About

I am a Ph.D. candidate at KAIST, advised by Prof. Yong Man Ro. I work on full-duplex spoken dialogue models, which listen and speak at the same time. My goal is to make them hold natural, real-time conversations with people, which requires them to:

• See the user: perceive the user's face in real time (expressions, nods, gaze) and use these nonverbal cues to decide what to say and when to say it.

• Stay stable over long conversations: understand why full-duplex models collapse into repetition, silence, and gibberish over tens of minutes, and prevent it without retraining.

• Remember what matters: keep track of the user's underlying context across many turns, such as the root problem in an emotional-support conversation.

• Hold up in the real world: remain robust to noisy audio by using audio-visual speech cues.

I am looking for research internships in 2027.

Education

• KAIST, Daejeon, Korea (2024.03-Present)

Ph.D. in Electrical Engineering (Advisor: Prof. Yong Man Ro)

• KAIST, Daejeon, Korea (2022.03-2024.02)

M.S. in Electrical Engineering (Advisor: Prof. Yong Man Ro)

• SKKU, Suwon, Korea (2016.03-2020.02)

B.S. in Electrical Engineering

Publication

2026

1. Why Full-Duplex Speech Models Collapse in Long Conversations
   Yeonju Kim, Bella Godiva, Se Jin Park, Yong Man Ro
   Under Review

2. CounselPlex: A Stressor-Aware Full-Duplex Counseling Model
   Yeonju Kim, Bella Godiva, Yong Man Ro
   Under Review

3. Noise Adaptive Streaming Audio-Visual Speech Token Enhancement for Robust Full-Duplex Spoken Dialogue Models
   Bella Godiva, Yeonju Kim, Yong Man Ro
   Conference on Empirical Methods in Natural Language Processing (EMNLP), 2026

4. Emotion-Coherent Reasoning for Multimodal LLMs via Emotional Rationale Verifier
   Hyeongseop Rha, Jeong Hun Yeo, Yeonju Kim, Yong Man Ro
   The Association for the Advancement of Artificial Intelligence (AAAI), 2026

2024

5. Advancing Causal Intervention in Image Captioning with Causal Prompt

Youngjoon Yu, Yeonju Kim, and Yong Man Ro

IEEE Transactions on Neural Network and Learning Systems (TNNLS), 2024

6. What if...?: Thinking Counterfactual Keywords Helps to Mitigate Hallucination in Large Multi-modal Models
   Junho Kim, Yeon Ju Kim, Yong Man Ro
   Findings of the Association for Computational Linguistics: EMNLP (EMNLP), 2024

7. AV-EmoDialog: Chat with Audio-Visual Users Leveraging Emotional Cues
   Se Jin Park, Yeonju Kim, Hyeongseop Rha, and Yong Man Ro
   arXiv preprint, 2024
   
   

8. CODE: Contrasting Self-generated Description to Combat Hallucination in Large Multi-modal Models
   Junho Kim*, Hyun Jun Kim*, Yeon Ju Kim, Yong Man Ro (* Co-First Authors)
   Advances in Neural Information Processing Systems (NeurIPS), 2024

2023

9. Mitigating Dataset Bias in Image Captioning through CLIP Confounder-free Captioning Network
   YeonJu Kim, Junho Kim, Byung-Kwan Lee, Sebin Shin, Yong Man Ro
   IEEE International Conference on Image Processing (ICIP), 2023.

Teaching Experience

• EE202 Signals and Systems, KAIST (2023 Fall)

• EE305 Introduction to Electronics Design Lab, KAIST (2022 Fall, 2023 Spring, 2024 Spring, 2026 Spring)

• EE837 Special Topics on Signal Processing: Multimedia Processing and Learning (2024 Fall, 2026 Fall)

• EE474 Introduction to Multimedia (2025 Spring)

Skills

• Pytorch

• Opencv

• English, Korean

Academic Service

Conference Reviewer 

• ICLR, CVPR, EMNLP, AAAI, ICIP

Journal Reviewer 

• • • IEEE Transactions on Circuits and Systems for Video Technology (TCSVT)

    • IEEE Transactions on Image Processing (TIP)

Work Experience

Samsung Electronics / Global Technology Research(2020.03 ~ 2022.01)
Software Engineer

• Developed visual inspection software for manufacturing equipment using OpenCV and PyTorch. 

• Built user-friendly GUIs for equipment operation (MFC, WinForms, PyQt). 

• Developed software to control inspection hardware.