Large Language Models (LLMs) are rapidly moving from research settings into the wild, deployed on enterprise infrastructure, personal devices, and edge platforms. While cloud deployments offer scalable compute, concerns over data sovereignty, compliance, latency, and third-party dependence are driving organizations toward edge and on-premise LLMs. This shift introduces new security and privacy challenges: limited compute and memory force aggressive optimizations, including quantization, pruning, model partitioning, and parameter-efficient adaptation, each of which can introduce vulnerabilities and reshape the threat landscape. We describe this tension as the Security-Efficiency Paradox, mechanisms that improve efficiency may weaken robustness, expose new attack surfaces, or increase privacy risks. We examine how compression can degrade safety alignment, how partitioned inference enables reconstruction attacks, and how continuous local adaptation may cause privacy leakage and model drift. To analyze these risks, we introduce a deployment-centric taxonomy organized around three architectural constraints: the Memory Wall, the Quadratic Wall, and the Compute Wall. We derive a unified constraint model that quantifies when unsafe optimizations become unavoidable, linking each wall to specific attack surfaces. Building on this model, we propose the Secure Operational Efficiency Score (SOES), a holistic metric balancing task accuracy, jailbreak resistance, and privacy against energy, memory, and latency, enabling practitioners to configure edge LLMs under real-world hardware limits. We further present a practical decision procedure and targeted mitigations for each optimization-induced vulnerability. Together, these contributions provide a co-designed framework for jointly evaluating security, privacy, and efficiency, laying a foundation for securing edge-native intelligent systems.

Federated Learning (FL) enables collaborative model training while preserving data privacy by keeping raw data locally stored on client devices, preventing access from other clients or the central server. However, recent studies reveal that sharing model gradients creates vulnerability to Model Inversion Attacks, particularly Deep Leakage from Gradients (DLG), which reconstructs private training data from shared gradients. While Homomorphic Encryption has been proposed as a promising defense mechanism to protect gradient privacy, fully encrypting all model gradients incurs high computational overhead. Selective encryption approaches aim to balance privacy protection with computational efficiency by encrypting only specific gradient components. However, the existing literature largely overlooks a theoretical exploration of the spectral behavior of encrypted versus unencrypted parameters, relying instead primarily on empirical evaluations. To address this gap, this paper presents a framework for theoretical analysis of the underlying principles of selective encryption as a defense against model inversion attacks. We then provide a comprehensive empirical study that identifies and quantifies the critical factors, such as model complexity, encryption ratios, and exposed gradients, that influence defense effectiveness. Our theoretical framework clarifies the relationship between gradient selection and privacy preservation, while our experimental evaluation demonstrates how these factors shape the robustness of defenses against model inversion attacks. Collectively, these contributions advance the understanding of selective encryption mechanisms and offer principled guidance for designing efficient, scalable, privacy-preserving federated learning systems.

Link prediction in graph data uses various algorithms and Graph Nerual Network (GNN) models to predict potential relationships between graph nodes. These techniques have found widespread use in numerous real-world applications, including recommendation systems, community/social networks, and biological structures. However, recent research has highlighted the vulnerability of GNN models to adversarial attacks, such as poisoning and evasion attacks. Addressing the vulnerability of GNN models is crucial to ensure stable and robust performance in GNN applications. Although many works have focused on enhancing the robustness of node classification on GNN models, the robustness of link prediction has received less attention. To bridge this gap, this article introduces an unweighted graph poisoning attack that leverages meta-learning with weighted scheme strategies to degrade the link prediction performance of GNNs. We conducted comprehensive experiments on diverse datasets across multiple link prediction applications to evaluate the proposed method and its parameters, comparing it with existing approaches under similar conditions. Our results demonstrate that our approach significantly reduces link prediction performance and consistently outperforms other state-of-the-art baselines.

