Adapting Language-Audio Models as Few-Shot Audio Learners
This paper gets accepted by INTERSPEECH 2023.
Jinhua Liang
Jinhua Liang is now pursuing his doctoral degree at Queen Mary University of London, co-supervised by Emmanouil Benetos, Huy Phan, and Mark Sandler. Before that, he received his Bachelor and Master degrees, both from Tianjin University in 2018 and 2021, respectively.His research interests mainly focus on everyday sound perception, including sound event detection, acoustic scene classification, and audio captioning. He is also interested in model compression and accelerating.
Recent Posts
Learning from Taxonomy: Multi-label Few-shot Classification for Everyday Sound Recognition
This paper is in peer review.
[Project launch] Everyday sound recognition with limited annotations
The goal is to explore general-proposed audio understanding in the limited annotations.
Channel Compression: Rethinking Information Redundancy Among Channels in CNN Architecture
This paper is accepted by IEEE Access.
Research on efficient acoustic scene recognition based on deep learning
The goal is to explore high-performance acoustic scene classification (ASC) technologies at a low computational cost.