Enhanced Conformer-Based Speech Recognition via Model Fusion and Adaptive Decoding with Dynamic Rescoring
Speech recognition is widely applied in fields like security, education, and healthcare. While its development drives global information infrastructure and AI strategies, current models still face challenges such as overfitting, local optima, and inefficiencies in decoding accuracy and computational...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2024-12-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/14/24/11583 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|