Recently, deep neural networks have been outperforming conventional machine learning algorithms in many computer vision-related tasks. However, it is not computationally acceptable to implement these models on mobile and IoT devices and the majority of devices are harnessing the cloud computing methodology in which outstanding deep learning models are responsible for analyzing the data on the server. This can bring the communication cost for the devices and make the whole system useless in those times where the communication is not available. In this paper, a new framework for deploying on IoT devices has been proposed which can take advantage of both the cloud and the on-device models by extracting the meta-information from each sample's classification result and evaluating the classification's performance for the necessity of sending the sample to the server. Experimental results show that only 40 percent of the test data should be sent to the server using this technique and the overall accuracy of the framework is 92 percent which improves the accuracy of both client and server models.

Due to the event driven nature and the versatility of GUI designs in Android programs, it is challenging to generate event sequences with adequate code coverage within a reasonable time. A common approach to handle this issue is to rely on GUI models to generate event sequences. These sequences can be effective in covering GUI states, but inconsistent in exposing program behaviors that require specific inputs. A major obstacle to generate such specific inputs is the lack of a systematic GUI exploration process to accommodate the analysis requirements. In this paper, we introduce Android Path Explorer (APEX), a systematic input generation framework using concolic execution. APEX addresses the limitations of model-based sequence generation by using concolic execution to discover the data dependencies of GUI state transitions. Moreover, concolic execution is also used to prioritize events during the exploration of GUI, which leads to a more robust model and accurate input generation. The key novelty of APEX is that concolic execution is not only used to construct event sequences, but also used to traverse the GUI more systematically. As such, our experimental results show that APEX can be used to generate a set of event sequences that achieve high code coverage, as well as event sequences that reach specific targets.

Federated Learning (FL) is a privacy-preserving technique that allows clients to collaboratively train models while keeping their local data private from both other clients and the central server. However, the presence of a malicious server or attacker can lead to model inversion attacks, which compromise clients’ private information. To counteract these threats, homomorphic encryption (HE) has been proposed as a solution, enabling encrypted model aggregation by the server without exposing raw data. This ensures that a malicious server or attacker cannot access model parameters or reveal clients’ local information. However, HE introduces substantial computational and communication overhead, making it impractical for many real-world implementations. To address this, recent research suggests selectively encrypting only a subset of most important model parameters while leaving the rest unencrypted. While it is promising, the challenge lies in identifying these key parameters to maintain security while minimizing the additional costs associated with homomorphic operations. In addition, even with a subset of parameters, none of the existing methods offer truly efficient and practical HE-based solution. In general, these approaches fall short when applied to large, complex deep learning models and are impractical beyond basic implementations. This paper introduces a straightforward yet effective parameter selection and encryption strategy for HE-based federated learning. Our goal is to minimize this computational, and communication overhead caused by homomorphic operations while maintaining robust security, ensuring practical applicability in real-world scenarios. We begin by evaluating different implementations of CKKS (Cheon-Kim-Kim-Song)-based HE algorithm (which allows floating point operations on encrypted domain) across various model architectures, systematically analyzing the additional burden HE imposes in terms of processing time and size of the model parameters. Next, we introduce a magnitude-based selective encryption strategy that not only provide same security guarantees compared to existing methods but is also practical for scaling to larger, more complex models in FL environments. Hence, this approach enables developers to make informed decisions that strike an optimal balance between performance and privacy/security, making it feasible for real-world deployments.

This article presents a comprehensive survey of both attack and defense mechanisms within the federated learning (FL) landscape. Furthermore, it explores the challenges involved and outlines future directions for the development of a robust and efficient FL solution.

Neural Machine Translation (NMT) models have become successful, but their performance remains poor when translating on new domains with a limited number of data. In this paper, we present a novel approach Epi-Curriculum to address low-resource domain adaptation (DA), which contains a new episodic training framework along with denoised curriculum learning. Our episodic training framework enhances the model's robustness to domain shift by episodically exposing the encoder/decoder to an inexperienced decoder/encoder. The denoised curriculum learning filters the noised data and further improves the model's adaptability by gradually guiding the learning process from easy to more difficult tasks. Experiments on English-German and English-Romanian translation show that: (i) Epi-Curriculum improves both model's robustness and adaptability in seen and unseen domains; (ii) Our episodic training framework enhances the encoder and decoder's robustness to domain shift.

