Wei Song

As the search space and hence the optimum vary through time, dynamic optimization problems (DOPs) bring tremendous difficulties. Changes in DOPs often manifest as diverse dynamics. Consequently, regulating individuals’ search to adapt to diverse dynamics is crucial to tackle DOPs. Besides, due to the inherent population nature in dynamic optimization algorithms (DOAs), loss of global and local diversities is a critical issue deteriorating the performance of DOAs. Faced with these difficulties, this article proposes a neural network-aided differential evolution with double Q-learning (NNDE-DQ), in which an evolutionary regulation network (ERN) is designed to maintain high global and local diversities over time and regulate individuals’ search that can adapt to diverse dynamics. NNDE-DQ first partitions the search space into multiple subspaces and in each subspace distant individuals are selected as the centers of ERN’s hidden nodes activated by radial basis function. Every input individual is mutated with two randomly selected hidden node centers from different subspaces as differential individuals, facilitating the maintenance of a high global diversity due to very distinct differential terms of the population. Moreover, each mutated individual selects a hidden node center from the subspace located by the mutated individual to undergo crossover. Due to distant hidden node centers in each subspace, a high local diversity can be maintained by individuals’ crossover. Besides, DQ is introduced to acquire ERN’s desired output by interactively estimating individuals’ state-action information, enabling ERN to learn the regulation of individuals’ search in dynamic environments and hence adapt to diverse dynamics. The experimental results demonstrate that NNDE-DQ significantly improves the performance in solving various DOPs comparing to seven state-of-the-art DOAs.

With the growth of cloud computing infrastructure, its cost-efficient paradigm is driving a growing wave of small and medium-sized enterprises to migrate data and services to cloud based platforms. However, due to privacy concerns, data encryption prior to outsourcing is an important means of protection, which in turn requires performing queries over the encrypted data. While several approaches do offer support for privacy preserving skyline or top-k queries, they typically struggle with efficiency when extended to secure top-k dominating queries, due to the inherent nature of combining the advantages of both top k and skyline queries. Nevertheless, this nature renders them as a more practical and promising alternative for location-based services. To address this, we introduce STLD, a secure top-k location-based dominating query scheme. Specifically, we develop an innovative index structure called Secure Aggregate R-tree (SAR-tree) by utilizing the Paillier cryptosystem and introducing meticulously crafted noise, while also incorporating principles from aggregate R-trees and semi-blind R-trees. Leveraging this structure, we propose a series of secure sub-protocols to facilitate top-k dominating queries, accompanied by optimization techniques to mitigate latency associated with computationally intensive dominating operations. Given an encrypted query, STLD not only efficiently answers the query but also guarantees the privacy of data(sets), results, queries and access patterns. Finally, STLD undergoes rigorous theoretical security and complexity analysis, complemented by empirical evaluations that demonstrate its performance and feasibility, achieving a reduction in query cost by 40%-60% compared to multiple competing methods.

Faced with high-dimensional expensive optimization problems (HEOPs), existing high-dimensional expensive optimization algorithms (HEOAs) struggle to locate promising areas quickly due to a dramatically expanded search space. Besides, slow convergence is a critical issue causing inefficiency of HEOAs. To address these issues, we propose a multispace dimensionality reduction-based HEOA (MDR-HEOA) where two extreme learning machine-assisted autoencoders (ELM-AEs) are devised. Specifically, an autoencoder with online sequential learning ELM (OSL-ELM-AE) builds a compressed space, where individuals of low quality explore expecting to locate promising areas quickly. Afterwards, the individuals’ positions are reconstructed to the original space by the output layer of OSL-ELM-AE. Online sequential learning ELM continuously updates OSL-ELM-AE’s parameters whenever new individuals arrive, ensuring high generalization performance in reconstructing the positions that approximate the related target ones. In contrast, ELM-AE with an equal number of hidden nodes and the individuals input each time (ELM-EAE) compresses the search space, where medium-quality individuals exploit aiming to expedite convergence. Based on the universal approximation theorem, ELM-EAE can learn these input individuals with zero error, demonstrating high reconstruction capability. Thus, the individuals’ positions with small offsets caused by exploitation can be reconstructed to closely approximate their target ones. Besides, high-quality individuals conduct a refined search in the original space to find the optimal solution of a HEOP. Our experimental results, particularly comprehensive comparisons with nine state-of-the-art HEOAs, validate that MDR-HEOA significantly improves the performance in handling HEOPs.

