• Corpus ID: 244773241

Roadmap for Edge AI: A Dagstuhl Perspective

  title={Roadmap for Edge AI: A Dagstuhl Perspective},
  author={Aaron Yi Ding and Ella Peltonen and Tobias Meuser and Atakan Aral and Christian Becker and Schahram Dustdar and Thomas Hiessl and Dieter Kranzlmuller and Madhusanka Liyanage and Setareh Magshudi and Nitinder Mohan and J{\"o}rg Ott and Jan S. Rellermeyer and Stefan Schulte and Henning Schulzrinne and G{\"u}rkan Solmaz and Sasu Tarkoma and Blesson Varghese and Lars C. Wolf},
Based on the collective input of Dagstuhl Seminar (21342), this paper presents a comprehensive discussion on AI methods and capabilities in the context of edge computing, referred as Edge AI. In a nutshell, we envision Edge AI to provide adaptation for data-driven applications, enhance network and radio access, and allow the creation, optimisation, and deployment of distributed AI/ML pipelines with given quality of experience, trust, security and privacy targets. The Edge AI community… 

Figures from this paper

Oakestra white paper: An Orchestrator for Edge Computing
This work proposes a novel hierarchical orchestration framework specifically designed for supporting service operation over edge infrastructures that can consolidate multiple infrastructure operators and absorb dynamic variations at the edge.
Evolving 5G: ANIARA, an edge-cloud perspective
ANIARA attempts to enhance edge architectures for smart manufacturing and cities by providing en-ablers and solutions for services in the domains of smart cities and manufacturing deployed and operated at the network edge(s).
Autonomy and Intelligence in the Computing Continuum: Challenges, Enablers, and Future Directions for Orchestration
It is claimed that to support the constantly growing requirements of intelligent applications in the device-edge-cloud computing continuum, resource orchestration needs to embrace edge AI and emphasize local autonomy and intelligence.
Towards Trustworthy Edge Intelligence: Insights from Voice-Activated Services
This paper examines requirements for trustworthy Edge Intelligence in a concrete application scenario of voice-activated services and proposes a unified framing for trustworthyEdge Intelligence that jointly considers trustworthiness attributes of AI and the IoT.
Smart City Intersections: Intelligence Nodes for Future Metropolises
High-bandwidth, low- latency applications such as privacy preservation, cloud-connected vehicles, a real-time ”radar-screen”, traffic management, and monitoring of pedestrian behavior during pandemics are described.
A Roadmap To Post-Moore Era for Distributed Systems
  • Vincenzo De Maio, Atakan Aral, I. Brandić
  • Computer Science
    Proceedings of the 2022 Workshop on Advanced tools, programming languages, and PLatforms for Implementing and Evaluating algorithms for Distributed systems
  • 2022
The implications of the post-Moore era for distributed systems are discussed, where one expects the coexistence of multiple types of architectures specialized for different types of computation.


6G White Paper on Edge Intelligence
This white paper envision a transition from Internet of Things to Intelligent Internet of Intelligent Things and provide a roadmap for development of 6G Intelligent Edge, along with security, privacy, pricing, and end-user aspects.
Pruning Edge Research with Latency Shears
This paper performs extensive client-to-cloud measurements using RIPE Atlas, and shows that latency reduction as motivation for edge is not as persuasive as once believed; for most applications the cloud is already 'close enough' for majority of the world's population.
Machine Learning Systems in the IoT: Trustworthiness Trade-offs for Edge Intelligence
  • Wiebke Toussaint, A. Ding
  • Computer Science
    2020 IEEE Second International Conference on Cognitive Machine Intelligence (CogMI)
  • 2020
This paper analyzes the tradeoffs by covering the latest developments on scaling and distributing ML across cloud, edge, and IoT devices and position machine learning systems as a component of the IoT, and edge intelligence as a socio-technical system.
Computational Models of Human Decision-Making with Application to the Internet of Everything
A solution for a task offloading problem in fog computing and the implications of including humans in the loop are analyzed and a brief review of computational models of human decision-making is provided.
Revisiting the Arguments for Edge Computing Research
This article argues that low latency, high bandwidth, device proliferation, sustainable digital infrastructure, and data privacy and sovereignty continue to motivate the need for edge computing
Distributed Task Management in Cyber-Physical Systems: How to Cooperate Under Uncertainty?
It is proved that in the task allocation game, the strong sequential core is equivalent to Walrasian equilibrium under state uncertainty; consequently, it can be implemented by using the Walras' tatonnement process.
Data Poisoning Attacks Against Federated Learning Systems
This paper studies targeted data poisoning attacks against FL systems in which a malicious subset of the participants aim to poison the global model by sending model updates derived from mislabeled data, and proposes a defense strategy that can help identify malicious participants in FL to circumvent poisoning attacks.
Enabling Joint Communication and Radar Sensing in Mobile Networks—A Survey
A broad picture of the motivation, methodologies, challenges, and research opportunities of realizing perceptive mobile network is presented, by providing a comprehensive survey for systems and technologies developed mainly in the last ten years.
Characterising the Role of Pre-Processing Parameters in Audio-based Embedded Machine Learning
It is shown that pre-processing parameter tuning is a remarkable tool that model developers can adopt to trade-off between a model's accuracy, fairness, and system efficiency, as well as to make an embedded ML model resilient to unseen deployment conditions.
Participation Behavior and Social Welfare in Repeated Task Allocations
Task allocation problems have focused on achieving one-shot optimality. In practice, many task allocation problems are of repeated nature, where the allocation outcome of previous rounds may