The proliferation of end devices has led to a distributed computing paradigm, wherein on-device machine learning models continuously process diverse data generated by these devices. The dynamic nature of this data, characterized by continuous changes or data drift, poses significant challenges for on-device models. To address this issue, continual learning (CL) is proposed, enabling machine learning models to incrementally update their knowledge and mitigate catastrophic forgetting. However, the traditional centralized approach to CL is unsuitable for end devices due to privacy and data volume concerns. In this context, federated continual learning (FCL) emerges as a promising solution, preserving user data locally while enhancing models through collaborative updates. Aiming at the challenges of limited storage resources for CL, poor autonomy in task shift detection, and difficulty in coping with new adversarial tasks in FCL scenario, we propose a novel FCL framework named SacFL. SacFL employs an Encoder-Decoder architecture to separate task-robust and task-sensitive components, significantly reducing storage demands by retaining lightweight task-sensitive components for resource-constrained end devices. Moreover, $\rm{SacFL}$ leverages contrastive learning to introduce an autonomous data shift detection mechanism, enabling it to discern whether a new task has emerged and whether it is a benign task. This capability ultimately allows the device to autonomously trigger CL or attack defense strategy without additional information, which is more practical for end devices. Comprehensive experiments conducted on multiple text and image datasets, such as Cifar100 and THUCNews, have validated the effectiveness of $\rm{SacFL}$ in both class-incremental and domain-incremental scenarios. Furthermore, a demo system has been developed to verify its practicality.

Fine-tuning plays a crucial role in enabling pre-trained LLMs to evolve from general language comprehension to task-specific expertise. To preserve user data privacy, federated fine-tuning is often employed and has emerged as the de facto paradigm. However, federated fine-tuning is prohibitively inefficient due to the tension between LLM complexity and the resource constraint of end devices, incurring unaffordable fine-tuning overhead. Existing literature primarily utilizes parameter-efficient fine-tuning techniques to mitigate communication costs, yet computational and memory burdens continue to pose significant challenges for developers. This work proposes DropPEFT, an innovative federated PEFT framework that employs a novel stochastic transformer layer dropout method, enabling devices to deactivate a considerable fraction of LLMs layers during training, thereby eliminating the associated computational load and memory footprint. In DropPEFT, a key challenge is the proper configuration of dropout ratios for layers, as overhead and training performance are highly sensitive to this setting. To address this challenge, we adaptively assign optimal dropout-ratio configurations to devices through an exploration-exploitation strategy, achieving efficient and effective fine-tuning. Extensive experiments show that DropPEFT can achieve a 1.3-6.3\times speedup in model convergence and a 40%-67% reduction in memory footprint compared to state-of-the-art methods.

The proliferation of IoT devices, AI agents, and robotics has redefined the nature of workloads in modern computing systems. With the emergence of optimized AI models and ongoing hardware advancements, most smart devices including smartphones, PCs, and IoT devices are already capable of running narrow AI models efficiently. While upcoming device upgrades will further enhance AI capabilities, current devices are sufficient for handling most inference workloads, making a device-first approach not only feasible but highly relevant for agentic workflows [2], [3]. These workloads are often Pareto-distributed [4], [5], [6], [7], [8], [9], [10] where a small percentage of high-resource tasks dominate computational resources, while most tasks are lightweight. Centralized cloud systems, originally designed for web browsing and app-based transactions, struggle to meet the demands of dynamic, context-aware applications. This paper explores the implications of HEC, which can process tasks locally on end devices when possible and offloads only high-resource tasks to the cloud or dedicated cloud gateways. To provide a comprehensive view, we analyze both traditional workloads which reflect typical smart devices with less intelligence and agentic workloads emerging in AI-driven systems like autonomous vehicles and robotics.

