Ling Wang

AAAI Conference 2026 Conference Paper

Context-aware Graph Meta-learning

• Ningbo Huang
• Gang Zhou
• Meng Zhang
• Shunhang Li
• Ling Wang
• Shiyu Wang
• Yi Xia

Developing a universal graph model capable of generalizing across diverse graph domains has consistently been a key objective in graph learning. Recently, many studies have focused on achieving in-context learning (ICL) on graphs, which can generalize to novel tasks without the need for fine-tuning, similar to large language models (LLMs) such as GPT-3. These researches can be primarily divided into graph-based methods and LLM-based methods. However, the generalization performance of the former is limited by the representation capability of GNNs, while the latter faces the challenge of LLMs understanding graph structures. Therefore, we propose CAGML, a context-aware graph meta-learning model, which learns to generalize to cross-domain and cross-granularity graph tasks using a meta-trained Transformer. Firstly, we formulate graph few-shot learning tasks as a structure-aware sequence modeling problem to unify cross-domain and cross-granularity tasks. Then, a structure-aware Transformer (SAT) is introduced as a graph in-context learner to make predictions with a few labels and the task-specific structural context. Finally, we pre-train SAT in a meta-optimization manner on large-scale citation network and knowledge graph. Experiments on 6 cross-domain graph datasets show that, without fine-tuning, CAGML can achieve state-of-the-art (SOTA) performance in terms of average performance across cross-granularity tasks on adopted datasets.

ICRA Conference 2025 Conference Paper

DAP-LED: Learning Degradation-Aware Priors with Clip for Joint Low-Light Enhancement and Deblurring

• Ling Wang
• Chen Wu
• Lin Wang

Autonomous vehicles and robots often struggle with reliable visual perception at night due to the low illumination and motion blur caused by the long exposure time of RGB cameras. Existing methods address this challenge by sequentially connecting the off-the-shelf pretrained lowlight enhancement and deblurring models. Unfortunately, these methods often lead to noticeable artifacts (e. g. , color distortions) in the over-exposed regions or make it hardly possible to learn the motion cues of the dark regions. In this paper, we interestingly find vision-language models, e. g. , Contrastive LanguageImage Pretraining (CLIP), can comprehensively perceive diverse degradation levels at night. In light of this, we propose a novel transformer-based joint learning framework, named DAP-LED, which can jointly achieve low-light enhancement and deblurring, benefiting downstream tasks, such as depth estimation, segmentation, and detection in the dark. The key insight is to leverage CLIP to adaptively learn the degradation levels from images at night. This subtly enables learning rich semantic information and visual representation for optimization of the joint tasks. To achieve this, we first introduce a CLIPguided cross-fusion module to obtain multi-scale patch-wise degradation heatmaps from the image embeddings. Then, the heatmaps are fused via the designed CLIP-enhanced transformer blocks to retain useful degradation information for effective model optimization. Experimental results show that, compared to existing methods, our DAP-LED achieves state-of-the-art performance in the dark. Meanwhile, the enhanced results are demonstrated to be effective for three downstream tasks. For demo and more results, please check the project page: https://vlislab22.github.io/dap-led/.

EAAI Journal 2025 Journal Article

Federated Reinforcement Learning for smart and privacy-preserving energy management of residential microgrids clusters

• Mao Tan
• Jie Zhao
• Xiao Liu
• Yongxin Su
• Ling Wang
• Rui Wang
• Zhuocen Dai

Real-time energy management optimizes energy utilization and manages electrical loads, which is crucial for improving the operational efficiency of residential microgrids. However, existing management methods suffer from model complexity and slow training speed. To solve this problem, we introduce Federated Reinforcement Learning to manage residential microgrids by training a control strategy in a decentralized and privacy-preserving manner. Specifically, a residential microgrid energy optimization management model is first established based on the Proximal Policy Optimization (PPO) method. Then, we propose a cooperative training strategy for multiple Residential microgrids based on Federated Reinforcement Learning (RFRL). The proposed method improves the training speed of residential microgrid models by sharing parameter information, such as network weights, while protects users’ usage data. Finally, clustering analysis is introduced in the case of heterogeneous residential microgrid data. Extensive experimental evaluation shows that our method outperforms the alternative residential microgrid management methods in terms of cost efficiency.

AIIM Journal 2025 Journal Article

Medical multimodal foundation models in clinical diagnosis and treatment: Applications, challenges, and future directions

• Kai Sun
• Siyan Xue
• Fuchun Sun
• Haoran Sun
• Yu Luo
• Ling Wang
• Siyuan Wang
• Na Guo

Recent advancements in deep learning have significantly revolutionized the field of clinical diagnosis and treatment, offering novel approaches to improve diagnostic precision and treatment efficacy across diverse clinical domains, thus driving the pursuit of precision medicine. The growing availability of multi-organ and multimodal datasets has accelerated the development of large-scale Medical Multimodal Foundation Models (MMFMs). These models, known for their strong generalization capabilities and rich representational power, are increasingly being adapted to address a wide range of clinical tasks, from early diagnosis to personalized treatment strategies. This review offers a comprehensive analysis of recent developments in MMFMs, focusing on three key aspects: datasets, model architectures, and clinical applications. We also explore the challenges and opportunities in optimizing multimodal representations and discuss how these advancements are shaping the future of healthcare by enabling improved patient outcomes and more efficient clinical workflows.

