About

I am a Ph.D. student in Artificial Intelligence at Music and Audio Research Group, Seoul National University. My research interests lie in robust and efficient speech processing. In particular, I am working on domain shift and uncertainty (robustness); knowledge distillation (architectural efficiency); semi-supervised learning and training acceleration (data efficiency).

Education

[2021 - now]
Ph.D. in Artificial Intelligence, Seoul National University
Advisor : Prof. Kyogu Lee

[2017 - 2021]
B.S. in Systems Management Engineering, Sungkyunkwan University

Experience

[Sep 2023 - Feb 2024]
Research Intern, NAVER Papago

Publications (google scholar)

• Guiding Frame-Level CTC Alignments Using Self-Knowledge Distillation [link]
  Eungbeom Kim, Hantae Kim, Kyogu Lee
  INTERSPEECH 2024
• Debiased Automatic Speech Recognition for Dysarthric Speech via Sample Reweighting with Sample Affinity Test [link]
  Eungbeom Kim, Yunkee Chae, Jaeheon Sim, Kyogu Lee
  INTERSPEECH 2023
• Exploring Train and Test-Time Augmentations for Audio-Language Learning [link]
  Eungbeom Kim, Jinhee Kim, Yoori Oh, Kyungsu Kim, Minju Park, Jaeheon Sim, Jinwoo Lee, Kyogu Lee
  arXiv 2022
• Representation Selective Self-distillation and wav2vec 2.0 Feature Exploration for Spoof-aware Speaker Verification [link]
  Jin Woo Lee, Eungbeom Kim, Junghyun Koo, Kyogu Lee
  INTERSPEECH 2022

Challenge

• Automated Audio Captioning Using Parameter Efficient Fine-Tuning and Merging for LLMs
  Eungbeom Kim, Jaeheon Sim, Jinwoo Lee, Kyogu Lee
  4th Place at Automated Audio Captioning on DCASE 2024