Voice Signal Processing for Machine Learning. The Case of Speaker Isolation

arXiv:2403.20202v1 Announce Type: cross
Abstract: The widespread use of automated voice assistants along with other recent technological developments have increased the demand for applications that process audio signals and human voice in particular. Voice recognition tasks are typically performed using artificial intelligence and machine learning models. Even though end-to-end models exist, properly pre-processing the signal can greatly reduce the complexity of the task and allow it to be solved with a simpler ML model and fewer computational resources. However, ML …

abstract applications artificial artificial intelligence arxiv assistants audio automated case cs.lg cs.sd demand eess.as human intelligence machine machine learning machine learning models process processing recognition signal speaker tasks type voice voice assistants voice recognition