IJCAI Conference 2025 Conference Paper

PNAct: Crafting Backdoor Attacks in Safe Reinforcement Learning

• Weiran Guo
• Guanjun Liu
• Ziyuan Zhou
• Ling Wang

Reinforcement Learning (RL) is widely used in tasks where agents interact with an environment to maximize rewards. Building on this foundation, Safe Reinforcement Learning (Safe RL) incorporates a cost metric alongside the reward metric, ensuring that agents adhere to safety constraints during decision-making. In this paper, we identify that Safe RL is vulnerable to backdoor attacks, which can manipulate agents into performing unsafe actions. First, we introduce the relevant concepts and evaluation metrics for backdoor attacks in Safe RL. It is the first attack framework in the Safe RL field that involves both Positive and Negative Action sample (PNAct) is to implant backdoors, where positive action samples provide reference actions and negative action samples indicate actions to be avoided. We theoretically point out the properties of PNAct and design an attack algorithm. Finally, we conduct experiments to evaluate the effectiveness of our proposed backdoor attack framework, evaluating it with the established metrics. This paper highlights the potential risks associated with Safe RL and underscores the feasibility of such attacks. Our code and supplementary material are available at https: //github. com/azure-123/PNAct.

EAAI Journal 2024 Journal Article

A comprehensive survey of weapon target assignment problem: Model, algorithm, and application

• Jinrui Li
• Guohua Wu
• Ling Wang

This paper provides an overview of the weapon target assignment problem, which aims to optimize the assignment of weapons to targets in order to maximize weapon damage to targets. The weapon target assignment problem can be viewed as a specialized instance of the optimal resource assignment problem. With the advancement of weapons technology, high-speed and high-lethality missiles have become more advanced, and their tactical applications more diverse. These missiles can strike targets with greater accuracy and improved concealment, posing a significant threat to both attackers and defenders. Consequently, the weapon target assignment problem has become a pressing concern in the field of military offense and defense. Subsequently, researchers worldwide are devoting significant efforts to address the weapon target assignment problem through the utilization of exact algorithms, heuristic algorithms, meta-heuristic algorithms, and artificial intelligence methods. This paper provides a brief review of the weapon target assignment problem development history, formula, solution techniques, and applications. We categorize weapon target assignment problems into four different formulas, considering the complexity of combat scenarios, and summarize various solution methods for each category. Furthermore, we also emphasize the relevance of weapon target assignment problems in national defense applications. Lastly, we conclude by discussing potential avenues for future research in addressing the weapon target assignment problem.

EAAI Journal 2024 Journal Article

Constrained multi-objective optimization evolutionary algorithm for real-world continuous mechanical design problems

• Fei Ming
• Wenyin Gong
• Huixiang Zhen
• Ling Wang
• Liang Gao

During the past two decades, evolutionary algorithms have seen great achievements in solving complex optimization problems owing to the advantages brought by their properties, especially constrained multi-objective optimization problems (CMOPs) with multiple conflicting objective functions and constraints which widely exist in industry, scientific research, and daily life. Among the real-world CMOPs, mechanical design problems (MDPs) from the industry widely exist and are important, while unfortunately, most constrained multi-objective evolutionary algorithms (CMOEAs), developed based on benchmark CMOPs, neglect the specific features and challenges of MDPs and thus cannot solve them well to provide the practitioners promising Pareto optimal solutions for decision making. To overcome this limitation, this paper analyzes the features and challenges of MDPs, including badly scaled objective space, decision space properties, and decision variable linkages. Then, we propose a new CMOEA named CMORWMDP. First, instead of the homogeneous operator in existing CMOEAs, a heterogeneous operator strategy is adopted to use the operator of Genetic Algorithm to enhance the convergence and the operator of Differential Evolution to tackle variable linkages. In addition, an improved fitness function that considers normalization is designed for environmental and mating selections. The proposed algorithm is simple, parameter-free, and easy to implement. Experiments on 21 real-world MDPs show its superiority compared to 20 state-of-the-art CMOEAs under the Friedman test and Wilcoxon test on different metrics, demonstrating the effectiveness of the heterogeneous operator and normalization-based fitness for selections for real-world MDPs. Moreover, the effectiveness of the proposed algorithm in solving other real-world CMOPs is also verified, revealing that our methods are very promising in tackling real-world problems.

YNIMG Journal 2024 Journal Article

Contrastive voxel clustering for multiscale modeling of brain network

• Zhiyuan Ding
• Yulang Huang
• Xiangzhu Zeng
• Shiyin Jiang
• Shuyang Feng
• Zhenduo Wang
• Ling Wang
• Zeng Wang

