Tony Woo
Hello everyone! My name is Tony Woo (or if you prefer my Korean name, Sang Hoon Woo) and I’m a research intern at SK Telecom and the Vision & Learning Lab at Seoul National University, where I work on the Sovereign AI Foundation Model Project, a national initiative supported by the South Korean government. I am also applying to PhD programs in the United States this year.
My works cover a range of topics, but my central research area is conversational AI. Broadly, my research goal is to create natural, context-aware conversational agents that improve human capability. Some of my current subtopics of interest include:
- Spoken Dialogue Interface: Humans perform most of their communication through speech, not text. Conversational systems should therefore support natural, voice-based interaction, not just by converting text to speech, but by understanding the characteristics of speech as a medium and adapting or exploiting them for more effective communication.
- Multimodal Context: Human perception is inherently multimodal, shaped by simultaneous cues from vision, audio, and more. I am interested in developing conversational agents that can interpret these diverse signals, integrate them meaningfully, and, when appropriate, generate multimodal expressions themselves.
- Dynamic Conversational Context: Conversations are dynamic. Users reveal new information about themselves, revise opinions, and shift preferences over time. Even external factors, e.g. temporal, social, or situational, change as interactions unfold. A robust conversational agent must track these evolving contexts and adapt its responses accordingly.
I’m currently advised by Professor Gunhee Kim at Seoul National University. Before my current role, I worked at a couple of Korean startups for my alternative military service, focusing on speech and language technologies. I received my B.S. in Computer Science from University of California Davis, where I had the opportunity to work with Professor Prem Devanbu and Professor Chen-Nee Chuah, that introduced me to the world of research.
news
| Oct 13, 2025 | Just started my research internship at SK Telecom! |
|---|---|
| Aug 21, 2025 | Our paper Think, Verbalize, then Speak: Bridging Complex Thoughts and Comprehensible Speech got accepted to EMNLP 2025! |
selected publications
- DExTER: Can Omnimodal Language Models Resolve Audio-Visual Deixis?Manuscript under review
-