantages of scientific outreach efforts will be discussed.
12:30 – 13:30
Lunch & poster set-up (Group 2: Last name’s starting letter: N - Z)13:30 – 15:30
Poster session & Best Poster Award announcement15:45 – 18:00
Treasure Hunt in Quartier Latin (11 Pl. Marcelin Berthelot, 75231 Paris)
Participants will be grouped in 15 teams, see list here
Friday 30/08
Venue: Porte d’Orléans meeting point - 8:45am (firm)
8:45
Meeting point at Porte d’Orléans (look for Ekhine and Julien!)
9:00
Shuttle to Institut Polytechnique Paris9:30 – 10:00
Welcome Session at Institut Polytechnique Paris10:00 – 11:00
Tiphaine Viard (IPP) - AI as a Sociotechnical construct