Resting-state functional magnetic resonance imaging (fMRI) provides an efficient way to analyze the functional connectivity between brain regions. A comprehensive understanding of brain functionality requires a unified description of multi-scale layers of neural structure. However, existing brain network modeling methods often simplify this property by averaging Blood oxygen level dependent (BOLD) signals at the brain region level for fMRI-based analysis with the assumption that BOLD signals are homogeneous within each brain region, which ignores the heterogeneity of voxels within each Region of Interest (ROI). This study introduces a novel multi-stage self-supervised learning framework for multiscale brain network analysis, which effectively delineates brain functionality from voxel to ROIs and up to sample level. A Contrastive Voxel Clustering (CVC) module is proposed to simultaneously learn the voxel-level features and clustering assignments, which ensures the retention of informative clustering features at the finest voxel-level and concurrently preserves functional connectivity characteristics. Additionally, based on the extracted features and clustering assignments at the voxel level by CVC, a Brain ROI-based Graph Neural Network (BR-GNN) is built to extract functional connectivity features at the brain ROI-level and used for sample-level prediction, which integrates the functional clustering maps with the pre-established structural ROI maps and creates a more comprehensive and effective analytical tool. Experiments are performed on two datasets, which illustrate the effectiveness and generalization ability of the proposed method by analyzing voxel-level clustering results and brain ROIs-level functional characteristics. The proposed method provides a multiscale modeling framework for brain functional connectivity analysis, which will be further used for other brain disease identification. Code is available at https://github.com/yanliugroup/fmri-cvc.

EAAI Journal 2024 Journal Article

Evolutionary computation and reinforcement learning integrated algorithm for distributed heterogeneous flowshop scheduling

• Rui Li
• Ling Wang
• Wenyin Gong
• Jingfang Chen
• Zixiao Pan
• Yuting Wu
• Yang Yu

With the advancement of the global economy, there is a growing focus on distributed manufacturing. This study addresses the complex challenges posed by the distributed heterogeneous flow shop scheduling problem (DHFSP), wherein multiple machine processing speeds are taken into account. The primary objectives involve the simultaneous minimization of both makespan and total energy consumption. To tackle this intricate problem, we propose an evolutionary computation and reinforcement learning integrated algorithm (ECRLIA) approach. Initially, an optimization framework is meticulously crafted to synergistically integrate both evolutionary computation and reinforcement learning solvers. Subsequently, a multi-rule cooperation initialization is devised to expedite the pre-search process across all solvers. Following this, a competition-based cooperative evolutionary algorithm is introduced to conduct a global search, thereby providing an initial solution to the DHFSP. The interplay of competition and cooperation among individuals enhances convergence. Further, a Q-learning approach employing dual agents is designed to perform a local search, supplementing solutions that evolutionary algorithms may struggle to uncover. This learning method incorporates an auxiliary agent to evaluate the action predictions of the primary agent, ensuring more stable learning. The effectiveness of the proposed algorithm is assessed through numerical experiments, which validate the efficacy of the cooperation framework, initialization cooperation, and the enhanced Q-learning method. Furthermore, ECRLIA is benchmarked against five state-of-the-art algorithms for DHFSP, and the results affirm the significant superiority of the proposed ECRLIA in addressing DHFSP compared to other algorithms.

EAAI Journal 2024 Journal Article

Spatio-temporal features for fast early warning of unplanned self-extubation in ICU

• Yang Chen
• Ling Wang
• Guorong Wang
• Shuang Yang
• Yingying Wang
• MingFang Xiang
• Xuan Zhang
• Hui Chen

Patients’ behaviors in the Intensive Care Units (ICU) have garnered research attention, particularly regarding the impact of Unplanned Extubation (UEX). However, there is currently no existing report on methods for early warning of UEX action in RGB video. Applying traditional human action recognition algorithms to UEX in the complex ICU environment proves challenging. To address the above issue, we propose a novel feature for early warning of UEX action in patients using RGB videos. Firstly, we employ the YOLOv3 detection method to extract the region of interest (ROI), which corresponds to the region where the patient is located. Subsequently, we develop a spatio-temporal (ST) feature for human action tracking by using the L-K optical flow algorithm. This ST feature encompasses optical flow corner number, trajectory distance, and wavelet transform features. Finally, we utilize support vector machine (SVM) for patient action classification and early warning. Experimental results on the ICU monitoring dataset demonstrate the superior performance of the proposed feature in UEX prediction.

ICML Conference 2024 Conference Paper

Subequivariant Reinforcement Learning in 3D Multi-Entity Physical Environments

• Runfa Chen
• Ling Wang
• Yu Du
• Tianrui Xue
• Fuchun Sun 0001
• Jianwei Zhang 0001
• Wenbing Huang 0001

Learning policies for multi-entity systems in 3D environments is far more complicated against single-entity scenarios, due to the exponential expansion of the global state space as the number of entities increases. One potential solution of alleviating the exponential complexity is dividing the global space into independent local views that are invariant to transformations including translations and rotations. To this end, this paper proposes Subequivariant Hierarchical Neural Networks (SHNN) to facilitate multi-entity policy learning. In particular, SHNN first dynamically decouples the global space into local entity-level graphs via task assignment. Second, it leverages subequivariant message passing over the local entity-level graphs to devise local reference frames, remarkably compressing the representation redundancy, particularly in gravity-affected environments. Furthermore, to overcome the limitations of existing benchmarks in capturing the subtleties of multi-entity systems under the Euclidean symmetry, we propose the Multi-entity Benchmark (MEBEN), a new suite of environments tailored for exploring a wide range of multi-entity reinforcement learning. Extensive experiments demonstrate significant advancements of SHNN on the proposed benchmarks compared to existing methods. Comprehensive ablations are conducted to verify the indispensability of task assignment and subequivariance.

