Research interest

• Speech recognition

  • Low resource language end-to-end automatic speech recognition

  • Word alignment for end-to-end automatic speech recognition

• Speaker verification/identification

  • Deep learning-based end-to-end speaker verification system

• Speech representation learning

  • Unsupervised speech embedding system training

  • Uncertainty robust embedding system for short-duration speech signals

  • Nuisance attribute disentangled speaker embedding

• Language/age/gender identification

  • I-vector-based age/gender estimation

  • Open-set language identification

• Anti-spoofing countermeasure

  • End-to-end logical access spoof attack detection