Research on Sound Event Detection

Last updated on Sep 28, 2019

My contributions and achievements go as follows:

  1. Proposed a bi-domain (on both time domain and frequency domain) attention model for rare sound event detection.
  2. Proposed a hierarchical pooling structure for weakly-labeled sound event detection.
  3. Outperformed state-of-the-art methods on evaluation dataset of Task 2 and Task 4 of Detection and Classification of Acoustic Events and Scenes (DCASE) 2017.
  4. Two Papers published in Proceedings of Interspeech 2019 as first or co-first author.
Audio
Avatar
Yuhan Shen
沈宇寒

PhD student

My research interests include machine learning, computer vision and natural language processing.