Hai-Lin Liu's research while affiliated with GuangDong University of Technology and other places

Differentiable bilevel Neural Architecture Search (NAS) has emerged as a powerful approach in automated machine learning (AutoML) for efficiently searching for neural network architectures. However, the existing differentiable methods encounter challenges, such as the risk of becoming trapped in local optima and the computationally expensive Hessian matrix inverse calculation performed when solving the bilevel NAS optimization model. In this paper, a novel-but-efficient stochastic bilevel optimization approach, called STO-DARTS, is proposed for the bilevel NAS optimization problem. Specifically, we design a hypergradient estimate, which is constructed using stochastic gradient descent from the gradient information contained in the Neumann series. This estimate alleviates the issue of local optima traps, enabling searches for exceptional network architectures. To validate the effectiveness and efficiency of the proposed method, two versions of STO-DARTS with different hypergradient estimators are constructed and experimentally tested on different datasets in NAS-Bench-201 and DARTS search spaces. The experimental results show that the proposed STO-DARTS approach achieves competitive performance with that of other state-of-the-art NAS methods in terms of determining effective network architectures. To support our approach, we also provide theoretical analyses.

Deploying in-building distributed antenna systems (IB-DAS) is a crucial step towards providing ubiquitous wireless services. In this paper, we study the multiobjective network planning problem, aiming to minimize both construction costs and average power loss. The main challenge in solving this problem is efficiently representing the network structure. To address this, we encode the network structure as a spanning tree, with the root node connecting to the signal source, and leaf and non-leaf nodes representing all floors and power devices, respectively. Compared to existing encodings, this tree encoding offers several advantages, including improved locality and the ability to easily generate valid solutions. Additionally, we propose a tree-encoded evolutionary algorithm called TMOEA. Since the standard operators cannot be applied, we devise problem-specific crossover and mutation operators to produce tree-encoded solutions. Moreover, the Tchebycheff approach is employed to update solutions. Comprehensive experiments on 11 test instances with up to 30 floors demonstrate that the proposed algorithm outperforms four compared algorithms in terms of both the hypervolume indicator and the inverted generational distance indicator for each test instance.

This article addresses the pricing scheme in a wireless-powered cooperative mobile edge computing (WP-CoMEC) system, focusing on personalized service customization. Traditional pricing schemes in such systems often assume a passive mode, with the service provider leading, and the device owner following. However, with the rise of personalized requirements, this paper proposes a novel approach where the device owner becomes an active participant in the pricing scheme, leading to personally customized services. The proposed pricing model formulates a bilevel multi-objective optimization problem, considering task offloading, resource allocation, and energy harvesting. This comprehensive approach ensures a more holistic optimization process. To address the computational challenges posed by the bilevel pricing model, this article proposes a bilevel periodically interactive evolutionary algorithm (BL-PIEA), which efficiently handles mixed variables, complex objective conflicts, and the inner nested structure of the bilevel pricing model. The proposed BL-PIEA is tested on ten instances, and the results indicate that BL-PIEA can effectively solve the proposed pricing model, showcasing superior performance in terms of reduced run time and saved evaluation budgets compared to other algorithms. With the proposed bilevel pricing model solved by BL-PIEA, the service provider can make out better pricing schemes that satisfy the device owner's requirements, so as to achieve a good personalized service customization.

Evolutionary computation (EC) based method for Neural Architecture Search (NAS) is a thriving research field. Current NAS research typically focuses on concurrently searching for architecture and given operational hyperparameters (e.g. kernel size, stride, and padding). The architecture and its operational hyperparameters obtained through this method may not necessarily be the most optimal match, ignoring the exploration of operational hyperparameters space. To address this challenge, we propose an EC-based NAS method, namely EA-FCNet, to strike a balance between global search and local search. Specifically, we design a hierarchical encoding strategy distinguishing architecture and operational hyperparameters so that the algorithm can search both in a nested manner. Meanwhile, unlike existing methods based on convolutional neural networks (CNNs), our search space combines convolutional and fully connected operations to extract comprehensive features and enhance feature association. Furthermore, we address the challenge of combining these two operations that have mismatched input and output shapes by introducing a repairment strategy that allows the algorithm to handle such situations seamlessly. To evaluate the performance of the proposed algorithm, we conducted extensive comparison and ablation experiments on two publicly available datasets: DRIVE and CHASE_DB1, which demonstrate the effectiveness of the proposed algorithm, with the obtained architecture achieving competitive results. Additionally, the proposed updating strategy has increased the search speed tenfold.