I am now a Ph.D. Candidate at The Chinese University of Hong Kong (CUHK), Department of Computer Science and Engineering from 2022, advised by Prof. Farzan Farnia.
I obtained my Bachelor’s Degree in Mathematics and Information Engineering, a double-major program jointly offered by the Department of Mathematics and the Department of Information Engineering at The Chinese University of Hong Kong (CUHK), within the Honor Class ELITE Stream. It’s my pleasure to be kindly advised by Prof. Chandra Nair and Prof. Anthony MC So.
My research interests include Machine Learning Theory, Trustworthy AI (Robustness, Fairness, and Interpretability) and Diffusion-based Generative Models. I seek relevant applications in Computer Vision, Decision Making, and Planning.
🔥 News
- 2025.11: 🎉 Our paper is accepted by AAAI 2026!
- 2025.09: 🎉 Two papers are accepted by NeurIPS 2025!
- 2025.07: 🎉 Our paper is accepted by ACMMM 2025!
- 2025.06: 🎉 Join Huawei Hong Kong Research Center, Theory Lab as a research intern!
- 2025.06: 🎉 Honored to be selected as the Outstanding Intern in Noah’s Ark Lab!
- 2025.01: 🎉 Our paper is accepted by AAAI 2025 GenPlan Workshop (Spotlight)!
- 2025.01: 🎉 Our paper is accepted by ICLR 2025!
- 2024.07: 🎉 Join Huawei Noah’s Ark Lab as a research intern!
- 2024.04: 🎉 Our paper is accepted by UAI 2024!
- 2022.08: 🎉 Start the journey as a Ph.D. student at CSE, CUHK!
- 2021.05: 🎉 Our paper is accepted by ICASSP 2021!
📖 Educations
- 2022.08 - 2026.07 (now): Ph.D. Candidate in Computer Science and Engineering Department, CUHK, Hong Kong SAR.
- 2017.09 - 2022.07: Bachelor of Science in Mathematics and Information Engineering (double major), CUHK, Hong Kong SAR.
💻 Experiences
- 2025.06 - Now: Research intern, Huawei Hong Kong Research Center, Theory Lab, Hong Kong SAR, China
- 2024.07 - 2025.01: Research intern, Huawei Noah’s Ark Lab, Decision Making & Reasoning Group, Shenzhen, China
- 2021.08 - 2022.04: Research Assistant, advised by Prof. Anthony MC So, Department of SEEM, CUHK, Hong Kong SAR
- 2021.05 - 2021.08：Research Assistant, advised by Prof. Eliza Mik, Department of Law, CUHK, Hong Kong SAR
- 2020.08 - 2020.12: Research Assistant, advised by Prof. Anthony MC So, Department of SEEM, CUHK, Hong Kong SAR
- 2019.01 - 2019.03: Financial Internship, Guotai Junan Securities, Hunan Changsha
📝 Publications
Two-Steps Diffusion Policy for Robotic Manipulation via Genetic Denoising
Mohammad Jalali*, Haoyu Lei*, Amin Gohari, Farzan Farnia (*Equal Contribution)
- We propose the Scalable Prompt-Aware Renyi Kernel Entropy Diversity Guidance (SPARKE) method to improve prompt-aware diversity in text-to-image diffusion models, allowing large-scale generation settings with linear inference computational overhead.
SPARKE: Scalable Prompt-Aware Diversity and Novelty Guidance in Diffusion Models via RKE Scores
Mohammad Jalali*, Haoyu Lei*, Amin Gohari, Farzan Farnia (*Equal Contribution)
- We propose the Scalable Prompt-Aware Renyi Kernel Entropy Diversity Guidance (SPARKE) method to improve prompt-aware diversity in text-to-image diffusion models, allowing large-scale generation settings with linear inference computational overhead.
Boosting Generalization in Diffusion-Based Neural Combinatorial Solver via Energy-guided Sampling
Haoyu Lei, Kaiwen Zhou, Yinchuan Li, Zhitang Chen, Farzan Farnia
Paper/[Code]
- We propose a general energy-guided sampling framework during inference time that enhances both the cross-scale and cross-problem generalization capabilities of diffusion-based NCO solvers without requiring additional training.
MESH - Understanding Videos Like Human: Measuring Hallucinations in Large Video Models
Garry Yang, Zizhe Chen, Man Hon Wong, Haoyu Lei, Yongqiang Chen, Zhenguo Li, Kaiwen Zhou, James Cheng
Paper/[Code]
- We introduce MESH, a benchmark designed to evaluate hallucinations in LVMs systematically. MESH uses a Question-Answering framework with binary and multi-choice formats, incorporating target and trap instances.
pFedFair: Towards Optimal Group Fairness-Accuracy Trade-off in Heterogeneous Federated Learning
Haoyu Lei, Shizhan Gong, Qi Dou, Farzan Farnia
Paper/[Code]
- We propose the pFedFair to address personalized federated learning under heterogeneous feature distribution, while applying image embeddings to improve the performance of group-fairness learning methods on computer vision datasets.
Boosting the visual interpretability of CLIP via adversarial fine-tuning
Shizhan Gong, Haoyu Lei, Qi Dou, Farzan Farnia
Paper/[Code]
- We propose an unsupervised adversarial fine-tuning (AFT) with norm-regularization to enhance the visual interpretability of CLIP.
On the Inductive Biases of Demographic Parity-based Fair Learning Algorithms
Haoyu Lei, Amin Gohari, Farzan Farnia
- We theoretically show that Inductive Biases widely exist in DP-based fair supervised learning algorithms, and further propose a distributionally robust optimization framework to mitigate the biases.
An Efficient Alternating Direction Method For Graph Learning From Smooth Signals
Xiaolu Wang, Chaorui Yao, Haoyu Lei, Anthony Man-Cho So
- We cast the graph learning formulation as a nonsmooth, strictly convex optimization problem and develop an efficient alternating direction method of multipliers (ADMM) to solve it.
🎖 Honors and Awards
- 2025: Huawei Noah’s Ark Lab Outstanding Intern
- 2022: Postgraduate Studentship Scholarship Scheme
- 2021: Chung Chi College Departmental Prize (Graduation Top 10%)
- 2019, 2021: Dean’s List (Academic Year Top 10%)
- 2019 - 2021: Chung Chi College Class Scholarship (College Top 10%)
- 2019 - 2021: ELITE Stream Scholarship
- 2017: Top 0.15% in Hunan Province’s Gaokao (University Entrance Exam)
⌨️ Languages and Skills
- Languages: English (Advanced), Cantonese (Intermediate), Putonghua (Native)
- Programming Skills: Python (Main), Pytorch (Main), Matlab, R, C, C++, C#, Julia, LaTeX
💬 Teaching
- 2025-26 Fall: AIST3010/ESTR3114 Numerical Optimization (Undergraduate)
- 2024-25 Spring: CSCI5130 Game Theory (Postgraduate)
- 2023-24 Spring: CSCI5030 Machine Learning Theory (Postgraduate)
- 2023-24 Fall: ENGG2760A Probability for Engineers (Undergraduate)
- 2022-23 Spring: ENGG1003SG Digital Literacy and Computational Thinking (Undergraduate)
- 2022-23 Fall: AIST3010 Numerical Optimization (Undergraduate)