Beomseok Lee

PhD Student

    Short bio

    Beomseok Lee is a PhD student at the University of Trento, conducting research at Fondazione Bruno Kessler and NAVER LABS Europe.

    Before starting his PhD, Beomseok worked as a full-time research engineer at SAMSUNG Research (Samsung Electronics R&D hub) Global AI Center, Seoul where he specialized in End-to-end (E2E) Speech-to-text translation. His current research focuses on E2E Spoken Language Understanding with an emphasis on multi-task, multi-lingual and multi-modal approaches. He holds a Computer Science (CS) Master's degree from Korea Advanced Institute of Science & Technology (KAIST, Korea) and a CS Bachelor's degree from Sungkyunkwan University (SKKU, Korea).

    Research topics

    Spoken Language Understanding, Multi-modality

    Publications

    1. Lee, Beomseok; Calapodescu, Ioan; Gaido, Marco; Negri, Matteo; Besacier, Laurent,
      in ยซยป,
      Proceedings of Interspeech2024,
      ,
      vol. ,
      n. ,
      2024
      , pp. 817-
      821

    ๐Ÿš€ New tech report out! Meet FAMA, our open-science speech foundation model family for both ASR and ST in ๐Ÿ‡ฌ๐Ÿ‡ง English and ๐Ÿ‡ฎ๐Ÿ‡น Italian.

    The models are live and ready to try on @huggingface ๐Ÿ‘‡
    ๐Ÿ”—

    #ASR #ST #OpenScience #MultilingualAI

    ๐Ÿš€ New shared task at #WMT2025 (co-located with @emnlpmeeting ): Model Compression for Machine Translation!
    Can you shrink an LLM and keep translation quality high?๐Ÿ”ง
    Submit by July 3 and push the limits of efficient NLP!
    ๐Ÿ‘‰ https://www2.statmt.org/wmt25/model-compression.html #NLP #ML #LLM #ModelCompression

    More great news! ๐ŸŽ‰
    Our paper โ€œEchoes of Phonetics: Unveiling Relevant Acoustic Cues for ASR via Feature Attributionโ€ was accepted at #Interspeech2025!

    Interested in interpretability for speech models? Preprint coming soon!

    โœ๐Ÿผ @mgaido91, @negri_teo, M.Cettolo, @luisabentivogli

    Load More