Objective: Enable low-latency, energy-efficient, privacy-aware intelligence on sensors and edge devices.

Novelty: Establish design principles for dynamically optimizing models, representations, and task partitioning under resource and network constraints.

Methods: Model compression (distillation/quantization), edge–cloud collaboration, federated learning, and uncertainty-aware inference/control.

Results: Implementable prototypes and evaluations via simulations and small-scale demonstrations.

Next: Advance toward network-aware Edge AI that adapts to changing network conditions and deployments.

Objective: Design scalable communication and control foundations for networked infrastructures with low latency and high reliability.

Novelty: Jointly consider communication (QoS/QoE) with control/optimization to achieve resilient operation under failures, congestion, and dynamics.

Methods: Distributed optimization, reinforcement learning, GNN-based control, traffic engineering, network slicing, and service function chaining (SFC) placement.

Results: Quantitative evaluations (latency/reliability/cost) under network constraints and systematic policy design.

Next: Integrate with the Edge AI foundation for unified architectures that optimize task partitioning and representations.

Objective: Build dependable smart systems for AgeTech (monitoring/assistance) and field IoT under real constraints.

Novelty: Map heterogeneous sensors into a shared semantic representation and handle missingness, non-IID, and privacy constraints robustly.

Methods: Semantic representation learning, privacy-preserving learning, real-time edge inference, and adaptive operation under varying network quality.

Results: Validation mainly through open datasets/simulations, complemented by small-scale prototypes.

Next: Bring resilience and distributed control principles into deployments that remain robust during disasters and congestion.