Workshop “Edge AI meets swarm intelligence”
18 September 2024, Dubrovnik, Croatia
On this page:
Workshop overview
Venue
Agenda
Committees
Important dates
Topics of interest
As part of the SmartEdge & AIoTwin Summer School, the “Edge AI meets Swarm Intelligence” technical workshop was held on 18 September 2024.
Among various initiatives and projects linked to The European Cloud, Edge & IoT Continuum and EU-US alliance on swarming computing and distributed intelligence, two Horizon Europe projects – SmartEdge and AIoTwin – are jointly organizing the ‘Edge AI meets Swarm Intelligence’ workshop to disseminate on-going research outcomes not only of these two projects, but also of related projects such as SMARTY, Clever and MLSysOps.
Workshop overview
Edge AI represents a novel computing paradigm designed to facilitate local data storage and processing, with AI algorithms enabling data treatment directly at the edge of the network. This approach aims to bring intelligence to the end-devices, facilitating real-time decision-making and empowering devices to operate autonomously, reducing reliance on external cloud services.
One of the key topics of interest is to develop low-code programming tool chains or platforms for edge intelligence to enable swarm computing paradigms. The tools will reduce the efforts of building smart systems requiring a collective of heterogeneous devices, sensors, vehicles and robots to collaborate towards a common goal. Such a solution aims at enabling Swarm Intelligence concepts, which are a form of AI that mimics the collective behavior of decentralized, self-organized systems, like those observed in nature.
Researchers working on the two projects share similar research interests in terms of enabling AI at the edge and dealing with the heterogeneity of such environments. Heterogeneity can refer here to differences in network capabilities, processing power, software stacks, APIs, and security protocols across the Cloud-Edge Continuum infrastructure, e.g. continuum reference architecture proposed by EUCloudEdgeIoT.
The primary goal of this workshop was to foster collaboration and the exchange of ideas among researchers and stakeholders. The workshop provided a platform for participants to share their experiences, best practices, case studies, and to identify emerging research areas and potential solutions to existing challenges.
Venue
University of Zagreb Centre for Advanced Academic Studies (CAAS)
Don Frana Bulica 4,
20000 Dubrovnik, Croatia
Agenda
| Time | Topic |
|---|---|
| 08:30-08:50 | Morning coffee |
| 08:50-09:00 | Opening ceremony |
| 09:00-10:30 | “Techniques for safe and highly available cloud applications” Keynote by Carla Ferreira (TarDIS) |
| 10:30-11:00 | Coffee break + posters session |
| 11:00-11:30 | Paper: Swarm Intelligence for Green Data Orchestration – A Vision for Energy-Efficient Computing in the Edge Continuum Authors: Aidan O Mahony, Fred Buining, Péter Forgács and Melanie Schranz |
| 11:30-12:00 | Paper: Performance Evaluation of ROS2-DDS middleware implementations facilitating Cooperative Driving in Autonomous Vehicle Authors: Sumit Paul, Danh Le-Phuoc and Manfred Hauswirth |
| 12:00-12:30 | Paper: Experimental comparison of graph-based approximate nearest neighbor search algorithms on edge devices Authors: Ali Ganbarov, Jicheng Yuan, Anh Le-Tuan, Manfred Hauswirth and Danh Le-Phuoc |
| 12:30-14:00 | Lunch break + posters session |
| 14:00-14:30 | Paper: Not All RDF is Created Equal: Investigating RDF Load Times on Resource-Constrained Devices Authors: Piotr Sowiński, Anh Le-Tuan, Paweł Szmeja and Maria Ganzha Speaker: Piotr Sowiński |
