EAI4IA was a multidisciplinary workshop addressing AI technologies at the edge, giving an overview of silicon-born AI components to advance Moore?s law and accelerate the adoption of AI-based edge processing in different industries. Technological advances regarding edge AI embedded in electronic components and systems were presented, and the challenges of introducing embedded AI technologies for digitising industrial sectors were discussed. It was co-organised by three large-scale ECSEL JU projects: AI4DI, ANDANTE and TEMPO. EAI4IA comprised technical presentations, keynotes and panel discussions focusing on industrial-edge AI hardware, software and AI.
Nine papers from AI4DI were accepted for presentation at the workshop. All presentations are listed below. The workshop publications are bundled in an open-access book: Industrial Artificial Intelligence Technologies and Applications (riverpublishers.com).
EAI4IA workshop agenda
Session: Trustworthy, Dependable AI for Digitising Industry
Frank Badstübner (Infineon Technologies AG, DE) Artificial Intelligence for Digitizing Industry (AI4DI)
Björn Debaillie (imec, BE) Technologies and hardware for neuromorphic computing (TEMPO)
Mario Diaz Nava (ST Microelectronics, FR) AI for New Devices and Technologies at the Edge (ANDANTE)
Session: Trustworthy, Dependable AI for Digitising Industry
David Kaufmann (AVL LIST, AT) A framework for integrating automated diagnosis into simulation
Lucas Mohimont (Université de Reims Champagne Ardenne, FR) AI-Driven Strategies to Implement a Grapevine Downy Mildew Warning System
Dinu Purice (Cognition Factory GmbH, DE) An End-to-End AI-based Automated Process for Semiconductor Device Parameter Extraction
Session: Embedded AI for semiconductor manufacturing
Annika Wandesleben (Fraunhofer IMPS, DE) Feasibility of wafer exchange for European Edge AI pilot lines
Saad Al-Baddai (Infineon Technologies AG, DE) Failure detection in semiconductor package
Lan Liu (Know-Center GmbH, AT) S2ORC-SemiCause: Annotating and Analysing Causality in the Semiconductor Domain
Dmitry Morits (VTT, FI) AI machine vision system for wafer defect detection
Session: AI model development workflow and HW target platforms deployment
Ovidiu Vermesan (SINTEF, NO) Embedded Edge Intelligent Processing for End-To-End Predictive Maintenance in Industrial Applications
Simon Narduzzi (CSEM, CH) Deploying a Convolutional Neural Network on Edge MCU and Neuromorphic Hardware Platforms
Ruben Prokscha (OTH Amberg-Weiden, DE) Efficient Edge Deployment Demonstrated on YOLOv5 and Coral Edge TPU
Preetha Vijayan (imec & TU Delft, NL) Temporal Delta Layer: Exploiting Temporal Sparsity in Deep Neural Networks for Time-Series
Data Session: Strategic Vision and Road mapping
Yves Gigase (KDT, BE) The KDT Joint Undertaking. A tutorial for the next generation of policy-makers and decision-takers
Dylan Muir (SynSense, CH) Disruptive technology development through SMEs. A view from Neuromorphic edge AI
Gerhard Friedrich (Universität Klagenfurt, AT) Strategic vision and road mapping: Industrial success stories of answer set / constraint programming. What?s still open?
Session: Explainable AI in the Embedded Electronics Industry Environment
Georgi Kuzmanov (KDT, BE) Key Digital Technologies JU. ECSEL and KDT Industrial AI Area Session: Verification, Validation and Benchmarking of AI Systems and Technologies
Loreto Mateu (Fraunhofer IIS, DE) Benchmarking Neuromorphic Computing for Inference
Amirreza Yousefzadeh (imec, NL) Benchmarking the Epiphany processor as a reference neuromorphic architecture
Franz Wotawa (Graz University of Technology, AT) On the Verification of Diagnosis Models
Session: Emerging AI Technologies in Industrial Applications
Reiner John (AVL, AU) The Future of Decision Making in SAR Quantum technology & distributed intelligence