Binghui (Alan) Wang

Assistant Professor
Department of Computer Science
Illinois Institute of Technology
Email: bwang70@illinoistech.edu
Office: Stuart Building, 216C, 10 W 31st St, Chicago, IL 60616
Research areas: Trustworthy AI, Data-Driven Security and Privacy, and AI/Data Science
Site Co-PI: Chicago-area IDEAL Institute
IEEE Senior Member

Openings: I'm always looking for highly motivated Ph.D. students, visiting scholars/students, and research interns to join my group. If you are interested, please send an email to alanwbh@gmail.com with your CV and transcripts attached.

About Me

I have been an Assistant Professor in the Department of Computer Science at Illinois Institute of Technology (Illinois Tech) since August 2021. I am interested in Trustworthy AI and Security & Privacy.

I did my postdoc at Duke University from August 2019 to July 2021, working with Neil Gong and Yiran Chen. I earned my Ph.D. in May 2019 from Iowa State University, where I was advised by Neil Gong. I received both my M.Sc. (2015) and B.E. (2012) from Dalian University of Technology, China, graduating with the highest honor (Qu Bochuan Scholarship).

I am a recipient of the NSF CAREER (2024) and CRII (2023) Awards, as well as research awards from Cisco (2022) and Amazon (2020). My work has received best paper awards at CCS’24 and CVPRW’20 and an honorable mention at NDSS’19. I was named one of Baidu Scholar’s Global Top 50 Chinese Rising Stars in AI + X (2022), received the Dean’s Excellence in Research Award (2024), and was selected for Stanford University’s World’s Top 2% Scientists List (2024, 2025).

Research Areas and Recent/Selected Publications

Trustworthy LLM/GenAI

Security Attacks

[Security’25a] Knowledge Corruption Attacks to RAG of LLMs

Trustworthy Representation Learning

[AAAI’25a] Robust and Privacy-Preserving Representation Learning
[Security’24] Privacy-Preserving Representations Learning against Inference Attacks
[AAAI’24] Privacy-Preserving Representation Learning for FL Against Attribute Inference Attacks
[KDD’21a] Privacy-Preserving Representation Learning on Graphs

Trustworthy (Explainable) Graph Learning

Security and Privacy Attacks

[AAAI’25b] Black-box Attacks on Dynamic GNNs for Link Prediction
[ICML’24] Black/Gray-box Attacks on Explainable GNNs
[CVPR’23a] Certified Robustness Inspired Attack Framework on GNNs
[CVPR’22] Score-based Black-box Attacks on GNNs with Guaranteed Performance
[CCS’21] Hard-label Black-box Attacks on GNNs
[SACMAT’21] Backdoor Attacks on GNNs
[CCS’19] Attacking Graph-based Collective Classification

Security and Privacy Defenses

[CVPR’25] Provably Robust GNNs against against Poisoning Attacks with Arbitrary Perturbations
[Security’25b] Provably Robust GNNs against Arbitrary Perturbations with Deterministic Guarantees
[ICLR’25] Provably Robust Explainable GNNs against Graph Perturbations with Deterministic Guarantees
[CCS’24a] Provably Robust FedGL against Distributed Backdoor Attacks with Deterministic Guarantees
[ICLR’24] Provably Robust GNNs against Graph Perturbations with Deterministic Guarantees
[KDD’21b] Provably Robust GNNs against Graph Perturbations with Probabilistic Guarantees
[WWW’20] Provably Robust Community Detection against Graph Perturbations with Probabilistic Guarantees

[KDD’21a] Privacy-Preserving Representation Learning on Graphs

Trustworthy Federated Learning

Security and Privacy Attacks

[CCS’24a] Distributed Backdoor Attacks on FedGL
[CIKM’24a] Optimization-based Attack on SOTA Poisoning Defenses to FL

