Sangkeun Jung

Pohang University of Science and Technology

Papers

Total Citations

112

H-Index

About

Sangkeun Jung is a leading researcher in spoken dialog systems, with a focus on data-driven user simulation and robust dialog management. His seminal work, "Data-driven user simulation for automated evaluation of spoken dialog systems" (2009, 74 citations), pioneered the use of simulated users to automatically assess system performance, reducing the need for costly human evaluations. This contribution has become foundational in the field, enabling scalable testing of conversational agents. Jung further advanced dialog management through his hybrid approach, combining agenda-based control with dialog examples to handle ambiguous or erroneous speech recognition outputs. His 2008 and 2009 papers (with 15 and 23 citations, respectively) introduced an agenda-based method that leverages n-best recognition hypotheses and discourse interpretation algorithms to maintain coherent dialog states, significantly improving robustness in noisy environments. These innovations have influenced the design of modern conversational AI, making Jung a key figure in developing more reliable and efficient spoken dialog systems. His work continues to impact researchers and engineers building next-generation interactive agents.