IJCAI Conference 2024 Conference Paper

Trade When Opportunity Comes: Price Movement Forecasting via Locality-Aware Attention and Iterative Refinement Labeling

• Liang Zeng
• Lei Wang
• Hui Niu
• Ruchen Zhang
• Ling Wang
• Jian Li

Price movement forecasting, aimed at predicting financial asset trends based on current market information, has achieved promising advancements through machine learning (ML) methods. Most existing ML methods, however, struggle with the extremely low signal-to-noise ratio and stochastic nature of financial data, often mistaking noises for real trading signals without careful selection of potentially profitable samples. To address this issue, we propose LARA, a novel price movement forecasting framework with two main components: Locality-Aware Attention (LA-Attention) and Iterative Refinement Labeling (RA-Labeling). (1) LA-Attention, enhanced by metric learning techniques, automatically extracts the potentially profitable samples through masked attention scheme and task-specific distance metrics. (2) RA-Labeling further iteratively refines the noisy labels of potentially profitable samples, and combines the learned predictors robust to the unseen and noisy samples. In a set of experiments on three real-world financial markets: stocks, cryptocurrencies, and ETFs, LARA significantly outperforms several machine learning based methods on the Qlib quantitative investment platform. Extensive ablation studies confirm LARA's superior ability in capturing more reliable trading opportunities.

EAAI Journal 2023 Journal Article

A cooperative memetic algorithm for energy-aware distributed welding shop scheduling problem

• Jing-jing Wang
• Ling Wang
• Xia Xiu

Facing globalization trends and sustainable industrial development, energy-aware distributed manufacturing has become an emerging topic. Meanwhile, welding is a kind of indispensable processing in the development of manufacturing and its effective scheduling can improve production efficiency and reduce energy consumption. However, it is difficult to solve the energy-aware distributed welding shop scheduling problem (EADWSP) due to the characteristics of large scale and multiple objectives. Thus, this paper presents a mathematical model and a cooperative memetic algorithm (CMA) to addresses the EADWSP with minimization both makespan and total energy consumption. To improve the quality and diversity of initial population, a hybrid initialization is developed with a modified NEH based heuristic. Via taking full advantage of historical information, a cooperative search based on feedback is designed and a cooperative selection strategy is employed to balance the exploration and exploitation. In addition, multiple problem-specific operators are presented and a local intensification with Q-learning is designed to enhance exploitation capability. Numerical experiments are carried out and the results demonstrate the effectiveness of the above specific designs. The comparisons to the existing algorithms show superiority of the proposed CMA. Moreover, the application to a real-life case also verifies the effectiveness and practicability in solving the EADWSP.

EAAI Journal 2023 Journal Article

Bi-level optimization of charging scheduling of a battery swap station based on deep reinforcement learning

• Mao Tan
• Zhuocen Dai
• Yongxin Su
• Caixue Chen
• Ling Wang
• Jie Chen

With the rapid increase of in the number of electric vehicle (EV), battery swapping is becoming a promising idea because of its short service waiting time. However, in the face of the uncertainty of the power grid and EV behavior, it is difficult to achieve a forward-looking and fast-response scheduling in a large scale battery swap station (BSS). A new bi-level scheduling model is proposed to solve this problem, in which the upper level is built on a deep reinforcement learning (DRL) framework to optimally allocate power among the chargers, and the lower level is modeled as a series of MILP subproblems for dispatching power among the batteries in a charger. A prediction module is included in the DRL framework improve the foresight of the algorithm, and a safety module is designed to avoid unsafe actions. Experimental results indicate that the proposed approach has excellent performance in large scale problem solving. It reduces the operating costs of the BSS significantly while satisfying the maximum power demand constraint. This is able to provide more economic benefits for the BSS and help peak shaving and valley filling for the power grid.

YNIMG Journal 2023 Journal Article

Learning of irrelevant stimulus-response associations modulates cognitive control

• Xiaokai Xia
• Mingqian Guo
• Ling Wang

It has been shown that manipulating the proportion of congruent to incongruent trials in conflict tasks (e.g., Stroop, Simon, and flanker tasks) can vary the size of conflict effects, however, by two different mechanisms. One theory is the control learning account (the brain learns the probability of conflict and uses it to proactively adjust the control demand for future trials). The other is the irrelevant stimulus-response learning account (the brain learns the probability of irrelevant stimulus-response associations and uses it to prepare responses). Previous fMRI studies have detected the brain regions that contribute to the control-learning-modulated conflict effects, but it is less known what neural substrates underlie the conflict effects modulated by irrelevant S-R learning. We here investigated this question with a model-based fMRI study, in which the proportion of congruent to incongruent trials changed dynamically in the Simon task and the models learned the probability of irrelevant S-R associations quantitatively. Behavioral analyses showed that the unsigned prediction errors (PEs) of responses generated by the learning models correlated with reaction times irrespective of congruent and incongruent trials, indicating that large unsigned PEs associated with slow responses. The fMRI results showed that the regions of fronto-parietal and cingulo-opercular network involved in cognitive control were significantly modulated by the unsigned PEs, also irrespective of congruent and incongruent trials, indicating that large unsigned PEs associated with transiently increased activity in these regions. These results together suggest that learning of irrelevant S-R associations modulates reactive control, which demonstrates a new way to modulate cognitive control compared to the control learning account.

