Yongcan Luo

Yongcan Luo

Auto-generated profile, based on publicly available information

Yongcan Luo

Department of Computer Science, City University of Hong Kong, Hong Kong SAR, China

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
10 August 2023 in IEEE Internet of Things Journal
Verified
Journal Article
Closed access
SQShuxin Qin
Shuxin Qin
LCLin Chen
Lin Chen
YLYongcan Luo
Yongcan Luo
GTGaofeng Tao
Gaofeng Tao
Internet of Things (IoT) systems typically generate large amounts of sensory signals that get involved to represent the states of the systems. Most existing methods focus on learning the temporal patterns of the signals to detect anomalies. However, the performance is limited due to two critical problems. First, the relationships between different signals are rarely considered, resulting in missing important information for representation. Second, the high sensitivity of time series makes it difficult to use conventional methods for data augmentation, which limits the improvement of representation and generalization capabilities. In this work, we propose a novel reconstruction-based framework with contrastive learning from multiple views to address these two issues. Specifically, intra-signal and inter-signal graph structures are formed and learned in parallel to model the temporal context and capture the dependency relationships between signals, respectively. Multi-view graph contrastive learning strategy is designed to improve the graph representation. We also provide an adaptive data augmentation method to generate graph views for contrastive learning, which helps to accurately capture valuable intrinsic patterns from two different perspectives. Finally, the contrastive learning task and the reconstruction task are jointly trained. Extensive experiments on four real-world datasets demonstrate that our method outperforms the existing state-of-the-art baselines.

13 August 2025 in IEEE Transactions on Network Science and Engineering
Verified
Journal Article
Closed access
YLYongcan Luo
Yongcan Luo
JCJingxuan Chen
Jingxuan Chen
SQShuxin Qin
Shuxin Qin
DWDapeng Oliver Wu
Dapeng Oliver Wu
In Internet of Underwater Things (IoUT) systems, the unpredictable underwater environment and unstable transmission conditions often lead to anomalies in the generated sensory signals. While numerous anomaly detection methods have emerged, many fail to effectively address the challenges posed by underwater settings, particularly the temporal dependencies and inter-signal relationships for accurate anomaly detection. To overcome these limitations, we propose a novel reconstruction-based framework incorporating a dual graph learning strategy tailored for underwater scenarios. Specifically, our method simultaneously learns intra-signal graph structures to capture temporal patterns and inter-signal graph structures to model dependencies between sensor signals. To enhance the fusion of these graph branches, we introduce a parametric matrix fusion module that dynamically optimizes the influence of each graph through training. Furthermore, a multi-head attention-based decoder is employed to efficiently process the fused features and improve decoding performance. We validated our method through extensive experiments on real-world datasets and simulate different error rates under varying transmission conditions, and the results demonstrate that our approach maintains high anomaly detection accuracy even in the presence of higher bit error rates underwater settings.

17 December 2024 in IEEE Transactions on Computational Social Systems
Verified
Journal Article
Closed access
SQShuxin Qin
Shuxin Qin
YLYongcan Luo
Yongcan Luo
JZJing Zhu
Jing Zhu
GTGaofeng Tao
Gaofeng Tao
JZJingya Zheng
Jingya Zheng
ZMZhongjun Ma
Zhongjun Ma
Detecting abnormal nodes from attributed networks plays an important role in various applications, including cybersecurity, finance, and social networks. Most existing methods focus on learning different scales of graphs or using augmented data to improve the quality of feature representation. However, the performance is limited due to two critical problems. First, the high sensitivity of attributed networks makes it uncontrollable and uncertain to use conventional methods for data augmentation, leading to limited improvement in representation and generalization capabilities. Second, under the unsupervised paradigm, anomalous nodes mixed in the training data may interfere with the learning of normal patterns and weaken the discrimination ability. In this work, we propose a novel multiview and multiscale contrastive learning framework to address these two issues. Specifically, a network augmentation method based on parameter perturbation is introduced to generate augmented views for both node–node and node–subgraph level contrast branches. Then, cross-view graph contrastive learning is employed to improve the representation without the need for augmented data. We also provide a cycle training strategy where normal samples detected in the former step are collected for an additional training step. In this way, the ability to learn normal patterns is enhanced. Extensive experiments on six benchmark datasets demonstrate that our method outperforms the existing state-of- the-art baselines.