[CVPR’21] Privacy Leakage in FL from Representation Perspective

Security and Privacy Defenses

[CCS’24a] Provably Robust FedGL against Distributed Backdoor Attacks

[CODASPY’25] Universally Harmonizing DP Mechanisms for FL: Boosting Accuracy and Convergence
[NeurIPS’24] Provably Robust Watermarking for FedGL
[AAAI’24] Privacy-Preserving FL Against Attribute Inference Attacks
[CVPR’21] Privacy-Preserving FL against Data Reconstruction Attacks

Trustworthy Deep Learning

Security and Privacy Attacks

[Security’25c]] Rectifying the Privacy and Efficacy Measurements for Machine Unlearning
[CCS’24b] Certifiable Black-Box Attacks with Randomized Adversarial Examples
[EuroSP’23] Certified Radius-Guided Attack Framework to Image Segmentation Models
[NeurIPS’20] Perturbing Feature Hierarchy to Improve Blackbox Attack Transferability

[SP’24b] Stealing DNN Model Architectures through Power Side Channel
[AsiaCCS’21] Robust and Verifiable Information Embedding Attacks to DNNs
[SP’18] Stealing Hyperparameters in Machine Learning

Security and Privacy Defenses

[SP’24a] Generalized Certified Robustness against Textual Adversarial Attacks
[ECCV’22] Universally Approximated Certified Robustness
[ICLR’22] Certified Robustness of Top-k Predictions against L0-Perturbation
[ICLR’20] Certified Robustness for Top-k Predictions against L2-Perturbation
[CVPRW’20] Certified Robustness against Backdoor Attacks

[AAAI’25a] Learning Robust and Privacy-Preserving Representations
[Security’24] Learning Privacy-Preserving Representations against Inference Attacks
[ACSAC’22] Lightweight Neuron-Guided Defense against MIAs

Graph-based Machine Learning for Security and Privacy

Graph-based ML for Security

[KDD’21c] An Unsupervised Approach for Unveiling Fake Accounts
[ACSAC’21] Detecting Growing-Up Behaviors of Malicious Accounts
[NDSS’19] Graph-based Security and Privacy Analytics
[RAID’18] Sybil Detection in Social Web Services without Manual Labels
[ICDM’17] Sybil Detection in Directed Social Networks via Guilt-by-Association
[DSN’17] Random Walk based Fake Account Detection in Online Social Networks
[INFOCOM’17] Sybil Detection in Online Social Networks via Local Rule based Propagation

Graph-based ML for Privacy

[TDSC’25] Personalized Photo-Sharing and Automatic Deletion in Social Networks
[AsiaCCS’22] A Graph-based Cross-Device Tracking Framework
[NDSS’19] Graph-based Security and Privacy Analytics
[WWW’17] Inferring User Attributes in Online Social Networks Using Markov Random Fields

Explainable Machine Learning

[ECCV’24] GNN Causal Explanation via Neural Causal Models
[CIKM’24b] An Information Propagation Approach for Improving DNN Model Explanations
[CVPR’23b] A Framework to Eliminate Explanation Noise from Integrated Gradients

Binghui (Alan) Wang

About Me

Recent News

Students

Ph.D. Student

Ph.D. Alumni

Master Student

Research Intern

Teaching

Research Areas and Recent/Selected Publications

Trustworthy LLM/GenAI

Security Attacks

Trustworthy Representation Learning

Trustworthy (Explainable) Graph Learning

Security and Privacy Attacks

Security and Privacy Defenses

Trustworthy Federated Learning

Security and Privacy Attacks

Security and Privacy Defenses

Trustworthy Deep Learning

Security and Privacy Attacks

Security and Privacy Defenses

Graph-based Machine Learning for Security and Privacy

Graph-based ML for Security

Graph-based ML for Privacy

Explainable Machine Learning

Other Topics

Federated Learning (FL) & Graph Learning (GL)

Prototype Learning & Single Sample per Person (SSPP)

Myoelectric Control & Brain-Computer Interface