ICML Conference 2023 Conference Paper

MetaModulation: Learning Variational Feature Hierarchies for Few-Shot Learning with Fewer Tasks

• Wenfang Sun
• Yingjun Du
• Xiantong Zhen
• Fan Wang
• Ling Wang
• Cees G. M. Snoek

Meta-learning algorithms are able to learn a new task using previously learned knowledge, but they often require a large number of meta-training tasks which may not be readily available. To address this issue, we propose a method for few-shot learning with fewer tasks, which we call MetaModulation. The key idea is to use a neural network to increase the density of the meta-training tasks by modulating batch normalization parameters during meta-training. Additionally, we modify parameters at various neural network levels, rather than just a single layer, to increase task diversity. To account for the uncertainty caused by the reduced number of training tasks, we propose a variational MetaModulation where the modulation parameters are treated as latent variables. We also introduce learning variational feature hierarchies by the variational MetaModulation, which modulates features at all layers and can take into account task uncertainty and generate more diverse tasks. The ablation studies illustrate the advantages of utilizing a learnable task modulation at different levels and demonstrate the benefit of incorporating probabilistic variants in few-task meta-learning. Our MetaModulation and its variational variants consistently outperform state-of-the-art alternatives on four few-task meta-learning benchmarks.

YNICL Journal 2023 Journal Article

Role of hippocampal subfields in neurodegenerative disease progression analyzed with a multi-scale attention-based network

• Hongbo Xu
• Yan Liu
• Ling Wang
• Xiangzhu Zeng
• Yingying Xu
• Zeng Wang

BACKGROUND AND OBJECTIVE: Both Alzheimer's disease (AD) and Parkinson's disease (PD) are progressive neurodegenerative diseases. Early identification is very important for the prevention and intervention of their progress. Hippocampus plays a crucial role in cognition, in which there are correlations between atrophy of Hippocampal subfields and cognitive impairment in neurodegenerative diseases. Exploring biomarkers in the prediction of early cognitive impairment in AD and PD is significant for understanding the progress of neurodegenerative diseases. METHODS: A multi-scale attention-based deep learning method is proposed to perform computer-aided diagnosis for neurodegenerative disease based on Hippocampal subfields. First, the two dimensional (2D) Hippocampal Mapping Image (HMI) is constructed and used as input of three branches of the following network. Second, the multi-scale module and attention module are integrated into the 2D residual network to improve the diversity of the extracted features and capture significance of various voxels for classification. Finally, the role of Hippocampal subfields in the progression of different neurodegenerative diseases is analyzed using the proposed method. RESULTS: Classification experiments between normal control (NC), mild cognitive impairment (MCI), AD, PD with normal cognition (PD-NC) and PD with mild cognitive impairment (PD-MCI) are carried out using the proposed method. Experimental results show that subfields subiculum, presubiculum, CA1, and molecular layer are strongly correlated with cognitive impairment in AD and MCI, subfields GC-DG and fimbria are sensitive in detecting early stage of cognitive impairment in MCI, subfields CA3, CA4, GC-DG, and CA1 show significant atrophy in PD. For exploring the role of Hippocampal subfields in PD cognitive impairment, we find that left parasubiculum, left HATA and left presubiculum could be important biomarkers for predicting conversion from PD-NC to PD-MCI. CONCLUSION: The proposed multi-scale attention-based network can effectively discover the correlation between subfields and neurodegenerative diseases. Experimental results are consistent with previous clinical studies, which will be useful for further exploring the role of Hippocampal subfields in neurodegenerative disease progression.

EAAI Journal 2022 Journal Article

A heuristic and meta-heuristic based on problem-specific knowledge for distributed blocking flow-shop scheduling problem with sequence-dependent setup times

• Fuqing Zhao
• Haizhu Bao
• Ling Wang
• Tianpeng Xu
• Ningning Zhu
• Jonrinaldi

The distributed production scenario with the sequence-dependent setup times (SDST) widely exists in the modern manufacturing system. This paper investigates the distributed blocking flow-shop scheduling problem with sequence-dependent setup times (SDST/DBFSP). Considering the complexity of the distributed scenario and SDSTs, a discrete heuristic and meta-heuristic is proposed by exploring the problem-specific knowledge. First, a knowledge-incorporated construction heuristic is proposed to reduce the blocking times and idle times generated by SDSTs. In the first stage of the meta-heuristic, an insertion-based neighborhood operator of different factories is developed to explore promising regions in the decision space. In the second stage, a local search operator is embedded to enhance the exploitation ability. Additionally, a simulated annealing-like acceptance criterion of the iterated greedy algorithm is employed to keep the diversity of the population. Finally, an insertion operation for critical factories is introduced to further improve the accuracy of the solutions. Moreover, a speedup method for the insertion neighborhood is expanded to reduce the computational complexity of SDST/DBFSP. In the part of the experiment, a deconstruction process is designed to gain insight into the contribution of each component in the proposed meta-heuristic. The proposed meta-heuristic is assessed through comparing with five state-of-the-art algorithms to demonstrate its effectiveness. The experimental results testified that the proposed meta-heuristic outperforms other algorithms regarding the significance of the SDST/DBFSP.