5 August 2025 in IEEE Internet of Things Journal
Verified
Journal Article
Closed access
JCJingxuan Chen
Jingxuan Chen
TSTianli Shi
Tianli Shi
YLYongcan Luo
Yongcan Luo
PYPeng Yang
Peng Yang
DWDapeng Oliver Wu
Dapeng Oliver Wu
The Internet of Underwater Things (IoUT) connects underwater devices for applications like environmental monitoring, marine exploration, and disaster response. Accurate localization of acoustic sources is vital to IoUT, enabling tasks such as sensor deployment and vehicle navigation. The Time Difference of Arrival (TDOA) method, which uses differences in signal arrival times at multiple receivers, is widely employed for this purpose. However, traditional TDOA approaches assume a constant sound speed, overlooking depth-dependent variations caused by changes in salinity, temperature, and pressure. Additionally, multipath propagation complicates localization, as the first received signal may include reflections rather than the direct path. This paper presents a novel synchronization-free localization method, Synchronized-Transmission TDOA (ST-TDOA), tailored for IoUT environments. The proposed method eliminates the need for clock synchronization among receivers, while addressing sound speed variability and mitigating multipath effects. A dynamic model is developed to adaptively adjust for sound speed changes, and a synchronized transmission algorithm ensures accurate TDOA measurements, reducing errors caused by clock discrepancies. Simulation results demonstrate significant improvements in localization accuracy and reliability, highlighting the effectiveness of ST-TDOA in addressing critical challenges in underwater localization for IoUT systems.

1 May 2025 in IEEE Transactions on Neural Networks and Learning Systems
Verified
Journal Article
Closed access
YLYongcan Luo
Yongcan Luo
JZJiahao Zheng
Jiahao Zheng
Zhengjie YangZhengjie Yang
Zhengjie Yang
Ning ChenNing Chen
Ning Chen
DWDapeng Oliver Wu
Dapeng Oliver Wu
Predicting stock trends is a highly rewarding but high-risk endeavor due to the complex interplay of market dynamics, irrational behaviors, and diverse sentiments. Previous studies have used time-series analysis on historical prices or sentiment analysis on textual information. However, these methods often fail to capture the dynamic interactions between text and time-series modalities and overlook the different perspectives embedded in textual data. To address these limitations, we propose the pleno-alignment framework (PAFrame) that enhances multimodal stock information through intermodal and intramodal alignment to capture market dynamics. Our framework first integrates textual and time-series data in a shared representation space to learn modal-invariant information. To tackle divergent sentiments in textual data, we employ a contrastive learning approach to extract abstract semantic meanings from objective and subjective perspectives, thereby improving the robustness of language representations. Finally, we use a hybrid approach that explicitly combines cross-attention mechanisms to create a unified representation and utilizes prompts to implicitly guide language models with numerical financial indicators for final prediction. Our comprehensive experiments on five real-world datasets show that PAFrame outperforms existing methods in predicting stock trends.

12 May 2026 in IEEE Transactions on Network Science and Engineering
Verified
Journal Article
Closed access
YLYongcan Luo
Yongcan Luo
JZJiahao Zheng
Jiahao Zheng
SQShuxin Qin
Shuxin Qin
TCTianyi Chen
Tianyi Chen
DWDapeng Oliver Wu
Dapeng Oliver Wu
Multi-cloud networking enables enterprises to orchestrate applications across heterogeneous cloud platforms, offering improved scalability, resilience, and cost-effectiveness. However, complex interactions among distributed cloud resources introduce intricate interdependencies, particularly through inter-VM traffic flows and shared network pathways. Current forecasting methods for system metrics often overlook these fine-grained traffic dynamics and the causal relationships inherent in multi-cloud environments, failing to capture how resource usage in one VM can influence others. In this paper, we propose a causality-aware dual-graph neural network designed explicitly for multivariate system metric forecasting in multi-cloud settings. Our framework employs an adaptive dual-graph architecture to simultaneously model spatial dependencies among metrics across different VMs and temporal dependencies within individual VM metrics. To explicitly encode directional causal relationships, we construct a Transfer Entropy-based causality matrix optimized through Bayesian search. A causally guided fusion module integrates these representations using contrastive and triplet learning, aligning latent spaces to emphasize causality-aware structures. Extensive experiments on real-world multi-cloud traces demonstrate that our approach consistently outperforms strong spatio-temporal and Transformer-based forecasting baselines, achieving relative error reductions of at least 8% (across MAE/RMSE/MAPE) in the 15-VM setting.