IEEE Computational Intelligence Society

 Motivation

The rapid evolution of artificial intelligence (AI), particularly deep learning (DL), foundation models, and generative AI, is transforming how intelligent applications and services are developed and deployed. Meanwhile, the proliferation of Internet of Things (IoT) devices, mobile computing, intelligent sensors, autonomous systems, and next-generation communication networks is generating unprecedented volumes of data at the network edge. Processing these data close to their sources offers significant opportunities for real-time intelligence, enhanced responsiveness, improved privacy, and reduced dependence on centralised cloud infrastructure.

However, bringing increasingly sophisticated AI capabilities to the edge presents substantial challenges. Modern AI models, particularly large language models (LLMs), multimodal foundation models, and generative models, demand considerable computational resources, memory, energy, and communication bandwidth. These requirements are often beyond the capabilities of resource-constrained edge devices. Furthermore, the heterogeneity, mobility, and dynamic operating conditions of edge environments introduce additional challenges in resource management, distributed learning, system orchestration, security, privacy, and reliability.

Deep Edge Intelligence (DEI) represents an emerging research paradigm that integrates AI, distributed and edge computing, communication networks, and IoT to enable intelligent, adaptive, and trustworthy services across interconnected devices and distributed computing infrastructures. Going beyond conventional on-device AI, DEI encompasses the collaborative development, deployment, execution, and evolution of intelligent models and services across the entire edge–cloud continuum, from resource-constrained devices and edge nodes to fog infrastructure and cloud data centres

Recent advances in foundation models, generative AI, autonomous AI agents, federated learning, and heterogeneous computing are creating new opportunities to advance DEI towards more capable, collaborative, and autonomous intelligent systems. These developments also introduce new research challenges in efficient model adaptation, distributed inference, multi-agent coordination, energy-aware computing, trustworthy AI, and resilient system operation.

DEI aims to make advanced AI capabilities accessible and sustainable across diverse environments, enabling intelligent services for individuals, organisations, and industries at any place and at any time. A position paper introducing the DEI vision, architecture, enabling technologies, and research challenges can be found here.

Goals

The goal of this task force is to establish, advance, and promote Deep Edge Intelligence as an interdisciplinary research direction, bringing together researchers and practitioners from AI, machine learning, distributed and edge computing, communication networks, IoT, and related disciplines.

The task force will provide an international forum for academic, industrial, and government stakeholders to exchange research ideas, identify emerging challenges, develop innovative solutions, and foster collaborations in DEI. It will promote fundamental research, technological innovation, and real-world applications through joint research activities, technical discussions, workshops, special sessions, tutorials, publications, and industry engagement.

In particular, the task force aims to:

• Advance the foundations, architectures, algorithms, and systems of Deep Edge Intelligence.

• Promote efficient and scalable deployment of foundation models, generative AI, and intelligent agents across heterogeneous edge–cloud environments.

• Develop collaborative and distributed intelligence through federated learning, decentralised learning, and distributed AI inference.

• Address the efficiency, security, privacy, trustworthiness, reliability, and sustainability challenges of DEI.

• Foster interdisciplinary research and collaboration between academia, industry, and government.

• Accelerate the translation of DEI research into practical applications and real-world intelligent services.

Scope

The scope of this task force covers, but is not limited to, the following research topics:

1. Deep Edge Intelligence Architectures and Systems

- Architectures, frameworks, and platforms for DEI.

- Intelligent edge–fog–cloud continuum and distributed AI infrastructures.

- Heterogeneous and resource-constrained computing architectures.

- Edge-native AI services, microservices, and serverless intelligence.

- 5G/6G-enabled edge intelligence and integrated communication, computing, and sensing.

2. Foundation Models and Generative AI at the Edge

- Edge-efficient large language models and multimodal foundation models.

- Model compression, pruning, quantisation, distillation, and efficient adaptation.

- On-device training, fine-tuning, and inference of foundation models.

- Retrieval-augmented generation and knowledge-enhanced edge intelligence.

- Distributed and collaborative inference of generative AI models.

- Privacy-preserving and personalised foundation models for edge applications.

3. Distributed, Federated, and Collaborative Intelligence

- Federated, split, decentralised, and distributed machine learning.

- Collaborative learning and inference across heterogeneous edge devices.

- Communication-efficient and resource-aware distributed learning.

- Personalised and continual learning in dynamic edge environments.

- Multi-agent learning and collaborative intelligence.

- Cross-device and cross-domain knowledge transfer and model adaptation.

4. Agentic and Autonomous Edge Intelligence

- AI agents and multi-agent systems at the edge.

- Agentic AI for distributed decision-making and autonomous services.

- Collaborative planning, reasoning, and execution across edge–cloud systems.

- Intelligent orchestration and autonomous management of edge resources.

- Trustworthy, secure, and resource-efficient autonomous agents.

5. Resource and Performance Optimisation

- Energy, computation, memory, storage, and communication optimisation.

- Resource-aware AI model selection, scheduling, and placement.

- Adaptive workload distribution and task offloading.

- Efficient training, inference, caching, and model serving.

- Hardware–software co-design and AI acceleration.

- Energy-efficient and carbon-aware DEI systems.

6. Security, Privacy, and Trustworthy Edge Intelligence

- Security and privacy protection for distributed AI and edge systems.

- Adversarial machine learning and robust edge intelligence.

- Trustworthy, explainable, and responsible AI.

- Secure and privacy-preserving federated learning.

- AI model integrity, provenance, and intellectual property protection.

- Verification, validation, testing, and assurance of intelligent edge systems.

- Resilience against malicious attacks, failures, and unreliable data.

7. Intelligent Edge Networking and Communications

- AI-driven network management and optimisation.

- Joint optimisation of communication, computation, and sensing.

- Intelligent resource allocation in 5G/6G and beyond networks.

- Distributed intelligence for vehicular, mobile, and wireless networks.

- Edge intelligence for network slicing and adaptive service provisioning.

8. Applications and Emerging Technologies

- DEI for Industry 4.0 and smart manufacturing.

- Autonomous driving, intelligent transportation, and connected vehicles.

- Smart grids, renewable energy systems, and the Internet of Energy.

- Smart cities, intelligent healthcare, and ambient intelligence.

- Networked robotics, drones, and autonomous systems.

- Digital twins and cyber-physical systems.

- Environmental monitoring, precision agriculture, and sustainable infrastructure.

- Other emerging applications enabled by distributed and edge intelligence