EAAI Journal 2022 Journal Article

A self-learning hyper-heuristic for the distributed assembly blocking flow shop scheduling problem with total flowtime criterion

• Fuqing Zhao
• Shilu Di
• Ling Wang
• Tianpeng Xu
• Ningning Zhu
• Jonrinaldi

The distributed assembly blocking flow shop scheduling problem, which is a significant scenario in modern supply chains and manufacturing systems, has attracted significant attention from researchers and practitioners. To formulate the problem, a mixed-integer linear programming model is introduced to optimize the total flowtime. A constructive heuristic (HHNRa) and a self-learning hyper-heuristic (SLHH) are proposed to address the scheduling problem. HHNRa is designed based on the problem-specific knowledge to obtain initial solutions with high quality. A self-learning high-level strategy based on the historical success rate of low-level heuristics is presented to manipulate the low-level heuristics to operate in the solution space. In addition, a restart scheme with three distinct constructive heuristics is utilized to maintain the diversity of the solution. Based on 900 small-scale benchmark instances and 810 large-scale benchmark instances, comprehensive numerical experiments are conducted to evaluate the performance of the proposed SLHH algorithm. The results of the statistical analysis indicate that the proposed self-learning hyper-heuristic is superior to the compared state-of-the-art algorithms for the problem under consideration. Consequently, the proposed constructive heuristic and the self-learning hyper-heuristic are effective methods for the distributed assembly blocking flow shop scheduling problem.

EAAI Journal 2022 Journal Article

A surrogate-assisted Jaya algorithm based on optimal directional guidance and historical learning mechanism

• Fuqing Zhao
• Hui Zhang
• Ling Wang
• Ru Ma
• Tianpeng Xu
• Ningning Zhu
• Jonrinaldi

An improved Jaya algorithm named surrogate-assisted Jaya algorithm based on optimal directional guidance and historical learning mechanism (SDH-Jaya) is proposed in this study to address the continuous optimization problems. In the SDH-Jaya, a surrogate-assisted model combined with the polynomial model and radial basis model built by the individual with real fitness is introduced to decrease the expensive computational simulations and accelerate the convergence speed. Two co-evolutionary mechanisms, which are named assisted co-evolutionary mechanism and self-learning co-evolutionary mechanism, are proposed to optimize the surrogate model and evolutionary population. Search directions and steps of the SDH-Jaya are adjusted adaptively by the differential vector resulting from the best solution and worst solution in the candidates at each generation. The historical population stored in an archive is selected randomly to provide new search areas for improving the diversity of the population during the evolution process of the SDH-Jaya. The performance of SDH-Jaya is tested on CEC2017 benchmark problems. The experimental results reveal that the effectiveness of the SDH-Jaya algorithm outperforms the classical Jaya algorithm, its variants, and state-of-the-art algorithms in terms of the quality of solution and execution time.

EAAI Journal 2022 Journal Article

Multi-node load forecasting based on multi-task learning with modal feature extraction

• Mao Tan
• Chenglin Hu
• Jie Chen
• Ling Wang
• Zhengmao Li

Accurate multi-node load forecasting is the key to the safe, reliable, and economical operation of the power system. However, the dynamic nature of load and the coupling nature of networks are difficult to extract, making consistent and accurate forecasting of node load rather difficult. In this regard, this paper proposes a soft sharing multi-task deep learning method for multi-node load forecasting in the power system. It has the following aspects: (1) Considering the coupling characteristics of the node network, a multi-modal feature module, based on the inception strategy and gated temporal convolutional network (GTCN), is firstly designed to explore the coupling features implied in the node load data. (2) A novel multi-objective neural network model is proposed to achieve simultaneous prediction of multi-node load by integrating the multi-modal feature module and gated recurrent unit (GRU). For sharing the learning information of sub-networks, this paper uses the soft sharing mechanism to capture load features, which can better optimize the prediction task for each node load simultaneously. Load data from the New Zealand distribution network and AEMO are used to compare the proposed model’s performance in various scenarios using regression metrics such as mean absolute percentage error (MAPE), Weighted Mean Accuracy (WMA), root mean squared logarithmic error (RMSLE), and Diebold–Mariano (DM). The simulation results show that the proposed method can explore the spatial–temporal coupling characteristics in multi-node load data. Compared with existing state-of-the-art multi-node load prediction methods, our proposed method’s MAPE decrease 17. 04% and 3. 92% in Non-aggregation and Aggregation situations.

EAAI Journal 2021 Journal Article

A hierarchical knowledge guided backtracking search algorithm with self-learning strategy

• Fuqing Zhao
• Jinlong Zhao
• Ling Wang
• Jie Cao
• Jianxin Tang

To improve the performance of the backtracking search optimization algorithm (BSA), a multi-population cooperative evolution strategy guided BSA with hierarchical knowledge (HKBSA) is proposed in this paper. According to the domain knowledge of the candidates in objective space, the population is divided into the dominant population, the ordinary population and the inferior population. The information between the sub-populations has interacted with the evolution processes of the three sub-populations. The individuals in the dominant population are maintained as the optimal solutions and are utilized to guide the evolution of the other two sub-populations. A multi-strategy mutation mechanism is applied to solve non-separable problems. The distribution vector of inferior individuals is constructed by sampling, and a mechanism of the individual generation with feedback is proposed by combining self-learning strategy and elite learning strategy. The convergence of HKBSA is analyzed with the Markov model. Compared with the state-of-the-art BSA variants, HKBSA outperforms other algorithms in terms of the speed of convergence, solution accuracy and stability.

