Takyoung Kim

김탁영 (金卓映)

My main research interest is focused on solving problems arising from human-machine communication in the real-world. I’m also interested in conversational reasoning (e.g., dialogue system and question answering), knowledge-intensive NLP (e.g., retrieving information), broad range of long-context modeling, and making language models trustworthy.

[Jun 2024] Our Bias Analysis in Deployed TOD Systems paper has been accepted to NLP4ConvAI Workshop @ ACL 2024 as oral presentation. → Update: Our work has been awarded as Best Paper!🏆
[Feb 2024] Life update! I will start my Ph.D. journey at UIUC Computer Science starting from Fall 2024.
[Feb 2024] Our Entity Adaptive DST paper has been accepted to Knowledge-Based Systems Journal.
[Jun 2023] Our Entity Adaptive DST paper has been accepted to Knowledge NLP Workshop @ KDD 2023.
[May 2023] Our two papers have been accepted to ACL 2023 - Korean Sensitive QA Benchmark as oral presentation and Korean Bias Benchmark as Industry Track, respectively.
[Apr 2023] Excited to join LG AI Research as a Research Intern!
[Oct 2022] Our DST with Turnback paper has been accepted to SereTOD Workshop @ EMNLP 2022.
[Aug 2022] Graduated from Korea University.
[Jul 2022] Excited to join NAVER AI Lab as a Research Intern!
[May 2022] Will attend ACL 2022 in-person for poster presentation (May 22 - May 27, Dublin, Ireland).
[Feb 2022] Our Multi-turn DST Evaluation paper has been accepted to ACL 2022.
[Sep 2020] Started Master’s degree at Korea University.
[Aug 2020] Graduated from Sungkyunkwan University.
[Jan 2020] Started undergraduate internship at Korea University.

Takyoung Kim

News

Education

Work Experience

Publications

Learning to Explore and Select for Coverage-Conditioned Retrieval-Augmented Generation

Revealing User Familiarity Bias in Task-Oriented Dialogue via Interactive Evaluation

DSTEA: Improving Dialogue State Tracking via Entity Adaptive Pre-training

KoSBi: A Dataset for Mitigating Social Bias Risks Towards Safer Large Language Model Application

SQuARe: A Large-Scale Dataset of Sensitive Questions and Acceptable Responses Created Through Human-Machine Collaboration

Oh My Mistake!: Toward Realistic Dialogue State Tracking including Turnback Utterances

Mismatch between Multi-turn Dialogue and its Evaluation Metric in Dialogue State Tracking

Building an Integrated Framework of Korean Text Summarization and Text-to-Speech