The rise of End-Edge-Cloud Collaboration (EECC) offers a promising paradigm for Artificial Intelligence (AI) model training across end devices, edge servers, and cloud data centers, providing enhanced reliability and reduced latency. Hierarchical Federated Learning (HFL) can benefit from this paradigm by enabling multi-tier model aggregation across distributed computing nodes. However, the potential of HFL is significantly constrained by the inherent heterogeneity and dynamic characteristics of EECC environments. Specifically, the uniform model structure bounded by the least powerful end device across all computing nodes imposes a performance bottleneck. Meanwhile, coupled heterogeneity in data distributions and resource capabilities across tiers disrupts hierarchical knowledge transfer, leading to biased updates and degraded performance. Furthermore, the mobility and fluctuating connectivity of computing nodes in EECC environments introduce complexities in dynamic node migration, further compromising the robustness of the training process. To address multiple challenges within a unified framework, we propose End-Edge-Cloud Federated Learning with Self-Rectified Knowledge Agglomeration (FedEEC), which is a novel EECC-empowered FL framework that allows the trained models from end, edge, to cloud to grow larger in size and stronger in generalization ability. FedEEC introduces two key innovations: (1) Bridge Sample Based Online Distillation Protocol (BSBODP), which enables knowledge transfer between neighboring nodes through generated bridge samples, and (2) Self-Knowledge Rectification (SKR), which refines the transferred knowledge to prevent suboptimal cloud model optimization. The proposed framework effectively handles both cross-tier resource heterogeneity and effective knowledge transfer between neighboring nodes, while satisfying the migration-resilient requirements of EECC.

Accurate and timely prediction of heavy rainfall events is crucial for effective flood risk management and disaster preparedness. By monitoring, analysing, and evaluating rainfall data at a local level, it is not only possible to take effective actions to prevent any severe climate variation but also to improve the planning of surface and underground hydrological resources. However, developing countries often lack the weather stations to collect data continuously due to the high cost of installation and maintenance. In light of this, the contribution of the present paper is twofold: first, we propose a low-cost IoT system for automatic recording, monitoring, and prediction of rainfall in rural regions. Second, we propose a novel approach to regional heavy rainfall prediction by implementing graph neural networks (GNNs), which are particularly well-suited for capturing the complex spatial dependencies inherent in rainfall patterns. The proposed approach was tested using a historical dataset spanning 72 months, with daily measurements, and experimental results demonstrated the effectiveness of the proposed method in predicting heavy rainfall events, making this approach particularly attractive for regions with limited resources or where traditional weather radar or station coverage is sparse.

Balancing mutually diverging performance metrics, such as, processing latency, outcome accuracy, and end device energy consumption is a challenging undertaking for deep learning model inference in ad-hoc edge environments. In this paper, we propose EdgeRL framework that seeks to strike such balance by using an Advantage Actor-Critic (A2C) Reinforcement Learning (RL) approach that can choose optimal run-time DNN inference parameters and aligns the performance metrics based on the application requirements. Using real world deep learning model and a hardware testbed, we evaluate the benefits of EdgeRL framework in terms of end device energy savings, inference accuracy improvement, and end-to-end inference latency reduction. Deep learning models, particularly deep neural networks (DNN), are becoming increasingly important for mission-critical applications, such as public safety, tactical scenarios, search and rescue, and emergency triage, most of which are often edgenative. Unlike traditional edge that are typically part of the network infrastructure, a new paradigm of ad-hoc deployments of edge computing environments are currently being adopted by public safety agencies and armed forces [1]-[3] to support mission-critical use cases.

On-device intelligence (ODI) enables artificial intelligence (AI) applications to run on end devices, providing real-time and customized AI inference without relying on remote servers. However, training models for on-device deployment face significant challenges due to the decentralized and privacy-sensitive nature of users' data, along with end-side constraints related to network connectivity, computation efficiency, etc. Existing training paradigms, such as cloud-based training, federated learning, and transfer learning, fail to sufficiently address these practical constraints that are prevalent for devices. To overcome these challenges, we propose Privacy-Enhanced Training-as-a-Service (PTaaS), a novel service computing paradigm that provides privacy-friendly, customized AI model training for end devices. PTaaS outsources the core training process to remote and powerful cloud or edge servers, efficiently developing customized on-device models based on uploaded anonymous queries, enhancing data privacy while reducing the computation load on individual devices. We explore the definition, goals, and design principles of PTaaS, alongside emerging technologies that support the PTaaS paradigm. An architectural scheme for PTaaS is also presented, followed by a series of open problems that set the stage for future research directions in the field of PTaaS.

