In March 2024, the tinyML Foundation organized one of the most successful and impactful events in its history. The subject was the rise of Generative AI on the Edge.

This first virtual forum brought together experts on edge AI, tinyML, foundation models, and generative AI, bridging crucial connections that are helping the community chart a new frontier for AI.

Since then, new innovations have appeared, new models have been released publicly, and both Hugging Face and GitHub are hosting an incredible amount of foundational models, paving the way for new use cases. The more the use cases are refined the better the opportunity to focus the perimeters of the Large Generative Model becomes concrete – thus offering new support to every business and person in everyday operations. Now more than ever, questions about how to deploy foundation models to the edge and what use cases become possible call to the machine learning community to act and share their latest progress.

Generative AI technology is at the heart of new applications yet to be invented, including a new wave of solutions that combine a large amount of smart sensors with LLMs to provide a better understanding of an environment or situation as well as autonomous systems applied to manufacturing via the deployment of humanoid robotics.

Therefore this “second edition” of the Generative AI at the Edge Forum aims to review the progress appearing in the state-of-the-art. It will witness even more the transformative impacts that foundation models can bring to the edge pushing hardware, software, tooling, applications, and AI design methodologies at the next level.

Two broad themes will be discussed:

• The current status and future outlook for the scenarios for foundation models and generative AI on resource-constrained energy efficient embedded devices, and their role within the edge-to-cloud continuum.
• The foundation model and generative AI tools, methodologies and techniques that can assist tinyML, edge AI, hardware and software embedded developers, including EDA tools, assisted labeling, and synthetic data generation.

A key goal is to inspire the community further by creating a unique workspace to favor the AI community to move forward through more collaborative work!