In IEEE TIFS 2023, Chen et al. proposed a method called APFed, which leverages additive homomorphic encryption to encrypt each client’s gradient, aiming to prevent information leakage while effectively defending against poisoning attacks. In this paper, we demonstrate a fundamental flaw in the authors’ claim of security proof yielding that the proposed APFed method is insecure.

We propose Powerformer, an efficient homomorphic encryption (HE)-based privacypreserving language model (PPLM) designed to reduce computational overhead while maintaining model performance.Powerformer incorporates three key techniques to optimize encrypted computations: 1) A novel distillation technique that replaces softmax and layer normalization with computationally efficient power and linear functions, ensuring no performance degradation while enabling seamless encrypted computation.2) A pseudo-sign composite approximation method that accurately approximates GELU and tanh functions with minimal computational overhead.3) A homomorphic matrix multiplication algorithm specifically optimized for Transformer models, enhancing efficiency in encrypted environments.By integrating these techniques, Powerformer based on the BERT-base model achieves a 45% reduction in computation time compared to the state-of-the-art HE-based PPLM without any loss in accuracy.

Bootstrapping in approximate homomorphic encryption involves evaluating the modular reduction function. Traditional methods decompose the modular reduction function into three components: scaled cosine, double-angle formula, and inverse sine. While these approaches offer a strong trade-off between computational cost and level consumption, they lack flexibility in parameterization. In this work, we propose a new method to decompose the modular reduction function with improved parameterization, generalizing prior trigonometric approaches. Numerical experiments demonstrate that our method achieves near-optimal approximation errors. Additionally, we introduce a technique that integrates the rescaling operation into matrix operations during bootstrapping, further reducing computational overhead.

Recently, the standard ResNet-20 network was successfully implemented on the fully homomorphic encryption scheme, residue number system variant Cheon-Kim-Kim-Song (RNS-CKKS) scheme using bootstrapping, but the implementation lacks practicality due to high latency and low security level. To improve the performance, we first minimize total bootstrapping runtime using multiplexed parallel convolution that collects sparse output data for multiple channels compactly. We also propose the imaginary-removing bootstrapping to prevent the deep neural networks from catastrophic divergence during approximate ReLU operations. In addition, we optimize level consumptions and use lighter and tighter parameters. Simulation results show that we have 4.67x lower inference latency and 134x less amortized runtime (runtime per image) for ResNet-20 compared to the state-of-the-art previous work, and we achieve standard 128-bit security. Furthermore, we successfully implement ResNet-110 with high accuracy on the RNS-CKKS scheme for the first time.

The comparison operation for two numbers is one of the most frequently used operations in several applications, including deep learning. As such, lots of research has been conducted with the goal of efficiently evaluating the comparison operation in homomorphic encryption schemes. Recently, Cheonet al.(Asiacrypt 2020) proposed new comparison methods that approximated the sign function on homomorphically encrypted data using composite polynomials and proved that these methods had optimal asymptotic complexity. In this article, we propose a practically optimal method that approximates the sign function using compositions of minimax approximation polynomials. We prove that this approximation method is optimal with respect to depth consumption and the number of non-scalar multiplications. In addition, we propose a polynomial-time algorithm that determines the optimal composition of minimax approximation polynomials for the proposed homomorphic comparison operation using dynamic programming. The numerical analysis demonstrates that when minimizing runtime, the proposed comparison operation reduces the runtime by approximately 45 percent on average when compared to the previous algorithm. Likewise, when minimizing depth consumption, the proposed algorithm reduces the runtime by approximately 41 percent on average. In addition, when high precision in the comparison operation is required, the previous algorithm does not achieve 128-bit security, while the proposed algorithm does due to its small depth consumption.

Many optimization problems in recent engineering are complex and high-dimensional problems, a so-called Large-Scale Global Optimization (LSGO) problem, due to the increasing requirements for multidisciplinary approach. This paper proposes a novel Bare Bones Particle Swarm Optimization (BBPSO) algorithm to solve LSGO problems. The BBPSO is a variant of a Particle Swarm Optimization (PSO) and is based on Gaussian distribution. The BBPSO does not consider the selection of controllable parameters of the PSO and is a simple but powerful optimizer. This algorithm, however, is vulnerable to LSGO problems. This study has improved its performance for LSGO problems by combining the heterogeneous cooperation based on the information exchange between particles and the Gaussian jump strategy to avoid local optima. The CEC'2015 Special Session on Large-Scale Global Optimization has given 15 benchmark problems to provide convenience and flexibility for comparing various optimization algorithms specifically designed for large-scale global optimization. Simulations performed with those benchmark problems have verified the performance of the proposed optimizer and compared with the reference algorithm DECC-G of the CEC'2015 special session on LSGO.

The Efficient-Energy Coverage (EEC) problem is an important issue when implementing Wireless Sensor Networks (WSNs) because of the need to limit energy use. In this paper, we propose a new approach to solving the EEC problem using a novel Ant Colony Optimization (ACO) algorithm. The proposed ACO algorithm has a unique characteristic that conventional ACO algorithms do not have. The proposed ACO algorithm (Three Pheromones ACO, TPACO) uses three types of pheromones to find the solution efficiently, whereas conventional ACO algorithms use only one type of pheromone. One of the three pheromones is the local pheromone, which helps an ant organize its coverage set with fewer sensors. The other two pheromones are global pheromones, one of which is used to optimize the number of required active sensors per Point of Interest (PoI), and the other is used to form a sensor set that has as many sensors as an ant has selected the number of active sensors by using the former pheromone. The TPACO algorithm has another advantage in that the two user parameters of ACO algorithms are not used. We also introduce some techniques that lead to a more realistic approach to solving the EEC problem. The first technique is to utilize the probabilistic sensor detection model. The second method is to use different kinds of sensors, i.e., heterogeneous sensors in continuous space, not a grid-based discrete space. Simulation results show the effectiveness of our algorithm over other algorithms, in terms of the whole network lifetime.

BACKGROUND: The importance of biodiversity conservation has been increasing steadily due to its benefits to human beings. Recently, producing and managing biodiversity databases have become much easier because of the information technology (IT) advancement. This made the general public's participation in biodiversity conservation much more practical than ever. For example, an open free web service can be devised for a wider spectrum of people to collaborate with each other for sharing biodiversity information. Bird migration is one such area of the collaboration. Korean migratory birds are usually traceable in the important routes of the East Asian-Australia Flyway (EAAF), and they play a key role as an environmental change indicator of the Earth. Therefore, the preservation of migratory birds requires an information system which involves a broader range of voluntary and interactive knowledge network to process bird information production, circulation, and dissemination. RESULTS: The Korean Bird Information System (KBIS) aims to construct a cooperative partnership domestically and internationally through the acquisition, management, and sharing of Korean bird information involving both expert and non-expert groups. KBIS has six goals: data standard, system linkage, data diversity, utilization, bird knowledge network, and statistics. The key features of KBIS are to provide a simple search, gallery (photographs), and community to lead the participation of numerous non-experts, especially amateur bird watchers. The function of real-time observation data submission through the internet has been accomplished. It also provides bird banding database, statistics, and taxon network for experts. Especially, the statistics part provides the user with easy understanding of ecological trends of species based on the time and region. CONCLUSION: KBIS is a tool for the conservation and management of bird diversity and ecosystem that encourages users to participate by providing the open free data access and real-time data input web-interface. It will enhance bird knowledge networking activities locally, nationally, and internationally. In addition, it provides opportunities to enhance the public awareness for the preservation of bird diversity and species information in relevant localities through the database construction and networking activities. It can be found at http://korbird.naris.go.kr.