Project

(NLP/APP)  Yield Analysis Copilot Service with LLM-based Agentic System (with A-FAB, 2023. 07. 01 ~ )

• Building domain-specific LLM fine-tuning dataset for multi-turn multi-API-calling dialogues
  
  

(NLP/ML) Quantification of customer service data (with Memory CS, 2022. 09. 01 ~ 2024. 12. 31)

• Aspect-based sentiment analysis of business reviews in semiconductor domain
  
  

(ASR/NLP/ML) Feature engineering and stress detection model development based on voice, language, and biosensors  (Kor: 음성/언어/바이오센서 기반 feature engineering 기술 및 스트레스 모델 개발) (with SKT, 2021. 12. 01 ~ 2022. 06. 24)

• Development of stress detection model using voice and  biosensor features
  
  

(NLP/ML) Pilot study for persona-based self-aware and empathetic  dialogue agent (Kor: 페르소나 기반 Self-Aware한 공감형 대화 에이전트를 위한 파일럿 데이터 구축 및 연구) (with Smilegate, 2021. 11.18 ~ 2022. 01. 24)

• Task formulation and  management of dialogue data collection
  
  

(ASR/NLP/ML) Research on advanced Korean-English speech translation model (Kor: 한국어-영어 음성-번역 데이터 구축 고도화 및 번역 모델 연구 / 네이버 주식회사) (with NAVER PAPAGO, 2021. 01. 01 ~ 2021. 04. 30)

• Construction of Korean-English speech-to-text translation dataset and building baseline models
  
  

(ASR/NLP/ML) Construction of Korean-English parallel corpus for high quality speech-to-text translation (Kor: 음성 번역 품질 고도화를 위한 한국어-영어 음성-번역 데이터 구축 / 네이버 주식회사) (with NAVER PAPAGO, 2020. 09. 01 ~ 2020. 12. 31)

• Construction of Korean-English speech-to-text translation dataset
  
  

(NLP/TTS) Deep learning multi-speaker prosody and emotion cloning technology based on a high quality end-to-end model using small amount of data (Kor: 소량 데이터만을 이용한 고품질 종단형 기반의 딥러닝 다화자 운율 및 감정 복제 기술 개발 / 과학기술정보통신부 차세대인공지능핵심원천) (with IITP, 2020. 04. 01 ~ 2021. 12. 31)

• Developed Korean-English code-mixed G2P for speech synthesis
  
  

(ASR/NLP/ML) Creating a system for understanding non-canonical spoken utterances,  comprehensively utilizing sound, text, and linguistic context (Kor: 음향과 텍스트, 그리고 언어적 맥락을 통합적으로 활용한 비정형 음성 발화 의도 파악 시스템 제작 / 네이버 주식회사) (with NAVER, 2019. 11. 01 ~ 2020. 04. 30)

• Developed and enhanced  end-to-end spoken language understanding modules
  
  

(ASR/NLP/ML) Development of free-running speech recognition technologies for embedded robot system (Kor: 로봇용 free-running 임베디드 자연어 대화음성인식을 위한 원천 기술 개발 / 산업통상자원부 산업핵심기술개발사업) (with MOTIE, 2017. 04. 10  ~  2020. 12. 31)

• Developed guideline for annotating Korean speech act, created corpus, and implemented prototype models
  
  

(Acoustics/ML) Predicting patients with Suicidality: Machine-learning based Voice analysis (Prism-V)  (Kor: 기계 학습 알고리즘 기반 음성 분석을 이용한 자살 고위험군 예측 기술 개발 / 이공학개인기초연구 한국형 SGER) (with SNUH, 2018. 11. 01 ~ 2021. 10. 31)

• Analysis on the voice recordings of patients with depression; extracted acoustic features and implemented ML baselines
  
  

(Communication) Implementation and Research on Vocoder (Kor: 보코더 구현 및 특성 연구 / 국방과학연구소) (with ADD, 2015. 10. 15 ~ 2020. 12. 31)

• Analysis on baseline vocoder models (MELP)
  
  

(Acoustics) Integrated sound scene awareness technology (Kor: 통합 싸운드 씬 인지 기술 / 한국연구재단 중견연구자사업 성과확산) (with NRF, 2015. 11. 01 ~ 2017. 10. 31) 

• Implementation of virtual bass technology
  
  

(Acoustics) Development of Stereo Audio Acquisition/Reproduction Technology for Realistic Orchestra Service (Kor: 실감 오케스트라 서비스를 위한 입체 음향 취득/재현기술 개발 / 정보통신기술진흥센터 대학ICT연구센터육성) (with IITP, 2013. 06. 01 ~ 2016. 12. 31)

• Performed recording experiments for RIR measurement