Scholar
Hyoung-Kyu Song
Google Scholar ID: KR4U5YMAAAAJ
Applied Scientist (Member of Technical Staff) @ Mirage
Model Optimization
Image Generation
Multimodal Video Generation
Follow
Homepage
↗
Google Scholar
↗
Citations & Impact
All-time
Citations
391
H-index
7
i10-index
7
Publications
11
Co-authors
9
list available
Contact
GitHub
Open ↗
LinkedIn
Open ↗
Publications
1 items
Seeing Voices: Generating A-Roll Video from Audio with Mirage
2025
Cited
0
Resume
Academic Achievements
- Presented at CES and GTC
- Published papers at CVPR and ECCV
- Served as an Outstanding Reviewer for CVPR 2025
- Published papers such as 'Seeing Voices: Generating A-Roll Video from Audio with Mirage' and others
Research Experience
- Worked on Stable Diffusion, VLM optimization, and talking face generation at startups in South Korea
- Led the Brain team at MAUM.AI during its KOSDAQ listing in 2022
Background
Research Interests: Generative AI and model optimization, especially for LLMs and visual content.
Summary: An Applied Scientist at Captions, working on a diffusion-based video generation model called Mirage.
Co-authors
5 total
Bo-Kyeong Kim
Nota Inc.
Tae-Ho Kim
Nota Inc.
Junhyeok Lee
Johns Hopkins University, Center for Language and Signal Processing
Kang-wook Kim
Seoul National University
Sang Hoon Woo
Independent Researcher