Data-driven based method for navigation and positioning has absorbed attention in recent years and it outperforms all its competitor methods in terms of accuracy and efficiency. This paper introduces a new architecture called IMUNet which is accurate and efficient for position estimation on edge device implementation receiving a sequence of raw IMU measurements. The architecture has been compared with one dimension version of the state-of-the-art CNN networks that have been introduced recently for edge device implementation in terms of accuracy and efficiency. Moreover, a new method for collecting a dataset using IMU sensors on cell phones and Google ARCore API has been proposed and a publicly available dataset has been recorded. A comprehensive evaluation using four different datasets as well as the proposed dataset and real device implementation has been done to prove the performance of the architecture. All the code in both Pytorch and Tensorflow framework as well as the Android application code have been shared to improve further research.

This study examines the impact of class-imbalanced data on deep learning models and proposes a technique for data balancing by generating synthetic data for the minority class. Unlike random-based oversampling, our method prioritizes balancing the informative regions by identifying high entropy samples. Generating well-placed synthetic data can enhance machine learning algorithms accuracy and efficiency, whereas poorly-placed ones may lead to higher misclassification rates. We introduce an algorithm that maximizes the probability of generating a synthetic sample in the correct region of its class by optimizing the class posterior ratio. Additionally, to maintain data topology, synthetic data are generated within each minority sample's neighborhood. Our experimental results on forty-one datasets demonstrate the superior performance of our technique in enhancing deep-learning models.

Recent breakthrough technological progressions of powerful mobile computing resources such as low-cost mobile GPUs along with cutting-edge, open-source software architectures have enabled high-performance deep learning on mobile platforms. These advancements have revolutionized the capabilities of today's mobile applications in different dimensions to perform data-driven intelligence locally, particularly for smart health applications. Unlike traditional machine learning (ML) architectures, modern on-device deep learning frameworks are proficient in utilizing computing resources in mobile platforms seamlessly, in terms of producing highly accurate results in less inference time. However, on the flip side, energy resources in a mobile device are typically limited. Hence, whenever a complex Deep Neural Network (DNN) architecture is fed into the on-device deep learning framework, while it achieves high prediction accuracy (and performance), it also urges huge energy demands during the runtime. Therefore, managing these resources efficiently within the spectrum of performance and energy efficiency is the newest challenge for any mobile application featuring data-driven intelligence beyond experimental evaluations. In this paper, first, we provide a timely review of recent advancements in on-device deep learning while empirically evaluating the performance metrics of current state-of-the-art ML architectures and conventional ML approaches with the emphasis given on energy characteristics by deploying them on a smart health application. With that, we are introducing a new framework through an energy-aware, adaptive model comprehension and realization (EAMCR) approach that can be utilized to make more robust and efficient inference decisions based on the available computing/energy resources in the mobile device during the runtime.

One of the most challenging issues in federated learning is that the data is often not independent and identically distributed (nonIID). Clients are expected to contribute the same type of data and drawn from one global distribution. However, data are often collected in different ways from different resources. Thus, the data distributions among clients might be different from the underlying global distribution. This creates a weight divergence issue and reduces federated learning performance. This work focuses on improving federated learning performance for skewed data distribution across clients. The main idea is to adjust the client distribution closer to the global distribution using sample weights. Thus, the machine learning model converges faster with higher accuracy. We start from the fundamental concept of empirical risk minimization and theoretically derive a solution for adjusting the distribution skewness using sample weights. To determine sample weights, we implicitly exchange density information by leveraging a neural network-based density estimation model, MADE. The clients data distribution can then be adjusted without exposing their raw data. Our experiment results on three real-world datasets show that the proposed method not only improves federated learning accuracy but also significantly reduces communication costs compared to the other experimental methods.

