The simultaneous development of deep learning techniques and Internet of Things (IoT)/Cyber-physical Systems (CPS) technologies has afforded untold possibilities for improving distributed computing, sensing, and data analysis. Among these technologies, federated learning has received increased attention as a privacy-preserving collaborative learning paradigm, and has shown significant potential in IoT/CPS-driven large-scale smart-world systems. At the same time, the vulnerabilities of deep neural networks, especially to adversarial attacks, cannot be overstated and should not be minimized. Moreover, distributed nature of federated learning makes defense against such adversarial attacks a more challenging problem due to unavailability of local data and resource heterogeneity. To tackle these challenges, in this paper, we propose ZeKoC, a Zero Knowledge Clustering approach to mitigate adversarial attacks. Particularly, we first formulate the problem of resource-constrained adversarial mitigation. Specifically, noting that global server has no access to training samples, we reformulate the unsupervised weight clustering problem. Our proposed ZeKoC approach allows the server to automatically split and merge weight clusters for weight selection and aggregation. Theoretical analysis of our approach demonstrates that convergence is guaranteed. Our experimental results illustrate that, in a non-i.i.d. (i.e., independent-and-identically distributed) data setting, our ZeKoC approach can successfully mitigate general attacks while outperforming state-of-art schemes.

The rise of deep learning and the Internet of Things (IoT) has driven a number of smart-world applications, which are mostly deployed in distributed environments. Federated learning, a privacy-preserving collaborative learning paradigm, has shown considerable potential to leverage the rich distributed data at network edges. Nonetheless, the heterogeneity of IoT devices and their connected network environment impedes federated learning applications in IoT systems. Particularly, the stale gradients updated by slower local learners impact the effectiveness of federated learning. Transmitting weight updates with a large number of users leads to network congestion at the edge of the network and incurs unaffordable communication costs. To overcome these challenges, we propose a transfer knowledge based federated learning framework under a resource-limited distributed system. We formulate a knowledge distillation based federated learning optimization problem with the consideration of dynamic local resource. The proposed approach carries out federated learning with the help of knowledge distillation to avoid occupying the expensive network bandwidth or bringing a heavy burden to the network. Theoretical analysis demonstrates convergence of the learning process. The experimental results on three public datasets illustrate that the proposed framework is capable of substantially improving the efficiency of federated learning and outperforming state-of-the-art schemes.

Distributed machine learning paradigms have benefited from the concurrent advancement of deep learning and the Internet of Things (IoT), among which federated learning is one of the most promising frameworks, where a central server collaborates with local learners to train a global model. The inherent heterogeneity of IoT devices, i.e., non-independent and identically distributed (non-i.i.d.) data, and the inconsistent communication network environment results in the bottleneck of a degraded learning performance and slow convergence. Moreover, most weight averaging-based model aggregation schemes raise learning fairness concerns. In this paper, we propose a peer-to-peer decentralized learning framework to tackle the above issues. Particularly, each local client iteratively finds a learning pair to exchange the local learning model. By doing this, multiple learning objectives are optimized to advocate for learning fairness while avoiding small-group domination. The proposed fairness-aware approach allows local clients to adaptively aggregate the received model based on the local learning performance. The experimental results demonstrate that the proposed approach is capable of significantly improving the efficacy of federated learning and outperforms the state-of-the-art schemes under real-world scenarios, including balanced-i.i.d., unbalanced-i.i.d., balanced-non.i.i.d., and unbalanced-non.i.i.d. environments.

IoT-enabled consumer electronics can collect and analyze data to improve functionality and user experiences, increasingly becoming part of edge computing networks. Decentralized federated learning is envisioned as a promising AIoT framework to leverage the data generated by these interconnected consumer electronics without compromising user privacy, where computations are performed closer to the data source rather than in a centralized cloud-based data center. However, a number of obstacles bottleneck the deployment of decentralized learning frameworks under a large-scale AIoT scenario, such as heterogeneity issues, communication efficiency, large-scale issues and others. Specifically, heterogeneity issues include diverse learning models, non-i.i.d. data distribution and unbalanced datasets. In this paper, we propose a trustworthy and adaptive knowledge distillation based decentralized federated learning framework to overcome a number of heterogeneity and large-scale issues. To achieve this goal, we formulate a decentralized optimization problem which formally presents various heterogeneity issues. The proposed framework is composed of two components, including a knowledge distillation based algorithm to transfer knowledge over diverse learning models as well as a real-time evaluation module to adapt client drift issues. Convergence analysis demonstrates that the convergence of the proposed framework is guaranteed. Finally, we conduct extensive experiments over three public datasets to validate the effectiveness and efficiency of proposed framework in the large-scale IoT scenario, illustrating that the proposed framework could significantly improve the learning performance over various metrics and outperforming state-of-the-art baseline approaches.

Embodied AI has notably advanced the autonomy of physical agents such as robots, vehicles, and AAVs, expanding their application scope. However, existing systems are predominantly data-driven, relying on static programming and pre-trained models. This limits their adaptability to dynamic and unforeseen scenarios. Additionally, the high computational cost of training large-scale models locally hinders their practical deployment. One promising solution lies in integrating Large Language Models (LLMs) into Embodied AI frameworks. Although LLMs excel in reasoning, coding, and perception, most existing frameworks adopt a single-LLM architecture, which restricts their effectiveness in addressing complex, multimodal tasks. The diverse strengths of individual LLMs, ranging from natural language understanding, visual processing to code generation, are seldom utilized in a collaborative and structured manner. To address these challenges, we propose a knowledge-driven framework, called EUF, that incorporates multi-LLMs into the Embodied AI architecture for AAV-as-a-Service in Intelligent Transportation Systems. Each LLM is dedicated to a specific stage of the AAV task, including user intent interpretation, adaptive path planning with code generation, and error-feedback mechanisms. Our research explores both One-shot and Segmented Code Generation approaches using various LLM-driven models to identify the optimal strategy. We conduct an in-depth analysis of different code errors to evaluate the strengths and limitations of each approach, including the feedback capabilities of the LLM-driven models. Experiments conducted in the AirSim environment demonstrate the framework’s robustness and accuracy in complex AAV path planning tasks, highlighting its practical potential for real-world ITS deployments.