As a new and efficient ensemble learning algorithm, XGBoost has been widely applied for its multitudinous advantages, but its classification effect in the case of data imbalance is often not ideal. Aiming at this problem, an attempt was made to optimize the regularization term of XGBoost, and a classification algorithm based on mixed sampling and ensemble learning is proposed. The main idea is to combine SVM-SMOTE over-sampling and EasyEnsemble under-sampling technologies for data processing, and then obtain the final model based on XGBoost by training and ensemble. At the same time, the optimal parameters are automatically searched and adjusted through the Bayesian optimization algorithm to realize classification prediction. In the experimental stage, the G-mean and area under the curve (AUC) values are used as evaluation indicators to compare and analyze the classification performance of different sampling methods and algorithm models. The experimental results on the public data set also verify the feasibility and effectiveness of the proposed algorithm.

In unmanned retail store, providing smart customer service requires two stages: understanding customer needs, and guiding the customer to the product. In this paper, we propose an end-to-end (Customer-to-Shelf) software service framework for unmanned retail. The framework integrates visual recognition technology to detect retail objects, large language models to analyze customer shopping needs and make proper recommendations. First, deep neural network based image recognition models are studied for implementing effective stock keeping units (SKUs) object recognition on the shelf. Second, a novel method is proposed to fine-tune large language models (LLMs) with limited training dataset. Metaheuristic approaches are used to optimize the mask locations in a low dimensional parameter space, resulting a more efficient parameter updating method for limited downstream data. Third, by facilitating an automatic analysis of customer preferences powered by large language models, we present a smart recommender system based on domain-specific knowledge, which completes the Customer-to-Shelf software service framework. Experimental results show that our proposed fine-tuning method, is more efficient than other state-of-the-art training methods for limited downstream domain dataset. Using fine-tuned large models, we can successfully create a seamless shopping experience for customers by understanding personalized needs and providing shopping advice in the unmanned retail store.

Two basic security requirements in communication are confidentiality and authentication. Signcryption is an ideal technique to transmit encrypted and authenticated data. In view of the shortcomings of existing signcryption schemes and the high security of elliptic curve cryptography (ECC), we design a ECC-based signcryption scheme and evaluate it in terms of security, computational overhead, and communication overhead. Finally, we consider the application of our secure and efficient signcryption scheme in the smart lock key management system and analyze the bit-oriented performance of the designed key management scheme.

Federated learning, as a distributed machine learning framework, aims to protect data privacy while addressing the issue of data silos by collaboratively training models across multiple clients. However, a significant challenge to federated learning arises from the non-independent and identically distributed (non-iid) nature of data across different clients. non-iid data can lead to inconsistencies between the minimal loss experienced by individual clients and the global loss observed after the central server aggregates the local models, affecting the model’s convergence speed and generalization capability. To address this challenge, we propose a novel federated learning algorithm based on update bias (FedUB). Unlike traditional federated learning approaches such as FedAvg and FedProx, which independently update model parameters on each client before direct aggregation to form a global model, the FedUB algorithm incorporates an update bias in the loss function of local models—specifically, the difference between each round’s local model updates and the global model updates. This design aims to reduce discrepancies between local and global updates, thus aligning the parameters of locally updated models more closely with those of the globally aggregated model, thereby mitigating the fundamental conflict between local and global optima. Additionally, during the aggregation phase at the server side, we introduce a metric called the bias metric, which assesses the similarity between each client’s local model and the global model. This metric adaptively sets the weight of each client during aggregation after each training round to achieve a better global model. Extensive experiments conducted on multiple datasets have confirmed the effectiveness of the FedUB algorithm. The results indicate that FedUB generally outperforms methods such as FedDC, FedDyn, and Scaffold, especially in scenarios involving partial client participation and non-iid data distributions. It demonstrates superior performance and faster convergence in tasks such as image classification.

Immunoglobulin A nephropathy (IgAN) is the most common primary glomerular disease all over the world and it is a major cause of renal failure. IgAN prediction in children with machine learning algorithms has been rarely studied. We retrospectively analyzed the electronic medical records from the Nanjing Eastern War Zone Hospital, chose eXtreme Gradient Boosting (XGBoost), random forest (RF), CatBoost, support vector machines (SVM), k-nearest neighbor (KNN), and extreme learning machine (ELM) models in order to predict the probability that the patient would not reach or reach end-stage renal disease (ESRD) within five years, used the chi-square test to select the most relevant 16 features as the input of the model, and designed a decision-making system (DMS) of IgAN prediction in children that is based on XGBoost and Django framework. The receiver operating characteristic (ROC) curve was used in order to evaluate the performance of the models and XGBoost had the best performance by comparison. The AUC value, accuracy, precision, recall, and f1-score of XGBoost were 85.11%, 78.60%, 75.96%, 76.70%, and 76.33%, respectively. The XGBoost model is useful for physicians and pediatric patients in providing predictions regarding IgAN. As an advantage, a DMS can be designed based on the XGBoost model to assist a physician to effectively treat IgAN in children for preventing deterioration.

The advancement of deep neural networks (DNNs) has prompted many cloud service providers to offer deep learning as a service (DLaaS) to users across various application domains. However, in current DLaaS prediction systems, users’ data are at risk of leakage. Homomorphic encryption allows operations to be performed on ciphertext without decryption, which can be applied to DLaaS to ensure users’ data privacy. However, mainstream homomorphic encryption schemes only support homomorphic addition and multiplication, and do not support the ReLU activation function commonly used in the activation layers of DNNs. Previous work used approximate polynomials to replace the ReLU activation function, but the DNNs they implemented either had low inference accuracy or high inference latency. In order to achieve low inference latency of DNNs on encrypted data while ensuring inference accuracy, we propose a low-degree Hermite deep neural network framework (called LHDNN), which uses a set of low-degree trainable Hermite polynomials (called LotHps) as activation layers of DNNs. Additionally, LHDNN integrates a novel weight initialization and regularization module into the LotHps activation layer, which makes the training process of DNNs more stable and gives a stronger generalization ability. Additionally, to further improve the model accuracy, we propose a variable-weighted difference training (VDT) strategy that uses ReLU-based models to guide the training of LotHps-based models. Extensive experiments on multiple benchmark datasets validate the superiority of LHDNN in terms of inference speed and accuracy on encrypted data.