In recent years, machine learning as a service (MLaaS) has brought considerable convenience to our daily lives. However, these services raise the issue of leaking users’ sensitive attributes, such as race, when provided through the cloud. The present work overcomes this issue by proposing an innovative privacy-preserving approach called privacy-preserving class overlap (PPCO), which incorporates both a Wasserstein generative adversarial network and the idea of class overlapping to obfuscate data for better resilience against the leakage of attribute-inference attacks(i.e., malicious inference on users’ sensitive attributes). Experiments show that the proposed method can be employed to enhance current state-of-the-art works and achieve superior privacy–utility trade-off. Furthermore, the proposed method is shown to be less susceptible to the influence of imbalanced classes in training data. Finally, we provide a theoretical analysis of the performance of our proposed method to give a flavour of the gap between theoretical and empirical performances.

With the development of machine learning and data science, data sharing is very common between companies and research institutes to avoid data scarcity. However, sharing original datasets that contain private information can cause privacy leakage. A reliable solution is to utilize private synthetic datasets which preserve statistical information from original datasets. In this paper, we propose MC-GEN, a privacy-preserving synthetic data generation method under differential privacy guarantee for machine learning classification tasks. MC-GEN applies multi-level clustering and differential private generative model to improve the utility of synthetic data. In the experimental evaluation, we evaluated the effects of parameters and the effectiveness of MC-GEN. The results showed that MC-GEN can achieve significant effectiveness under certain privacy guarantees on multiple classification tasks. Moreover, we compare MC-GEN with three existing methods. The results showed that MC-GEN outperforms other methods in terms of utility.

The global landscape has been reshaped by the COVID-19 pandemic, causing unprecedented health, economic, and societal disruptions. In this transformative period, technology emerged as a vital lifeline, serving as both a shield against the virus and a catalyst for adapting to new challenges. Digital contact tracing frameworks rapidly emerged and became popular worldwide, thus enabling swift identification of potential exposure and containment of infections. In addition, the subsequent development of digital COVID certificates for vaccinations, immunity, and testing facilitated the safe resumption of daily activities and cross-border travel.

Convolutional neural networks (CNNs) have achieved the state-of-the-art performance in skin lesion analysis. Compared with single CNN classifier, combining the results of multiple classifiers via fusion approaches shows to be more effective and robust. Since the skin lesion datasets are usually limited and statistically biased, while designing an effective fusion approach, it is important to consider not only the performance of each classifier on the training/validation dataset, but also the relative discriminative power (e.g., confidence) of each classifier regarding an individual sample in the testing phase, which calls for an active fusion approach. Furthermore, in skin lesion analysis, the data of certain classes (e.g., the benign lesions) is usually abundant which makes them an over-represented majority, while the data of some other classes (e.g., the cancerous lesions) is deficient which makes them an underrepresented minority. It is more crucial to precisely identify the samples from an underrepresented (i.e., in terms of the amount of data) but more important minority class (e.g., cancerous skin lesions). In other words, misclassifying a more severe skin lesion to a benign or less severe skin lesion should have relative more cost (e.g., money, time and even lives). To address such challenges, we present CS-AF, a cost-sensitive multi-classifier active fusion framework for skin lesion classification. In the experimental evaluation, we prepared 96 base classifiers (of 12 CNN architectures) on the ISIC Challenge 2019 research dataset. Our experimental results show that our framework consistently outperforms both the static and the active fusion competitors in terms of the accuracy and total costs.

