Binghui Wu

Binghui Wu

Auto-generated profile, based on publicly available information

Binghui Wu

Department of Electrical and Computer Engineering, National University of Singapore, Queenstown, Singapore

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

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

Representative Publications

View All Publications
2 July 2025 in IEEE Transactions on Cognitive Communications and Networking
Verified
Journal Article
Closed access
BWBinghui Wu
Binghui Wu
DDDinil Mon Divakaran
Dinil Mon Divakaran
MGMohan Gurusamy
Mohan Gurusamy
As modern networks grow increasingly complex—driven by diverse devices, encrypted protocols, and evolving threats—network traffic analysis has become critically important. Existing machine learning models often rely only on a single representation of packets or flows, limiting their ability to capture the contextual relationships essential for robust analysis. Furthermore, task-specific architectures for supervised, semi-supervised, and unsupervised learning lead to inefficiencies in adapting to varying data formats and security tasks. To address these gaps, we propose UniNet, a unified framework that introduces a novel multi-granular traffic representation (T-Matrix) with rich contextual information, integrating session, flow, and packet-level features to provide comprehensive contextual information. Combined with T-Attent, a specially designed lightweight attention-based model, UniNet efficiently learns latent embeddings for diverse security tasks. Extensive evaluations across four key network security and privacy problems—anomaly detection, attack classification, IoT device identification, and encrypted website fingerprinting—demonstrate UniNet’s significant performance gain over state-of-the-art methods, achieving higher accuracy, lower false positive rates, and improved scalability across various datasets. By addressing the limitations of single-level models and unifying traffic analysis paradigms, UniNet sets a new benchmark for modern network security.

13 April 2026 in IEEE Transactions on Cognitive Communications and Networking
Verified
Journal Article
Closed access
BWBinghui Wu
Binghui Wu
PAPalleri Chettuparambil Amogh
Palleri Chettuparambil Amogh
NANalam Venkata Abhishek
Nalam Venkata Abhishek
MGMohan Gurusamy
Mohan Gurusamy
The rapid evolution of 5G and beyond-5G (B5G) networks, coupled with the emergence of network slicing, has introduced unprecedented challenges in resource allocation for Radio Access Networks (RAN). Existing methods primarily fall into two categories: i) reactive approaches, which fail to anticipate sudden traffic changes; ii) predictive methods, which struggle to adapt to real-time conditions. These limitations result in inefficient resource utilization, degraded Quality of Service (QoS), and increased latency. To address these challenges, we propose PreNS, a hybrid framework that integrates predictive modeling with real-time decision-making to optimize resource allocation. PreNS combines an attention-based Bi-LSTM model for accurate traffic demand forecasting with a Double Deep Q-Network (DDQN) for real-time adjustments. Its key innovation lies in the aggregated state representation, which fuses predicted demands and real-time data, enabling adaptive and efficient resource management. By balancing proactive foresight and reactive adaptability, PreNS addresses fluctuating network conditions and the rigidity of static optimization. Experimental evaluations on a realistic 5G testbed demonstrate that PreNS improves resource utilization by approximately 14%, reduces latency by 20%, and enhances QoS satisfaction by 5%, compared to state-of-the-art methods—demonstrating its potential for efficient, scalable resource allocation in future wireless networks.