Hello! I'm Luong
AI researcher exploring multimodal learning and scalable models.
I am an AI Research Resident at the FPT Software AI Center, working on multimodal systems and mixture-of-experts architectures. I earned my Bachelor's in Data Science and Artificial Intelligence (Advanced Program) at the Hanoi University of Science and Technology (HUST), graduating in 2024 with a GPA of 3.75/4.0.
I am especially excited about research that bridges robust representation learning with practical deployment: scaling CLIP-style models efficiently, designing aligned vision-language systems, and building trustworthy multimodal agents.
Publications
-
CLIP-FMoE: Scalable CLIP via Fused Mixture-of-Experts with Enforced SpecializationOpenReview
-
More Reliable Pseudo-labels, Better Performance: A Generalized Approach to Single Positive Multi-label LearningPaper
-
Improving Single Positive Multi-label Classification via Knowledge-based Label-weighted Large Loss RejectionPaper
Preprints
-
Precise Video-to-Audio Generation with Cross-Modal Alignment in Latent Space
-
LIBMoE: A Library for Comprehensive Benchmarking of Mixture of Experts in Large Language ModelsarXiv
Awards & Honors
-
Best Presentation Award
Recognized for thesis presentation in the 2024.2 Computer Vision Council graduation.
-
Track Winner - Cloud Application for Community
Co-created EcoFrenzy, a social app encouraging greener daily actions.