Haotian Qi [齊]
I am a third-year PhD student at TMH, KTH Royal Institute of Technology, where I am advised by Prof. Gabriel Skantze and Prof. Gustav Eje Henter My research is funded by the Wallenberg AI, Autonomous Systems and Software Program (WASP). Previously, I completed my Integrated Master's (MComp) in Computer Science with Artificial Intelligence at the University of Sheffield, graduating with Distinction.
My goal is to build real-time AI agents capable of natural, engaging human-agent interaction. To achieve this, my current research focuses on turn-taking dynamics, multimodal modeling, and conversational AI.
MuVAP: Multimodal Multiparty Voice Activity Projection for Turn-taking Prediction in the Wild
Turn-taking
Conversational AI
Multimodal
On the interplay between linguistic information and downstream performance: an analysis of tractable model compression
NLP
BERT
Probing
Linguistic Features
Positional Excitation Implementation for Human Action Recognition with ACTION-Net Model
Computer Vision
Self-attention
Action Recognition