Publications

Papers I am very proud of:

• Ums Considered Harmful (SIGBOVIK 2026)
• Language Model, 2025 (SIGBOVIK 2025)
• Grounding Language Models to their Physical Presence (SIGBOVIK 2024)

Papers I am utterly ashamed of:

• None yet, thankfully.

Papers:

• Automated Speech Analysis for Screening and Monitoring Bipolar Depression: Machine Learning Model Development and Interpretation Study (JMIR Med Inform 2025)
• Prompt Crossing: Evaluating Whether LLM Response Stems from Jailbreak or Normal Prompt (ICASSP 2025)
• RT-VQ²A²: Real Time Vector Quantized Question Answering with ASR (COLING 2024)
• Non-Essential is NEcessary: Order-agnostic Multi-hop Question Generation (COLING 2024)
• Unsupervised Extractive Dialogue Summarization In Hyperdimensional Space (ICASSP 2024)
• Unsupervised Dialogue Topic Segmentation in Hyperdimensional Space (Interspeech 2023)
• Cross-task Knowledge Transfer for Extremely Weakly Supervised Text Classification (ACL Findings 2023)
• LIME: Weakly Supervised Text Classification Without Seeds (COLING 2022)
• Leveraging Non-dialogue Summaries For Dialogue Summarization (Transcript Understanding Workshop at COLING 2022)
• Unsupervised Abstractive Dialogue Summarization with Word Graphs and POV Conversion (WIT Workshop at ACL 2022)
• Finetuning Pretrained Transformers into Variational Autoencoders (Insights Workshop at EMNLP 2021)
• Improving Distinction between ASR Errors and Speech Disfluencies with Feature Space Interpolation (preprint, 2021)
• Technology Adoption in Spectrum Sharing: Estimating the Impact on Incumbents in the 3.5GHz Band (Speaker at TPRC47)

Patents:

• 단어 그래프를 이용한 대화 요약 생성 방법 (KR 10-2444457)
• 음성 인식 오류 검출을 위해 학습 데이터를 생성하기 위한 방법 (KR 10-2478076)
• Method for generating learning data for speech recognition error detection (US 11,721,325)
• 자소 정보를 이용한 음성 인식 방법 (KR 10-2478763)
• Method for speech recognition with grapheme information (US 11,749,260)
• 인코더-디코더 구조를 가진 인공 신경망 모델을 활용한 데이터 생성 방법 (KR 10-2480423)
• Method of generating data by using artificial neural network model having encoder-decoder structure (US 11,887,002)
• 텍스트 정보에서 타겟 단어에 대응하는 단어를 식별하는 방법 (KR 10-2517661)
• Method for identify a word corresponding to a target word in text information (US 12,197,862)
• 비-대화형 텍스트 정보를 활용하여 대화 요약을 위한 학습 데이터를 생성하는 방법 (KR 10-2520167)
• Method for dialogue summarization with word graphs (US 11,640,493)
  주 화자에 기초하여 요약문을 생성하는 방법 (KR 10-2537173)
• Method of generating summary based on main speaker (US 17/962,011)
• 텍스트 분류를 위한 학습 데이터를 생성하는 방법 (KR 10-2545892)
• 초차원 컴퓨팅을 이용하여 텍스트를 분할하는 방법 (KR 10-2647510)
• テキスト情報の中からターゲット単語と対応する単語を識別する方法 (JP 7,475,592)