GeezerVison

GeezerVison is showing hungchiayu1.github.io — Chia Yu Hung

I am currently a 2nd year PhD student at Nanyang Technological University (NTU), fortunately to be advised Prof. Soujanya Poria. My research interest lies in MultiModals and Vision Language Action Models.

Previously, I earned my Bachelor’s degree in Computer Science from Singapore University of Technology and Design (SUTD) in May 2024. Throughout my undergraduate period, I was fortunate to have opportunity to conduct research under Prof. Roy Ka-Wei Lee and Prof. Soujanya Poria.

Current works: I am working on reinforcement learning with action conditioned world models!! Feel free to email me to chat about this!

news

Jun 10, 2026	I have joined Meta SuperIntelligence Lab (FAIR) as a Research Scientist Intern in NYC this Summer working on multimodal LLM training with my mentor Daniel Bolya!
Jan 27, 2026	TangoFlux has been accepted to ICLR 2026!
Nov 20, 2025	We are releasing NORA-1.5, a VLA post-trained using action conditioned world model as reward.
May 01, 2025	We are excited to release Nora, a VLA based on Qwen 2.5 VL with FAST+ tokenizer!.
Jan 24, 2025	Darwin has been accepted to NAACL 2025 (Oral)!

selected publications

ACM MM 2024 (Oral)

Tango 2: Aligning Diffusion-based Text-to-Audio Generations through Direct Preference Optimization

Navonil Majumder^*, Chia-Yu Hung^*, Deepanway Ghosal^*, Wei-Ning Hsu , and 2 more authors

2024
NAACL 2025 (Oral)

Inference Time Alignment with Reward-Guided Tree Search

Chia-Yu Hung, Navonil Majumder, Ambuj Mehrish, and Soujanya Poria

2024
ICLR 2026

TangoFlux: Super Fast and Faithful Text to Audio Generation with Flow Matching and Clap-Ranked Preference Optimization

Chia-Yu Hung, Navonil Majumder, Zhifeng Kong, Ambuj Mehrish , and 3 more authors

2024
NORA: A Small Open-Sourced Generalist Vision Language Action Model for Embodied Tasks

Chia-Yu Hung, Qi Sun, Pengfei Hong, Amir Zadeh , and 4 more authors

2025

chiayu001 at e.ntu.edu.sg