Short Bio

I am a tenure-track Assistant Professor in Computer Science at the School of Computing and Information Sciences (SCIS), Singapore Management University (SMU). Previously, I was a postdoctoral researcher at the department of computer science at TU Kaiserslautern in Prof. Dr. Marius Kloft's research group. I hold a Bachelor's and a Masters' degree in mathematics from Clare College, University of Cambridge. Previously, I worked on low-dimensional projections of high-dimensional stochastic differential equations at the University of Luxembourg (Mathematics Department), where I obtained my PhD in 2017.

My main research areas are  Statistical Learning Theory and Recommender Systems. In learning theory, I have provided generalization bounds for deep convolutional neural networks, pairwise and contrastive learning, as well as advanced the state-of-the-art in the statistical theory of Matrix Completion in the approximate recovery case.

Positions

I am actively looking for highly motivated PhD students at SMU. A good Bachelor's or Master's degree in Computer Science, Mathematics, or a related discipline is required. I am also happy to accept suitable CSC-funded visiting PhD students. Drop me an email!

SMU is a young and dynamic university on a clear upward trajectory and is ranked No. 41 globally (No. 16 in Asia) in the broad AI category and No. 85 globally (No. 15 in Asia) in the overall "Computer Science" discipline according to CSRankings.  SMU faculty members and PhD students frequently publish multiple papers at top-tier conferences such as NeurIPS, ICML and ICLR. 

Singapore is a vibrant, international city of 5-6 million with a rich cultural life at the crossroads of Asian and Western cultures. There are many universities and AI startups around, making it the main research hub for Artificial Intelligence in the Asia-Pacific region. 

I have available projects in Statistical Learning Theory, Matrix Completion or Recommender Systems, for either a mathematics-centred candidate with basic programming skills or a computer science-centred candidate with decent mathematical skills. The projects/topics can also be adapted to fit the potential students' strengths and interests, as long as it is not too far away from my existing and future research directions. 

The positions are fully funded for four years, with a possibility to extend to 5  with a teaching load allocation. There is a strong possibility to upgrade to more competitive scholarships, including our Presidential Scholarship (which comes with a highly competitive stipend) based on performance. The standard for acceptance to the Presidential Scholarship is high but still achievable. One of my current students, Hieu Nong, has obtained the prestigious AISG Scholarship (stipend above 6000 SGD a month) based on his early performance working with me. Further information regarding funding opportunities is available at: Funding Opportunities.

News

[06/25]  I have been nominated for the "Most Promising Teacher " award at SMU. I am very grateful to my students for their positive and constructive feedback, and thrilled to have been able to make a positive impact on their learning journey. Looking forward to a fruitful new academic term!

[06/25] I honoured to have been named as an "Outstanding Meta Reviewer (Area Chair)" at ICML 2025, a distinction awarded to only 19 people this year (top ≤ 2 percent).  ICML Programme Committee. 

[06/25] Congratulations to my PhD student, Hieu Nong, for obtaining the prestigious AISG Scholarship. Looking forward to our continued successful collaboration!

[05/25] Our paper entitled "Generalization Analysis for Contrastive Representation Learning under Non-IID Settings" has been accepted at ICML 2025.  Congratulations to PhD student Hieu Nong Minh! Full Text (arXiv) 

[02/25]  I am honoured to have been offered the opportunity to serve as Area Chair at NeurIPS 2025.

[12/24]  I am honoured to have been offered the opportunity to serve as Area Chair at ICML 2025.

[12/24] Two papers accepted at AAAI 2025: "Explainable Neural Networks with Guarantee: A Sparse Estimation Approach" and "Generalization Analysis for Deep Contrastive Representation Learning". A huge congratulations to incoming PhD student Hieu Nong Minh for getting his paper accepted before even formally matriculating. A fantastic start to his PhD journey!

[08/24] I am honoured to have been offered the opportunity to serve as Action Editor for TMLR.

[08/24] Expert reviewer at TMLR.

[05/24] Our Paper entitled "Generalization Analysis of Deep Nonlinear Matrix Completion" was accepted at ICML 2024. Full text (Openreview). 

[04/24] Our paper entitled "Interpretable Tensor Fusion" was accepted at the IJCAI 2024 conference. Full text (ArXiv). 

[03/24] I am honoured to have been offered the opportunity to serve as an Area Chair at NeurIPS 2024. 

[03/24] Our paper entitled "Context Aware REpresentation: Jointly Learning Item Features and Selection from Triplets" was accepted at TNNLS. Full Text. 

[01/24] Our paper entitled "Unraveling the Dynamics of Stable and Curious Audiences in Web Systems" was accepted at the Web Conference 2024 (oral). Full text (ACM). 

[01/24] Our paper entitled "Numerically Tight Generalization Bounds for Adversarial Risk in Stochastic Neural Networks" was accepted at the AiStats 2024 conference. Full Text 

[08/23] Certificate of Excellence in Reviewing at KDD.

[07/23] Our late-breaking results paper was accepted at RecSys 2023 on recommendation systems with joint outlier detection via Graph Neural Networks.  Congratulations to Visiting Research Student Petr Kasalický! Full Text 

[07/23] Our paper on uncertainty-adjusted matrix completion was accepted at TNNLS. Full text (PubMed). 

[12/22] Chaired a session at AAAI 2023. 

[11/22] Our paper on generalization bounds for low-noise IMC was accepted at AAAI 2023 (oral). Full-text (ArXiv). 

[10/22]  "Top Reviewers" and complementary registration at NeurIPS 2022. 

[07/22] Top ten percent of reviewers at ICML 2022. 

[07/22] Started work as an Assistant Professor at SCIS. 

[05/22] "Highlighted reviewer" (top 10 percent) at ICLR 2022.