As we transition from the mobile internet to the 'Cognitive Internet,' a significant shift occurs in how we engage with technology and intelligence. We contend that the Cognitive Internet goes beyond the Cognitive Internet of Things (Cognitive IoT), enabling connected objects to independently acquire knowledge and understanding. Unlike the Mobile Internet and Cognitive IoT, the Cognitive Internet integrates collaborative intelligence throughout the network, blending the cognitive IoT realm with system-wide collaboration and human intelligence. This integrated intelligence facilitates interactions between devices, services, entities, and individuals across diverse domains while preserving decision-making autonomy and accommodating various identities. The paper delves into the foundational elements, distinct characteristics, benefits, and industrial impact of the 'Cognitive Internet' paradigm. It highlights the importance of adaptable AI infrastructures and hybrid edge cloud (HEC) platforms in enabling this shift. This evolution brings forth cognitive services, a Knowledge as a Service (KaaS) economy, enhanced decision-making autonomy, sustainable digital progress, advancements in data management, processing techniques, and a stronger emphasis on privacy. In essence, this paper serves as a crucial resource for understanding and leveraging the transformative potential of HEC for Cognitive Internet. Supported by case studies, forward-looking perspectives, and real-world applications, it provides comprehensive insights into this emerging paradigm.

Federated Learning (FL) enables training Artificial Intelligence (AI) models over end devices without compromising their privacy. As computing tasks are increasingly performed by a combination of cloud, edge, and end devices, FL can benefit from this End-Edge-Cloud Collaboration (EECC) paradigm to achieve collaborative device-scale expansion with real-time access. Although Hierarchical Federated Learning (HFL) supports multi-tier model aggregation suitable for EECC, prior works assume the same model structure on all computing nodes, constraining the model scale by the weakest end devices. To address this issue, we propose Agglomerative Federated Learning (FedAgg), which is a novel EECC-empowered FL framework that allows the trained models from end, edge, to cloud to grow larger in size and stronger in generalization ability. FedAgg recursively organizes computing nodes among all tiers based on Bridge Sample Based Online Distillation Protocol (BSBODP), which enables every pair of parent-child computing nodes to mutually transfer and distill knowledge extracted from generated bridge samples. This design enhances the performance by exploiting the potential of larger models, with privacy constraints of FL and flexibility requirements of EECC both satisfied. Experiments under various settings demonstrate that FedAgg outperforms state-of-the-art methods by an average of 4.53\% accuracy gains and remarkable improvements in convergence rate.

"Deep neural network (DNN) partition" is a research problem that involves splitting a DNN into multiple parts and offloading them to specific locations. Because of the recent advancement in multi-access edge computing and edge intelligence, DNN partition has been considered as a powerful tool for improving DNN inference performance when the computing resources of edge and end devices are limited and the remote transmission of data from these devices to clouds is costly. This paper provides a comprehensive survey on the recent advances and challenges in DNN partition approaches over the cloud, edge, and end devices based on a detailed literature collection. We review how DNN partition works in various application scenarios, and provide a unified mathematical model of the DNN partition problem. We developed a five-dimensional classification framework for DNN partition approaches, consisting of deployment locations, partition granularity, partition constraints, optimization objectives, and optimization algorithms. Each existing DNN partition approache can be perfectly defined in this framework by instantiating each dimension into specific values. In addition, we suggest a set of metrics for comparing and evaluating the DNN partition approaches. Based on this, we identify and discuss research challenges that have not yet been investigated or fully addressed. We hope that this work helps DNN partition researchers by highlighting significant future research directions in this domain.