Applications of machine learning (ML) models and convolutional neural networks (CNNs) have been rapidly increased. Although state-of-the-art CNNs provide high accuracy in many applications, recent investigations show that such networks are highly vulnerable to adversarial attacks. The black-box adversarial attack is one type of attack that the attacker does not have any knowledge about the model or the training dataset, but it has some input data set and their labels. In this paper, we propose a novel approach to generate a black-box attack in sparse domain whereas the most important information of an image can be observed. Our investigation shows that large sparse (LaS) components play a critical role in the performance of image classifiers. Under this presumption, to generate adversarial example, we transfer an image into a sparse domain and put a threshold to choose only k LaS components. In contrast to the very recent works that randomly perturb k low frequency (LoF) components, we perturb k LaS components either randomly (query-based) or in the direction of the most correlated sparse signal from a different class. We show that LaS components contain some middle or higher frequency components information which leads fooling image classifiers with a fewer number of queries. We demonstrate the effectiveness of this approach by fooling six state-of-the-art image classifiers, the TensorFlow Lite (TFLite) model of Google Cloud Vision platform, and YOLOv5 model as an object detection algorithm. Mean squared error (MSE) and peak signal to noise ratio (PSNR) are used as quality metrics. We also present a theoretical proof to connect these metrics to the level of perturbation in the sparse domain.

Deep learning often requires a large amount of data. In real-world applications, e.g., healthcare applications, the data collected by a single organization (e.g., hospital) is often limited, and the majority of massive and diverse data is often segregated across multiple organizations. As such, it motivates the researchers to conduct distributed deep learning, where the data user would like to build DL models using the data segregated across multiple different data owners. However, this could lead to severe privacy concerns due to the sensitive nature of the data, thus the data owners would be hesitant and reluctant to participate. We propose LDP-DL, a privacy-preserving distributed deep learning framework via local differential privacy and knowledge distillation, where each data owner learns a teacher model using its own (local) private dataset, and the data user learns a student model to mimic the output of the ensemble of the teacher models. In the experimental evaluation, a comprehensive comparison has been made among our proposed approach (i.e., LDP-DL), DP-SGD, PATE and DP-FL, using three popular deep learning benchmark datasets (i.e., CIFAR10, MNIST and FashionMNIST). The experimental results show that LDP-DL consistently outperforms the other competitors in terms of privacy budget and model accuracy.

Convolutional neural networks (CNNs) have achieved great success in skin lesion classification. A balanced dataset is required to train a good model. However, due to the appearance of different skin lesions in practice, severe or even deadliest skin lesion types (e.g., melanoma) naturally have quite small amount represented in a dataset. In that, classification performance degradation occurs widely, it is significantly important to have CNNs that work well on class imbalanced skin lesion image dataset. In this paper, we propose SuperCon, a two-stage training strategy to overcome the class imbalance problem on skin lesion classification. It contains two stages: (i) representation training that tries to learn a feature representation that closely aligned among intra-classes and distantly apart from inter-classes, and (ii) classifier fine-tuning that aims to learn a classifier that correctly predict the label based on the learnt representations. In the experimental evaluation, extensive comparisons have been made among our approach and other existing approaches on skin lesion benchmark datasets. The results show that our two-stage training strategy effectively addresses the class imbalance classification problem, and significantly improves existing works in terms of F1-score and AUC score, resulting in state-of-the-art performance.

This article proposes a novel contrastive self-supervised neural architecture search (NAS) algorithm, which completely alleviates the expensive costs of data labeling inherited from supervised learning. Our algorithm capitalizes on the effectiveness of self-supervised learning for image representations, which is an increasingly crucial topic of computer vision. First, using only a small amount of unlabeled train data under contrastive self-supervised learning allows us to search on a more extensive search space, discovering better neural architectures without surging the computational resources. Second, we entirely relieve the cost for labeled data (by contrastive loss) in the search stage without compromising architectures’ final performance in the evaluation phase. Finally, we tackle the inherent discrete search space of the NAS problem by sequential model-based optimization via the tree-parzen estimator, enabling us to significantly reduce the computational expense response surface. An extensive number of experiments empirically show that our search algorithm can achieve state-of-the-art results with better efficiency in data labeling cost, searching time, and accuracy in final validation.