Olivier Jiyoun Jung
[ click anywhere to enter ]
· · · · · ·
Olivier Jiyoun Jung
01

Olivier Jiyoun Jung

Graduate researcher at the intersection of Speech and Language Processing and Human–AI Interaction, studying the quality of human conversation and how it breaks down.

Affiliation
Ewha Womans University
Division of Communication & Media
Focus
Speech & Language Processing
Human–AI Interaction
Computational Communication
Correspondence
olivierjiyounjung@gmail.com
Status
M.A. candidate · Graduate Research Assistant
02
About

Who I Am

My research studies the quality of human conversation — and the points where it breaks down — through speech.

I model dialogue from two directions at once: what is said, through language and LLM-based analysis, and how it is said, through prosody and timing. Each direction captures what the other misses, and reading them together surfaces what neither shows alone: depth and responsiveness in facilitated group dialogue, early traces of cognitive and affective decline in how a person talks, and the shifts in trust and memory that follow when a voice turns out to be synthetic.

Working across Speech and Language Processing, Human–AI Interaction, and Computational Communication, I build interpretable models to understand how people communicate, connect, and respond through speech.

Speech & Language Processing Human–AI Interaction Computational Communication
03
Education

Academic Path

Mar 2025

Aug 2027
M.A., Communication & Media
Ewha Womans University · SeoulAdvisor: Prof. Jiyoung Kang · GPA 4.44/4.5 · expected Aug 2027
Aug 2023

Feb 2024
Exchange Student, Applied Science
University of Applied Sciences and Arts Northwestern Switzerland · Basel
Mar 2021

Feb 2025
B.A., Communication & Media
Ewha Womans University · Scranton Honors ProgramGPA 4.32/4.5 · Summa Cum Laude
04
Research

Selected Works

01
Listening Between the Lines: Joint Learning of ASR Embeddings and LLM-Augmented Linguistics for Dementia Detection
Jung, O. J.†, Park, J.† & Oh, M. (†equal contribution)
INTERSPEECH · Accepted 2026 · arXiv
Accepted2026
02
LoRA-Tuned Large Language Models for Dementia Detection via Multi-View Speech-Derived Features
Park, J.†, Jung, O. J.† & Oh, M. (†equal contribution)
INTERSPEECH · Accepted 2026 · arXiv
Accepted2026
03
South Korean V-Tuber Visual Conventions within Korea–Japan Cultural Circuits
Jung, O. J., & Youk, S.
Asian Communication Research, 23(2), Article 014 · 2026 · DOI
Published2026
04
Comparing Neural Responses to Human and Social Robot Interlocutors: An EEG Study on CASA and Perceived Pleasantness, Importance, and Cooperativeness
Youk, S., Jung, O. J., Jie, Y., & Lee, S. K.
ICA 76th Annual Conference · 2026
Presented2026
05
Analyzing V-Tuber Visual Tropes in South Korea Through Korea–Japan Media Circuits
Jung, O. J., & Youk, S.
NCA 111th Annual Convention · 2025
Presented2025
06
Hearing Is Not Remembering: Labeled AI Exposure Protects Memory Without Improving Discrimination of Deepfake Audio
Jung, O. J.
ACM CHI 2027 · Under Review
Under Review2027
07
Toward a Neurocognitive Account of CASA: EEG Evidence from Human and ChatGPT-Enabled Social Robot Conversation
Youk, S., Jung, O. J., Jie, Y., & Lee, S. K.
Computers in Human Behavior · Under Review 2026
Under Review2026
08
How Policy Shapes Media Narratives of Environmental Risk: A Nine-Year Analysis of Particulate Matter News Coverage in South Korea
Youk, S., Kim, H., Jung, O. J., & Kim, Y.
Under revision for resubmission · 2026
In Progress2026
09
Meme Politics: Effects of AI-Generated Short-Form Content on Young Voters in South Korea
Jung, O. J.
Computational Social Science · In Progress
In Progress
10
End-to-end Multimodal Explanation Generation for Audio Deepfake Detection
Jung, O. J.
Speech & Audio · In Progress
In Progress
11
Multimodal Depression Detection from Clinical Interviews via Speech-Text Feature Fusion
Jung, O. J.
Clinical NLP · In Progress
In Progress
05
Experience

Research & Work

01
May 2026 — Present
Research Collaborator
MIT Media Lab · Center for Constructive Communication · Cambridge, MA
Operationalizing interpersonal communication constructs — self-disclosure, partner responsiveness, narrative structure — as measurable features of group dialogue.
Leading the audio analysis, building a speech pipeline (MFA, WhisperX, acoustic features) to test how prosody complements text-based measures.
02
Nov 2025 — May 2026
AI Research Scientist Intern
NAVER Cloud · HCX Application AI Modeling Team · Seongnam
Developed speech AI and LLM systems; extracted discourse and conversational features.
Implemented and fine-tuned deep models in PyTorch for speech & language processing.
03
Mar 2025 — Present
Graduate Research Assistant
Ewha Womans University · Division of Communication & Media · Seoul
Independent research on AI-mediated communication and computational media.
Interdisciplinary collaborations across communication, HCI, and AI.
04
2025
Teaching Assistant
Ewha Womans University · Communication and Society · Seoul
Supported instruction for Communication and Society.
06
Tools & Methods

Instruments of the Trade

Programming

  • Python
  • R
  • JavaScript
  • LaTeX

Machine Learning

  • PyTorch
  • Deep Learning
  • NLP
  • LLM Fine-tuning
  • Speech Processing
  • Computer Vision

Research

  • SPSS
  • PsychoPy
  • EEG
  • Unreal Engine (VR)

Audio & Creative

  • Ableton Live
  • Logic Pro
  • Melodyne
  • Adobe Photoshop
  • Microsoft Office
07
Recognition

Honors & Awards

2025–2026
Excellence Ewha-in Scholarship ($10,000)
Ewha Womans University
2025
Summa Cum Laude
Ewha Womans University
2025
Student Travel Grant, NCA 111th Convention
National Communication Association
2024
Departmental Academic Excellence Scholarship
Ewha Womans University
2023
FHNW International Merit Scholarship
University of Applied Sciences and Arts Northwestern Switzerland
2022
Excellence Award, 5th Investigative Reporting Contest
Korea News Agency Commission
2021–2025
Highest Honors (all semesters)
Ewha Womans University
08
· ✦ · ✦ · ✦ ·
Rate Your Experience
How was your visit?
Say Hello
Leave a message ↓
Guestbook
System
Guestbook initialized. Loading entries...
Contact
Get In Touch
Open to research collaborations, speaking invitations, and interesting conversations.