| 14:30-15:00 | Paper: Dynamic Knowledge Graph Based Swarm Networks Authors: Xuanchi Guo, Anh Le-Tuan and Danh Le Phuoc |
| 15:00-15:20 | Coffee break + poster sessions |
| 15:20-15:50 | Paper: Cost-Effective and Efficient Scene Understanding for Traffic Cameras Authors: Thuy-Duong Tran and Trung-Kien Tran |
| 15:50-16:20 | Paper: Co-Learning: Towards Semi-Supervised Object Detection with Road-side Cameras Authors: Jicheng Yuan, Anh Le-Tuan, Ali Ganbarov, Manfred Hauswirth and Danh Le-Phuoc |
| 16:20-16:40 | Coffee break + poster sessions |
| 16:40-17:10 | Paper: Towards Complete Input Representations for Vehicle Trajectory Prediction Models Authors: Marcel Milich, Trung-Kien Tran and Maximilian Naumann Speaker: Marcel Milich |
| 17:10-17:40 | Paper: A comparison of extended object tracking with multi-modal sensors in indoor environment Authors: Jiangtao Shuai, Martin Baerveldt, Manh Nguyen-Duc, Anh Le-Tuan, Manfred Hauswirth and Danh Le-Phuoc |
| 17:40-18:00 | Closing the workshop |
Committees
Workshop chairs
- Ivana Podnar Zarko (University of Zagreb, Croatia)
- Philippe Cudre-Mauroux (University of Fribourg, Switzerland)
- Trung-Kien Tran (Bosch, Germany)
- Danh Le-Phuoc (TU Berlin, Germany)
Publicity and proceeding chairs
- Ilir Murturi (TU Wien, Austria)
- Anh Le-Tuan (TU Berlin, Germany)
Technical program committee
- Alessio Carenini (CEFRIEL, Italy)
- Mario Scrocca (CEFRIEL, Italy)
- Kari Koskinen (CONVEQS, Finland)
- Mehrdad (CONVEQS, Finland)
- Alan Cueva Mora (DELL, Ireland)
- Christopher Krauß (FHG, Germany)
- Louay Bassbouss (FHG, Germany)
- Philippe Cudre-Mauroux (FRIB, Switzerland)
- Jean-Paul Calbimonte (HESSO – Switzerland)
- Davide Calvaresi (HESSO, Switzerland)
- JJ Vegas Olmos (NVIDIA, Denmark)
- Peng Qian (Oxford University, UK)
- Changgang Zheng (Oxford University, UK)
- Lodovico Giaretta (RISE, Sweden)
- Sarunas Girdzijauskas (RISE, Sweden)
- Kirill Dorofeev (Siemens, Germany)
- Damien Foucard (TU Berlin, Germany)
- Mario Kušek (University of Zagreb, Croatia)
- Gordan Ježić (University of Zagreb, Croatia)
- Marie-Claire Forgue (W3C/ERCIM)
- Michalis Mountantonakis (W3C/ERCIM)
- Panagiotis Papadakos (W3C/ERCIM)
- Filippo Cugini (CNIT, Italy)
- Michelangelo Guaitolini (CNIT, Italy)
Important dates
The call for papers opened on May, 1st and the submission deadline was 29 June 2024.
- Open for Submission: 1st , May
- Abstract Submission: 29th, June
- Full Submission: 6th, July
- Notifications: 15th, August
- Camera-ready: 23rd, August
Topics of interest
We welcomed submissions of original research, work-in-progress, and proof of concepts describing original ideas and presenting new directions in edge and swarm AI. We invited submissions related to the design, development, and evaluation of architectures, technologies, and applications for “Edge AI” and “Swarm Intelligence”, including but not limited to:
- Network optimization and interoperable protocols for Edge AI and Swarm Intelligence
- Novel system architectures and hardware designs in Edge AI and Swarm Intelligence
- Network security, data privacy, confidence, and trust in Edge AI and Swarm Intelligence
- AI-enabled resource allocation, federated learning, and swarm intelligence
- Next generation smart use-cases enabled by Edge AI and Swarm Intelligence
- Intelligence in distributed computing continuum systems
- Novel results in embedded AI, in-network computing or heterogeneous computing
- Methods, resources, and experimental findings in autonomous systems