뇌파(EEG)와 감정 모델링을 중심으로 AI 연구를 수행했습니다.
Research to Product: 연구실 안에서 잘 되는 AI를 넘어, 실제 환경에서 쓰이는 AI로 나아가고자 합니다.
AI 데이터 신뢰성 · 도메인 기반 AI · 감성 컴퓨팅(Affective Computing)

• 인하대학교 대학원 전기컴퓨터공학과 인공지능 전공 석사 2022.09—2024.08
• 인하대학교 산업경영공학과, 소프트웨어융합연계전공 2018.03—2022.08
  • 졸업 프로젝트: 열화상 이미지와 공정능력지수를 통한 배전설비 실시간 모니터링 시스템에서 GUI 담당 [code] — 🏅학술발표우수상(생산안전분야), 대한안전경영과학회 춘계학술대회 2021.10.22

EEG 기반 감정 인식에서의 데이터 신뢰성 : 비신뢰 쌍 탐지 및 제거 프레임워크 개발
감성인공지능 연구실 | 지도교수님: 김병형

뇌파(EEG) 데이터와 감정 레이블 간 상관성이 저하된 데이터 샘플들을 딥러닝의 OOD(Out-of-Distribution) 탐지 기법으로 식별·제거하는 데이터 정제 프레임워크를 제안하였습니다. 세 개의 공개 데이터셋에서 기존 방법(RP) 대비 분류 정확도 최대 6.14%, AUROC 최대 4.33% 향상시켰습니다.

• 2024.12 International Conference [pdf] [link] [code]
  Detecting Concept Shifts under Different Levels of Self-awareness on Emotion Labeling
  27th International Conference on Pattern Recognition (ICPR 2024)
  Hyoseon Choi, Dahoon Choi, Netiwit Kaongeon, Byung Hyung Kim*

• 2024.05 Domestic Journal [pdf] [link] 🏅KCC2023 우수논문상(인공지능분야) 초청논문 2023.06.19
  EEG 기반 감정 분류에서 MSP를 사용한 OOD 검출 적용
  정보과학회논문지 (Journal of KIISE)
  최효선, 최다훈, 김병형*

• 2024.02 International Conference [pdf] [link] [code]
  Multi-Output Regression for Integrated Prediction of Valence and Arousal in EEG-Based Emotion Recognition
  12th IEEE International Winter Conference on Brain-Computer Interface (WinterBCI)
  Hyoseon Choi, ChaeEun Woo, JiYun Kong, Byung Hyung Kim*

• 2023.08 International Journal [pdf] [link]
  Decoding Auditory-Evoked Response in Affective States using Wearable Around-Ear EEG System
  Biomedical Physics and Engineering Express
  Jaehoon Choi, Netiwit Kaongoen, Hyoseon Choi, Minuk Kim, Byung Hyung Kim*, Sungho Jo*

언어    |  딥러닝/ML    |  데이터/시각화    |  신호처리/통계    |  환경/도구    |  CS   2021.11.26  |  기타 경험  
RAG Agent 구축  

• 영어 어학연수   호주, 브리즈번 2024.09.09—2025.09.19 | Greenwich English College 1년, Cambridge FCE B2 취득 2025.06.10
• 국제 학회 포스터 발표 | ICPR 2024   인도, 콜카타 2024.12.01—2024.12.05
• 공부한 책
  • 『밑바닥부터 시작하는 딥러닝 1, 2』 인하대학교 AI 동아리 AiDL 스터디
  • 『혼자 공부하는 머신러닝+딥러닝』,『Do It! 딥러닝』,『파이토치로 배우는 LLM&AI』

Conducted AI research focused on EEG-based emotion modeling and data reliability.
Research to Product: Bridging the gap between lab-validated AI and AI that works in the real world.
AI Data Reliability · Domain-specific AI · Affective Computing

• Inha University M.S. in Artificial Intelligence, Dept. of Electrical & Computer Engineering 2022.09—2024.08
• Inha University B.S. in Industrial Engineering (Interdisciplinary: Software Engineering) 2018.03—2022.08
  • Graduation Project: Real-time Monitoring System for Electricity Distribution using Thermal Image and Process Capability Index — GUI Developer [code], 🏅Best Presentation Award, Korea Safety Management & Science 2021.10.22

Data Reliability in EEG-based Emotion Recognition: A Framework for Unreliable Data Pair Removal
Affective AI Lab | Advisor: Byung Hyung Kim

Proposed a data cleansing framework that identifies and removes unreliable EEG-label pairs — samples where the correlation between EEG signals and emotion labels is degraded — using deep learning-based Out-of-Distribution (OOD) detection. Achieved up to +6.14% improvement in classification accuracy and +4.33% in AUROC over the conventional method (Riemannian Potato) across three public datasets.

• 2024.12 International Conference [pdf] [link] [code]
  Detecting Concept Shifts under Different Levels of Self-awareness on Emotion Labeling
  27th International Conference on Pattern Recognition (ICPR 2024)
  Hyoseon Choi, Dahoon Choi, Netiwit Kaongeon, Byung Hyung Kim*

• 2024.05 Domestic Journal [pdf] [link] 🏅Invited from KCC 2023 Best Paper 2023.06.19
  Applying OOD Detection using MSP in EEG-based Emotion Classification
  Journal of KIISE
  Hyoseon Choi, Dahoon Choi, Byung Hyung Kim*

• 2024.02 International Conference [pdf] [link] [code]
  Multi-Output Regression for Integrated Prediction of Valence and Arousal in EEG-Based Emotion Recognition
  12th IEEE International Winter Conference on Brain-Computer Interface (WinterBCI)
  Hyoseon Choi, ChaeEun Woo, JiYun Kong, Byung Hyung Kim*

• 2023.08 International Journal [pdf] [link]
  Decoding Auditory-Evoked Response in Affective States using Wearable Around-Ear EEG System
  Biomedical Physics and Engineering Express
  Jaehoon Choi, Netiwit Kaongoen, Hyoseon Choi, Minuk Kim, Byung Hyung Kim*, Sungho Jo*

Languages    |  Deep Learning/ML    |  Data/Visualization    |  Signal Processing/Statistics    |  Environment/Tools    |  CS   2021.11.26  |  Other Experience  
RAG Agent  

• English Language Program   Brisbane, Australia 2024.09.09—2025.09.19 | Greenwich English College, Cambridge FCE B2 2025.06.10
• International Conference Poster Presentation | ICPR 2024   Kolkata, India 2024.12.01—2024.12.05
• Books I've Studied
  • Deep Learning from Scratch 1, 2 — Inha Univ. AI Club AiDL Study
  • Machine Learning + Deep Learning by Yourself, Do It! Deep Learning, AI and ML for Coders in Pytorch/Laurence Moroney