Edge AI technologies refer to a transformative AI approach that combines IoT and edge computing. In this approach, data processing and computation occur at the “edge,” closer to the data source and end users, rather than relying exclusively on centralized processing services. This paradigm shift enables various innovative capabilities and benefits, making it particularly significant in scenarios requiring real-time or near-real-time data processing and decision-making.
As generative AI continues to evolve and develop the capability to create innovative solutions across various domains, the integration of edge computing becomes increasingly significant. This workshop brings together experts and stakeholders to explore functional and non-functional requirements for deploying edge AI technologies effectively, considering the evolution of generative AI applications. The sessions will cover cutting-edge advancements, challenges, best practices, and cross-industry insights to facilitate robust, efficient, and scalable edge AI systems.
The functional requirements for edge AI systems topics address identifying and defining the core functionalities essential for deploying AI models at the edge, ensuring real-time data processing and decision-making capabilities, and exploring hardware and software design considerations for efficient AI inference at the edge.
Non-functional requirements discussions will emphasize system quality properties to ensure system dependability, robustness, reliability, security, availability, explainability, and interpretability in edge AI systems while ensuring scalability and adaptability to support diverse environments and workloads.
The workshop in “Charting the Future of Edge AI: Functional and Non-Functional Requirements in the Age of Generative AI” provided a comprehensive platform for stakeholders to exchange ideas, share experiences, and collaborate on advancing edge AI technologies. By addressing functional and non-functional requirements defined based on system engineering principles, participants gained valuable insights into deploying effective and responsible edge AI solutions using cross-disciplinary approaches for enhancing edge AI capabilities in the generative AI era.
The workshop combined presentations and panel discussions to allow the participants to share their insights, research findings, and best practices, facilitating a collaborative environment that promotes innovation in edge AI.
119 attendees from 77 institutions in 23 countries registered for the event that took place on 21 January 2025 at HiPEAC Conference in Barcelona, Spain and was organized by the Chips JU EdgeAI project in collaboration with HE dAIEDGE project.
Presentations:
- Edge AI Systems Functional and Non-Functional Requirements Engineering. Generative AI Future. Trends and Concepts, Ovidiu Vermesan, SINTEF, Norway
- Generative AI at the Edge for the Metaverse
- Challenges, Goals, and Trends for Smart Edge Agents,
- A Touch-free Gesture Recognition System Powered by Edge AI,
- Panel discussion Edge AI Technology
- Low-Power Near-Sensor Intelligence: Advancing Micro-Edge Computing Beyond CNNs.,
- Edge AI For Computer Vision: Challenges, Technological Advancements, and Future Trends
- Designing Efficient LLM Accelerators for Edge Devices, Jose Cano Reyes
- Edge AI Future Trends and Strategic Research Agenda
- A Versatile and Secure AI Platform with Reconfigurable RISC-V Based Parallel Processing
- Edge AI research topics and future trends
