AMLD Generative AI Workshops for the EPFL community
EPFL Campus, Lausanne, Switzerland
Tuesday, August 29th, 2023
To further the discussions from the main AMLD GenAI Conference, we’re bringing together another group of experts for a special session on themes in generative AI, including theory, science, and policy. Generative AI and foundation models promise to change the world as personal computers, the web and mobile computing did before them. Join us for an exciting day of presentations and discussions at this pivotal moment in the next technological revolution.
TRACK 1 – Technology of Foundation Models
Organized by Prof. Volkan Cevher (STI), Prof. Nicolas Flammarion (IC), Prof. Caglar Gulcehre (IC), Prof. Martin Jaggi (IC), Prof. Amir Zamir (IC), and Prof. Lenka Zdeborova (IC, SB)
This 3-fold workshop with world-leading experts will focus on the technology of recent foundation models such as large language models: How to build them – a systems and engineering perspective; How to understand them – towards a theory of foundation models; and how to integrate models across different data modalities.
TRACK 2 – AI Policy Track
Organized by Dr. Jan Kerschgens (CIS) and Prof. Tanja Käser (IC)
The realm of AI has emerged as a pivotal technology in the 21st century. Recent breakthroughs in training Large Language Models such as ChatGPT, built upon foundational models applicable to a diverse range of human activities, are already reshaping numerous facets of our society, promising substantial boosts to economic productivity (McKinsey). Our AI policy workshop on this day aims to delve into these transformative dynamics. Through illuminating presentations from experts in science, economy, and policy-making, we will explore how the forthcoming EU AI Act and its subsequent implementation via industrial standards will need to balance risk mitigation with the enablement of positive AI-driven contributions to our societies.
TRACK 3 – Foundation Models for Science
Organized by Prof. Maria Brbic (IC), Prof. Pascal Frossard (STI), Prof. Philippe Schwaller (SB), and Dr. Dorina Thanou (STI, SB, IC)
Foundation models – powerful AI models that have been trained on massive datasets and can be used to solve many different downstream tasks – have been revolutionizing AI recently. In this workshop, we will discuss recent advances, opportunities and risks of these models in scientific applications. The workshop will feature renowned speakers from the fields of biology, medicine, and chemistry, bringing diverse perspectives and expertise to explore the potential of foundation models for scientific breakthroughs in these fields.
Program – Day two workshops
Technology of Foundation Models
CM3 auditorium (live feed in CM1 when CM3 is at full capacity)
08:30 – 09:00 Welcome & registration
09:00 – 09:10 Introduction
09:10 – 09:45 Armand Joulin, « Democratizing Large Language Models »
09:45 – 10:20 Julien Launay, « Challenges in Training Frontier Language Models »
10:20 – 10:50 Coffee break
10:50 – 11:15 Ce Zhang, « Building RedPajama »
11:15 – 11:50 Tim Dettmers, « QLoRA: Efficient Finetuning of Quantized LLMs »
11:50 – 12:30 Panel discussion #1 with Tim Dettmers, Johnathan Frankle, Julien Launay and Ce Zhang.
12:30 – 14:00 Lunch break (on your own)
14:00 – 14:35 Ani Kembhavi, « Is this the beginning or is this the end for end-to-end vision models ? »
14:35 – 15:10 Neil Houlsby, « Advances in Visual Pretraining for LLMS »
15:10 – 15:40 Coffee break
15:40 – 16:15 Emily Webber, « Pretrain Foundation Models on Public Clouds »(joining remotely)
16:15 – 16:50 Fei Xia (joining remotely), « Low-level Embodied Intelligence with Foundation Models »
16:50 – 17:30 Panel discussion #2 with Armand Joulin, Neil Houlsby, Ani Kembhavi and Fei Xia.
17:30 Closing
AI Policy Track
CM2 auditorium
08:30 – 09:00 Welcome & registration
09:00 – 09:10 Introduction
09:10 – 09:30 Pierre Vandergheynst, « A Super Simple Introduction to AI and Generative AI »
09:30 – 09:50 Roger Dubach, « Set the course: Crafting a legal and regulatory landscape for AI »
09:50 – 10:10 Adina Dorthe, « AI in the machine tool industry: opportunities & challenges »
10:10 – 10:30 Ayisha Piotti, « AI Policy – a strategic opportunity »
10:30 – 11:00 Coffee break
11:00 – 12:00 Round table with Daniel Dobos, Adina Dorthe, Roger Dubach, Adam M. Gontarz (remote), Marcel Knecht, Ayisha Piotti, Sabine Süsstrunk and Pierre Vandergheynst.
Moderated by Jan Kerschgens
12:00 Closing
Foundation Models for Science
CM2 auditorium
13:30 – 14:00 Welcome & registration
14:00 – 14:05 Introduction
14:05 – 14:40 Pranav Rajpurkar, « The Generalist Medical AI Will See You Now »
14:40 – 15:15 Payel Das, « Promises (and Pitfalls) of Foundation Models for Science »
15:15 – 15:45 Coffee break
15:45 – 16:20 Andrew White, « Language is the Future of Chemistry »
16:20 – 16:55 Shekoofeh Azizi, « Exploring Foundation Models for Generalist Biomedical AI »
16:55 – 17:30 Round table with Shekoofeh Azizi, Payel Das, Pranav Rajpurkar, Nigam Shah and Andrew White.
17:30 Closing
Speakers
Program Organization
School of Computer and Communication Sciences
School of Computer and Communication Sciences
School of Engineering
School of Computer and Communication Sciences
School of Engineering
School of Computer and Communication Sciences
School of Computer and Communication Sciences
School of Computer and Communication Sciences
Center for Intelligent Systems
School of Computer and Communication Sciences
School of Life Sciences
School of Life Sciences
School of Computer and Communication Sciences
School of Basic Sciences
School of Engineering
School of Basic Sciences
School of Computer and Communication Sciences
School of Computer and Communication Sciences
School of Computer and Communication Sciences School of Basic Sciences
Event Organization
Directions
If you’re flying in, Genève-Cointrin is the nearest international airport. By train, it takes ~45 minutes to reach Lausanne. Zürich Airport is about ~2.5 hours away by train. To get information on train schedules, please see the Swiss Rail website.
To reach the CM1 and CM2 venue from the EPFL M1 metro station, please check the map below.