×

About Me

I am currently a first-year Ph.D. student at the Ohio State University under the supervision of Assist. Prof. Wei-Lun Chao. Before that, I obtained my bachelor degree in Computer Engineering at University of Information Technology, Vietnam National University Ho Chi Minh City in late 2020, and was a research assistant at VinUni-Illinois Smart Health Center (VISHC), College of Engineering and Computer Science, VinUni, and AI research resident at FPT Software AI Residency program, mentored by Assist. Prof Dung D. Le (VinUniversity) between 2022 and 2024. More detailed can be found in my CV (update July 2024).

 

Research interests

My research interests include meta-learning, domain adaptation, and black-box optimization for lifelong and open-world machine learning, especially in computer vision. My research goal is to develop machine learning algorithms that continuously learn with minimal data or supervision and effectively adapt to other domains outside the training data.

I am also interested in exploring the connection between optimization and machine learning for effective and generalizable learning algorithms.

Details of my research interests are discussed here.

 

Highlighted Publications

Zero-Shot Object-Level Out-of-Distribution Detection with Context-Aware Inpainting
Quang-Huy Nguyen*, Jin Zhou*, Zhenzhen Liu, Huyen Bui, Kilian Q. Weinberger, Dung D. Le
Under review, 2024

Our paper tackles Object-level OOD Detection without access to the training data, considering object detector as a black-box function. We leverages off-the-shelf Diffusion model to replace detected object with in-context inpainting, drawing the input object closer to the in-distribution (ID) domain. Hence, we are able to recognize OOD objects that are erroneously predicted by the object detection model without any re-training effort.


Controllable Expensive Multi-objective Learning with Warm-starting Bayesian Optimization
Quang-Huy Nguyen*, Long P. Hoang*, Hoang V. Vu, Dung D. Le
Under review, 2024

Our work explores Multi-objective Black-box Optimization with Pareto Front Learning (PSL), aligning trade-off preferences with corresponding optimal solutions between conflicting objectives. As existing methods often suffer from unstable and inefficient performance when optimizing based on Gaussian Processes (GPs), we then tackle this by leveraging warm-starting Bayesian Optimization to adequately obtain a good approximation of the front first and re-initialize the Pareto Set Model during the opimization steps to stabilize the PSL.


Enhancing Few-shot Image Classification with Cosine Transformer
Quang-Huy Nguyen, Cuong Q. Nguyen, Dung D. Le, Hieu H. Pham
IEEE Access, 2023

We explore Few-shot Image Classification by proposing a new cross-attention mechanism based on cosine similarity, without using softmax, to further emphasizes the correlation between labeled supports and unlabeled query representations, thus enhancing ViT-based few-shot algorithms across various settings and scenarios compare to convention attention mechanism.

 

News