Search papers, labs, and topics across Lattice.
Tongji University
1
0
2
0
Fine-tuning large audio-language models for emotion recognition can be drastically improved by leveraging hyperbolic geometry, leading to better performance on class-imbalanced datasets.