In recent years, transfer learning-based dynamic constrained multi-objective optimization algorithms (TL-DCMOAs) have shown great promise in handling dynamic constrained multi-objective optimization problems (DCMOPs). In existing TL-DCMOAs, environmental similarity is typically assessed by measures based on Euclidean distance. However, such measures merely capture overall numerical differences between two environments, which are sensitive to local fluctuations in environmental information, potentially leading to misjudgments of environmental similarity. Besides, most existing TL-DCMOAs leverage historical constrained and unconstrained Pareto-optimal solutions as knowledge, but neglect relations and implicit solutions behind them for acquiring knowledge. To address the above issues, an environmental knowledge-driven transfer strategy with collaborative learning (EKTS-CL) is proposed in this article. Specifically, we design two structural similarity measures (SSMs) which consider the mean, variance, and covariance of environmental information, to comprehensively assess the environmental structural discrepancies with respect to concentration, divergence, and correlation, respectively. Using SSMs, the most structurally similar historical unconstrained Pareto-optimal solutions (UPS) and constrained Pareto-optimal solutions (CPS) can be obtained. Additionally, true CPS and UPS share partial solutions across all environments or the majority of environments of a DCMOP. From the obtained CPS and UPS that separately approximate the true CPS and UPS of the current environment, we extract implicit shared solutions and exclusive ones through collaborative learning, providing high-quality knowledge than the obtained plain solutions for knowledge transfer. Our experimental results include comprehensive comparisons with seven state-of-the-art DCMOAs, and show that the proposed EKTS-CL brings significant performance improvements in solving DCMOPs.

With the growing reliance on cloud services for large-scale data management, preserving the security and privacy of outsourced datasets has become increasingly critical. While encrypting data and queries can prevent direct content exposure, recent research reveals that adversaries can still infer sensitive information via access pattern and search path analysis. However, existing solutions that offer strong access pattern privacy often incur substantial performance overhead. In this paper, we propose a novel privacy-preserving range query scheme over encrypted datasets, offering strong security guarantees while maintaining high efficiency. To achieve this, we develop secure learned spatial index (SLS-INDEX), a secure learned index that integrates the Paillier cryptosystem with a hierarchical prediction architecture and noise-injected buckets, enabling data-aware query acceleration in the encrypted domain. To further obfuscate query execution paths, SLS-INDEXbased Range Queries (SLRQ) employs a permutation-based secure bucket prediction protocol. Additionally, we introduce a secure point extraction protocol that generates candidate results to reduce the overhead of secure computation. We provide formal security analysis under realistic leakage functions and implement a prototype to evaluate its practical performance. Extensive experiments on both real-world and synthetic datasets demonstrate that SLRQ significantly outperforms existing solutions in query efficiency while ensuring dataset, query, result, and access pattern privacy.

Recently, machine learning-embedded large-scale multiobjective evolutionary algorithms (LMOEAs) have shown great promise in solving large-scale multiobjective optimization problems (LMOPs). However, the fast convergence of the population to the true Pareto-optimal front (POF) and even distribution of the obtained Pareto-optimal solutions (POSs) on the POF are not adequately considered when tackling an LMOP. Besides, existing LMOEAs typically pair solutions with a matching rule and employ a network to learn the evolution pattern among the obtained solution pairs. It is difficult to learn various evolution patterns through a simple network, which hinders the collaboration of different patterns for enhancing the search capability. Facing such difficulties, this article proposes an LMOEA with multipattern learning and collaboration (LMOEA-MLC), where a single-hidden-layer multioutput network (SMN) is established to learn inductive and hybrid evolution patterns. Specifically, two inductive ones can be learned with the solution pairs built by two matching rules toward fast convergence and even distribution, respectively. Moreover, the solution pairs considering the fusion of the two inductive ones are collected, enabling SMN to learn a hybrid one and thus making a tradeoff between fast convergence and even distribution. Besides, the learned evolution patterns collaborate to enhance the search capability due to the distinct patterns. To enhance learning speed, SMN’s parameters are updated by an incremental random vector functional link (IRVFL). In our experiments, comprehensive comparisons with eight state-of-the-art LMOEAs demonstrate the significant performance improvement of LMOEA-MLC in handling LMOPs.

Recognizing distinct categories from the same super-category is a highly challenging task due to small inter-class and large intra-class variances. Despite the significant progress of deep convolutional neural network-based methods over the past few years, they are incapable of locating complete salient regions with respect to the object and parts (OPs) of an input image. Locating complete salient regions of OPs is very crucial to fine-grained visual categorization (FGVC), as it contributes to adequately capturing the salient features of OPs. Besides, most existing methods focus on extracting OPs’ features in a unidirectional way, neglecting potentially valuable information between OPs and that within themselves of different feature levels. Consequently, the quality of fine-grained features that are discriminative for FGVC is affected. To address these issues, herein an interaction feedback network with salient-region-aware augmentation (IFN-SA) is proposed. Specifically, a preliminary feature learning (PFL) module is first designed through the use of attention-based object and part streams, which preserve the contextual details semantically associated with OPs, facilitating the location of the complete salient regions and acquisition of the salient features of OPs. Following this, a progressive representation learning (PRL) module is built through the specially-designed vertical interaction mechanism (VIM) and horizontal feedback mechanism (HFM). VIM captures the high-level features via the interactions between OPs, while HFM calibrates the salient features with the feedbacks from the high-level ones and progressively excavates the latter ones. The experimental results especially the performance comparisons on four benchmark datasets validate the excellent performance of IFN-SA for FGVC.