Hello, I'm Ngoc Quan

I am a final-year student at Hanoi University of Science and Technology. I have gained valuable experience working as a Research Assistant at the BKAI Laboratory and as an AI Resident at VinAI Research (mentored by Prof. Trung Le), where I focused on topics including Bayesian Inference and Machine Learning Theory. Currently, I am an AI Research Resident at Qualcomm AI Research, my research will focus on model fine-tuning and test-time scaling techniques for image generation.

You can connect with me on LinkedIn linkedin.com/in/ngocquanai or Facebook fb.com/ngocquanofficial, and feel free to visit my blog to explore more images about me.


News

  • Apr 2025: I have officially joined Qualcomm AI Research as an AI Resident. My research will focus on model fine-tuning and test-time scaling techniques for image generation.
  • Dec 2024: I have successfully achieved an IELTS Academic Band Score of 7.0, marking another step toward my academic and professional goals.
  • Nov 2024: The LaTeX version of my book, 89 Functional Equations (Vietnamese version), is now available to read at vmo.ngocquan.com. The English version will be coming soon.
  • Aug 2024: I have officially joined the VinAI Residency Program as an AI Research Resident under the mentorship of Prof. Trung Le.
  • Mar 2024: Robust Pitch-Fusion Model has been accepted at ICASSP 2024, and the accompanying dataset, ViSEC, has also been released.
  • Sep 2023: I was awarded Second Prize in the SOICT Hackathon 2023 (AI-Powered BFIS Track).
  • May 2023: I achieved First Prize in the National Startup Contest for Students (SV.STARTUP 2023).
  • Jan 2021: I achieved Third Prize in the Vietnam Mathematics Olympiad 2021 (VMO 2021)

Publications

A Robust Pitch-Fusion Model for Speech Emotion Recognition in Tonal Languages

A Robust Pitch-Fusion Model for Speech Emotion Recognition in Tonal Languages

ICASSP, 2024

This paper presents Pitch-fusion, a novel speech emotion recognition (SER) model optimized for tonal languages by incorporating pitch features to enhance performance.

Improving Generalization With Flat Hilbert Bayesian Inference

Improving Generalization With Flat Hilbert Bayesian Inference

arXiv, 2024

This paper introduce Flat Hilbert Bayesian Inference (FHBI), an algorithm that enhances Bayesian inference generalization through an iterative two-step process involving adversarial functional perturbation and functional descent within reproducing kernel Hilbert spaces.

Blog Post