YNICL Journal 2021 Journal Article

Abnormal large-scale structural rich club organization in Leber's hereditary optic neuropathy

• Jiahui Zhang
• Ling Wang
• Hao Ding
• Ke Fan
• Qin Tian
• Meng Liang
• Zhihua Sun
• Dapeng Shi

OBJECTIVE: The purpose of this study was to investigate whether the large-scale structural rich club organization was abnormal in patients with Leber's hereditary optic neuropathy (LHON) using diffusion tensor imaging (DTI), and the associations among disrupted brain structural connectivity, disease duration, and neuro-ophthalmological impairment. METHODS: Nineteen acute, 34 chronic LHON patients, and 36 healthy controls (HC) underwent DTI and neuro-ophthalmological measurements. The brain structural network and rich club organization were constructed based on deterministic fiber tracking at the individual level. Then intergroup differences among the acute, chronic LHON patients and healthy controls (HC) in three types of structural connections, including rich club, feeder, and local ones, were compared. Network-based Statistics (NBS) was also used to test the intergroup connectivity differences for each fiber. Several linear and nonlinear curve fit models were applied to explore the associations among large-scale brain structural connectivity, disease duration, and neuro-ophthalmological metrics. RESULTS: Compared to the HC, both the acute and chronic LHON patients had consistently significantly lower fractional anisotropy (FA) and higher radial diffusion (RD) for feeder connections (p 0.05, FDR correction). NBS also identified reduced FA of three feeder connections and five local ones linking visual, auditory, and basal ganglia areas in LHON patients (p 0.05, FDR correction). A significant negative correlation was shown between the retinal nerve fiber layer (RNFL) thickness and disease duration (p < 0.05, FDR correction). CONCLUSIONS: Abnormal rich club organization of the structural network was identified in both the acute and chronic LHON. Furthermore, our findings suggest the coexistence of both primary and secondary connectivity damage in the LHON.

YNICL Journal 2019 Journal Article

Reorganization of the somatosensory pathway after subacute incomplete cervical cord injury

• Qian Chen
• Weimin Zheng
• Xin Chen
• Xuejing Li
• Ling Wang
• Wen Qin
• Kuncheng Li
• Nan Chen

OBJECTIVE: The main purpose of the present study was to investigate the possible somatosensory-related brain functional reorganization after traumatic spinal cord injury (SCI). METHODS: Thirteen patients with subacute incomplete cervical cord injury (ICCI) and thirteen age- and sex-matched healthy controls (HCs) were recruited. Eleven patients and all the HCs underwent both sensory task-related brain functional scanning and whole brain structural scanning on a 3.0 Tesla MRI system, and two patients underwent only structural scanning; the process of structural scanning was completed on thirteen patients, while functional scanning was only applied to eleven patients. We performed sensory task-related functional MRI (fMRI) to investigate the functional changes in the brain. In addition, voxel-based morphometry (VBM) was applied to explore whether any sensory-related brain structural changes occur in the whole brain after SCI. RESULTS: Compared with HCs, ICCI patients exhibited decreased activation in the left postcentral gyrus (postCG), the brainstem (midbrain and right pons) and the right cerebellar lobules IV-VI. Moreover, a significant positive association was found between the activation in the left PostCG and the activation in both the brainstem and the right cerebellar lobules IV-VI. Additionally, the decrease in gray matter volume (GMV) was detected in the left superior parietal lobule (SPL). The decrease of white matter volume (WMV) was observed in the right temporal lobe, the right occipital lobe, and the right calcarine gyrus. No structural change in the primary sensory cortex (S1), the secondary somatosensory cortex (S2) or the thalamus was detected. CONCLUSION: These functional and structural findings may demonstrate the existence of an alternative pathway in the impairment of somatosensory function after SCI, which consists of the ipsilateral cerebellum, the brainstem and the contralateral postCG. It provides a new theoretical basis for the mechanism of sensory-related brain alteration in SCI patients and the rehabilitation therapy based on this pathway in the future.

EAAI Journal 2015 Journal Article

Intelligent virtual reference feedback tuning and its application to heat treatment electric furnace control

• Ling Wang
• Haoqi Ni
• Ruixin Yang
• Panos M. Pardalos
• Li Jia
• Minrui Fei

Virtual Reference Feedback Tuning (VRFT) is a data-driven one-shot control method which is very attractive for engineering applications. However, it cannot design controllers with the optimal control performance based on the standard VRFT approach as performance indices are not explicitly represented in its objective function. To deal with this problem, this paper presents a novel intelligent VRFT (IVRFT) based on adaptive binary ant system harmony search (ABASHS) where the reference model of VRFT, which potentially determines the control performance, is coordinately optimized with the controller by ABASHS to achieve the best control performance. Finally, the proposed ABASHS-based intelligent virtual reference feedback tuning (ABASHS-IVRFT) method is applied to the temperature control of the heat treatment electric furnace. The simulation results demonstrate that ABASHS-IVRFT is valid and can implement the optimal non-overshoot control easily and efficiently. Considering the characteristics such as ease of implementation and no need of the model information of controlled objects, ABASHS-IVRFT is a promising approach for engineering applications.

