About me
Quandong is a senior algorithm engineer and also a researcher at Speech Group, Xiaomi AI Lab, with collaboration with Dr. Daniel Povey He is responsible for the exploration and implementation of cutting-edge speech technologies. His research focuses on far-field multi-channel / end-to-end speech recognition, speech enhancement, speech separation, and multimodal modeling.
He is passionate about advancing far-field speech processing technologies and solving key problems. He has a vision of opening the door to the ultimate natural voice experience in the era of the Internet of Things, where there would not exist any obstacles to human-machine and human-human communication.
Prior to joining Xiaomi, Quandong received his Ph.D. degrees in Signal and Information Proccessing from Institute of Acoustics, Chinese Academy of Sciences / University of Chinese Academy of Sciences in 2019. During the Ph.D. period, he was also a joint training Ph.D. student at Georgia Institute of Technology under the supervision of Prof. Chin-Hui Lee. He received his B.S. degree in Electrical Engineering from Harbin Engineering University in 2014.
Recent Projects / Research Highlight
- Multi-channel End-to-end Speech Recognition: Built Multi-channel End-to-end Speech Recognition from 0 to 1, and won the Excellence Award of Xiaomi Million Dollar Technology Award 2020, with produced three patents.
- Multimodal ASR / KWS: Won the 1st place in audio-visual wake word spotting and 2nd place in audio-visual speech recognition in the worldwide challenge MISP2021 under ICASSP 2022.
- Multi-channel Speech Enhancement in Video Conference: Won the 2nd place in the worldwide challenge ConferencingSpeech2021 under INTERSPEECH 2021.
- Personal ASR: Build personal ASR models for online accented users and disabled people. My work is accepted in World Internet Conference 2021.