This Research Topic explores the integration of large language models (LLMs) with edge computing to optimize generative AI for resource-constrained Internet of Things (IoT) environments. While LLMs excel in natural language understanding and generation, their significant computational and memory demands pose challenges for deployment on edge devices and IoT systems with limited resources. Addressing this gap is crucial for enabling real-time, low-latency, and privacy-preserving AI applications in decentralized environments. This topic seeks innovative strategies to adapt, compress, and efficiently deploy LLMs in edge computing, fostering breakthroughs in areas such as model optimization, distributed training, and secure inference. By advancing the deployment of LLMs in resource-constrained settings, this research aims to unlock new possibilities for IoT applications in domains such as healthcare, smart cities, and industrial automation, paving the way for transformative AI solutions at the edge.

The rapid evolution of LLMs has unlocked unprecedented capabilities in natural language understanding and generation. However, their deployment in resource-constrained environments, such as Edge and IoT devices, remains a significant challenge due to high computational demands, latency, and energy consumption. This research topic aims to explore innovative methodologies to adapt and optimize LLMs for edge computing and IoT systems, enabling real-time and efficient AI-powered decision-making in decentralized settings. To achieve this, the research topic will focus on developing lightweight, distributed, energy-efficient model architectures, model compression techniques, and federated learning approaches that ensure privacy and scalability. Furthermore, it will investigate the integration of multimodal AI capabilities for IoT, addressing the challenges of processing heterogeneous data streams in constrained environments. By bridging the gap between the computational needs of LLMs and the limitations of edge devices, this research topic aims to unlock new applications, from smart healthcare and autonomous vehicles to personalized edge AI solutions, advancing both the field of AI and its real-world adoption.

This Research Topic focuses on advancing the deployment of LLMs in Edge and IoT systems, addressing their integration into resource-constrained environments. We invite contributions that explore the following themes:

Efficient Model Design & Optimization
1.1. Quantization, pruning, knowledge distillation, and low-rank adaptation for reducing LLM complexity.
1.2. Novel architectures tailored for edge-specific tasks.

Compression & Efficient Storage
2.1. Weight compression, sparsity techniques, and efficient storage for edge AI.
2.2. Optimized retrieval of LLM knowledge for low-power devices.

Federated Learning with LLM
3.1. Privacy-preserving training approaches for decentralized IoT networks.
3.2. Strategies for minimizing communication overhead during model updates.

Energy-Efficient LLMs
4.1. Designing low-power inference methods for IoT devices.
4.2. Balancing model performance with energy consumption.

Applications and Case Studies
5.1. Multimodal LLMs for IoT data (e.g., vision, audio, and sensor data).
5.2. Use cases in healthcare, smart cities, and autonomous systems.

Security and Privacy
6.1 Safeguards against adversarial attacks and secure communication during training.

Submissions may include original research, reviews, case studies, and technical notes, with a focus on practical implementations and real-world challenges.

Are you interested in contributing a manuscript to this Research Topic?

Edge-Aware LLMs: Scaling Generative AI to Edge and IoT Environments