JMLR Journal 2015 Journal Article

Supervised Learning via Euler's Elastica Models

• Tong Lin
• Hanlin Xue
• Ling Wang
• Bo Huang
• Hongbin Zha

This paper investigates the Euler's elastica (EE) model for high-dimensional supervised learning problems in a function approximation framework. In 1744 Euler introduced the elastica energy for a 2D curve on modeling torsion-free thin elastic rods. Together with its degenerate form of total variation (TV), Euler's elastica has been successfully applied to low- dimensional data processing such as image denoising and image inpainting in the last two decades. Our motivation is to apply Euler's elastica to high-dimensional supervised learning problems. To this end, a supervised learning problem is modeled as an energy functional minimization under a new geometric regularization scheme, where the energy is composed of a squared loss and an elastica penalty. The elastica penalty aims at regularizing the approximated function by heavily penalizing large gradients and high curvature values on all level curves. We take a computational PDE approach to minimize the energy functional. By using variational principles, the energy minimization problem is transformed into an Euler-Lagrange PDE. However, this PDE is usually high-dimensional and can not be directly handled by common low-dimensional solvers. To circumvent this difficulty, we use radial basis functions (RBF) to approximate the target function, which reduces the optimization problem to finding the linear coefficients of these basis functions. Some theoretical properties of this new model, including the existence and uniqueness of solutions and universal consistency, are analyzed. Extensive experiments have demonstrated the effectiveness of the proposed model for binary classification, multi-class classification, and regression tasks. [abs] [ pdf ][ bib ] &copy JMLR 2015. ( edit, beta )

EAAI Journal 2014 Journal Article

MBPOA-based LQR controller and its application to the double-parallel inverted pendulum system

• Ling Wang
• Haoqi Ni
• Weifeng Zhou
• Panos M. Pardalos
• Jiating Fang
• Minrui Fei

As the performance of Linear Quadratic Regulator (LQR) controllers greatly depends on its weighting matrices, i. e. Q and R, designing these two matrices is one of the most important components in the LQR problem which is a tedious and challenging work in the applications of LQR. Hence, a novel LQR approach based on the Pareto-based Multi-objective Binary Probability Optimization Algorithm (MBPOA) is proposed in this paper, in which MBPOA is utilized to search for the optimal weighting matrices to relieve the effort of parameter settings and improve the control performance according to the pre-defined objective functions. By combining LQR with MBPOA, the optimal controllers can be obtained easily and effortless. Moreover, the control performance can be adjusted further conveniently to meet the requirements of applications as a set of Pareto-optimal LQR controllers is offered. The simulation and experiment results on the double-parallel inverted pendulum system demonstrate the effectiveness and efficiency of the developed MBPOA-based LQR method. Considering the characteristics such as robustness, the optimal dynamic performance and easy implementation without prior knowledge, the MBPOA-based LQR is a promising control approach for engineering applications.

ICML Conference 2012 Conference Paper

Total Variation and Euler's Elastica for Supervised Learning

• Tong Lin 0002
• Hanlin Xue
• Ling Wang
• Hongbin Zha

EAAI Journal 2007 Journal Article

An effective co-evolutionary particle swarm optimization for constrained engineering design problems

• Qie He
• Ling Wang

Many engineering design problems can be formulated as constrained optimization problems. So far, penalty function methods have been the most popular methods for constrained optimization due to their simplicity and easy implementation. However, it is often not easy to set suitable penalty factors or to design adaptive mechanism. By employing the notion of co-evolution to adapt penalty factors, this paper proposes a co-evolutionary particle swarm optimization approach (CPSO) for constrained optimization problems, where PSO is applied with two kinds of swarms for evolutionary exploration and exploitation in spaces of both solutions and penalty factors. The proposed CPSO is population based and easy to implement in parallel. Especially, penalty factors also evolve using PSO in a self-tuning way. Simulation results based on well-known constrained engineering design problems demonstrate the effectiveness, efficiency and robustness on initial populations of the proposed method. Moreover, the CPSO obtains some solutions better than those previously reported in the literature.

AIIM Journal 2005 Journal Article

Factors affecting automated syndromic surveillance

• Ling Wang
• Marco F. Ramoni
• Kenneth D. Mandl
• Paola Sebastiani

Objective: The increased threat of bioterroristic attacks and epidemic events requires the development of accurate and timely outbreak detection systems for early identification of anomalies in public health data. Material and methods: We propose an automated outbreak detection system based on syndromic data. This system uses an autoregressive model with seasonal components to monitor, online, the daily counts of chief complaints for respiratory syndromes at the emergency department of two major metropolitan hospitals. We evaluate this system by estimating the false positive rate in real data under the assumption that there were no outbreaks of disease, and the true positive rate in real baseline data in which we injected stochastically simulated outbreaks of different shape and size. We then use directed graphical models to account for the effect of exogenous factors on the detection performance of the system. Results: Our study shows that for a week-long outbreak, our model has an overall 84. 8% true detection accuracy across all shapes of outbreaks, while the outbreak size influences the earliness to detection. The false and true positive rates are also associated with the exogenous factors and knowledge about these factors can help to improve the detection accuracy. Conclusion: This study suggests that the integration of multiple data sources can significantly improve the detection accuracy of syndromic surveillance systems.