A glimpse on the EdgeAI project: Technologies for Optimized Embedded Processing

A glimpse on the EdgeAI project: Technologies for Optimized Embedded Processing Invited Paper Paola Busia University of Cagliari, DIEE Cagliari, Italy [email protected] Ovidiu Vermesan SINTEF Digital Oslo, Norway [email protected] ACM Reference Format: Paola Busia and Ovidiu Vermesan. 2025. A glimpse on the EdgeAI project: Technologies for Optimized Embedded Processing : Invited Paper. In 22nd ACM International Conference on Computing Frontiers (CF Companion ’25), May 28–30, 2025, Cagliari, Italy. ACM, New York, NY, USA, 2 pages. https: //doi.org/10.1145/3706594.3729340 1 AT-THE-EDGE AI: THE EDGEAI PROJECT At-the-edge Artificial Intelligence (AI) empowers Machine Learning (ML) and Deep Learning (DL) at the network’s periphery, closer to sensors and actuators, for localized data collection and processing, reducing latency, enhancing data privacy and security, and diminishing the need for cloud connectivity. However, this poses challenges related to the execution of a complex computing workload on resource-constrained platforms. Thus, it requires dealing with diverse technologies and optimizing energy usage. The EdgeAI project (https://edge-ai-tech.eu/), a joint effort of 42 partners, part of the Key Digital Technologies (KDT) Joint Undertaking (JU), aims to face such challenges, to play a pivotal role in Europe’s digital evolution towards smarter processing solutions at the edge. It focuses on creating fresh electronic parts and systems, refining processing setups, improving connectivity, and developing software, algorithms, and middle-layer technologies. 2 EDGEAI APPLICATIONS The main aim of EdgeAI is to advance solutions across various layers of AI technology, culminating in the creation of real-time performing multimodal edge AI implementations for diverse industrial sectors. The EdgeAI project partners work to demonstrate the applicability of the developed approaches in 20 demonstrators across five industrial value chains: •digital industry, We acknowledge the contribution of the whole Edge AI Consortium. "Edge AI Technologies for Optimised Performance Embedded Processing” is supported by the Chips Joint Undertaking and its members including top-up funding by Austria, Belgium, France, Greece, Italy, Latvia, Netherlands, and Norway under grant agreement No 101097300. Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the Chips Joint Undertaking. Neither the European Union nor the granting authority can be held responsible for them. Permission to make digital or hard copies of part or all of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for third-party components of this work must be honored. For all other uses, contact the owner/author(s). CF’ 25, May 28-30, 2025, Cagliari, Sardinia, Italy ©2025 Copyright held by the owner/author(s). ACM ISBN 979-8-4007-1393-4/2025/05 https://doi.org/10.1145/3706594.3729340 Micro edge Deep edge Meta edge sensors actuators cloud Embedded low-power microcontrollers, smart sensors Performant processors and microcontrollers, CPUs, GPUs, TPUs, and ASICs CPUs, GPUs, FPGAs Figure 1: EdgeAI vision of the computing continuum. •energy, •agri-food and beverage, •mobility, and •digital society, considering performance, security, trust, and energy efficiency demands inherently in each of these demonstrators. EdgeAI is designed to provide benefits across industrial sectors, to significantly contribute to the ubiquitous adoption of at-the-edge AI in society. 3 COMPUTING ASPECTS IN EDGEAI The EdgeAI vision spans across the the whole computing continuum, as represented in Figure 1, comprising: • the micro-edge (processing units in embedded microcontrollers, sensors, and actuators, etc.) • the deep-edge (processing units providing extended processing power, in gateways, mobile phones, programmable logic controllers, etc.) • the meta-edge (on-premises high-performance edge processing microservers combining different microcontrollers and processors for specific operations). To this aim, different research tasks focus on a wide scope of heterogeneous processing platforms, aiming to amalgamate different computing solutions such as central processing units, graphics processing units, tensor processing units, application-specific integrated circuits, neuromorphic processing units, system-on-chip (SoC). The EdgeAI strategic goals are realised through a set of objectives. These objectives reflect the way in which EdgeAI consortium delivers processing solutions for AI at the edge addressing the design stack and middleware across the target industrial sectors: • Objective 1 (O1) - Develop secure AI-based edge platforms for end-to-end hardware/software (HW/SW) solutions addressing the AI design stack and middleware (MW). • Objective 2 (O2) - Provide scalable edge AI-based energyefficient techniques, methods and frameworks supporting different OSs and hardware platforms. CF’ 25, May 28-30, 2025, Cagliari, Sardinia, Italy Busia et al. • Objective 3 (O3) - Advance multi-core SoC and SoM AI-based designs with hybrid architectures, embedded systems, and IoT devices designed for industrial environments. • Objective 4 (O4) - Integration of scalable and modular AI Co-design: hardware/software, algorithms, topologies into novel AI open architecture platforms. • Objective 5 (O5) - Implementation of reconfigurable AI-based architectures for increasing the re-use, updatability, upgradability, and service life of AI. • Objective 6 (O6) - Provide trustworthy and explainable edgeAI by design solutions with real-time operation capabilities and dynamic online learning. Publications deriving from the project are available at: https://zenodo.org/communities/edgeai_project/. The scope of the project presentation at the workshop will cover aspects related with computing techniques and methods, ranging from hardware to software, frameworks, and applications.