Cheongjae Lee

About

Cheongjae Lee is a leading researcher in spoken dialog systems and human-robot interaction, with a career dedicated to making machines more conversational and context-aware. His work centers on robust dialog management, user simulation, and affective computing. Lee’s most influential contribution is the development of data-driven user simulation for automated evaluation of spoken dialog systems (74 citations), a foundational method that enables scalable, objective testing without human intervention. He further advanced dialog robustness with a hybrid agenda-based approach that leverages dialog examples and n-best recognition hypotheses (23 and 15 citations), allowing systems to manage ambiguity and maintain coherent state tracking even under noisy conditions. Lee also explored emotion recognition from natural language dialogs (10 citations), integrating affective cues into user interfaces to improve human-robot interaction. His applied work includes the PHOPE intelligent building guide robot (4 citations), a multi-modal platform combining face recognition, voice communication, and navigation. Through these contributions, Lee has shaped how dialog systems are evaluated, made robust, and infused with emotional intelligence—impacting both academic research and real-world service robotics.

Research Focus

Key Achievements

Top Papers

1. 1
   Data-driven user simulation for automated evaluation of spoken dialog systems

   74 citations · 2009
2. 2
   Hybrid approach to robust dialog management using agenda and dialog examples

   23 citations · 2009
3. 3
   Robust Dialog Management with N-Best Hypotheses Using Dialog Examples and Agenda

   15 citations · 2008
4. 4
   Emotion Recognition for Affective User Interfaces using Natural Language Dialogs

   10 citations · 2007
5. 5
   A system architecture for intelligent building guide robot PHOPE

   4 citations · 2009

Key Collaborators

Contact & Links