Improvement of Accent Classification Models Through Grad-Transfer From Spectrograms and Gradient-Weighted Class Activation Mapping

[EN] Automatic accent classification is an active research field concerning speech processing. It can be useful to identify a speaker's region of origin, which can be applied in police investigations carried out by Law Enforcement Agencies, as well as for the improvement of current speech recog...

Descripción completa

Detalles Bibliográficos
Autores: Carofilis Vasco, Roberto Andrés, Alegre Gutiérrez, Enrique, Fidalgo Fernández, Eduardo, Fernández Robles, Laura
Tipo de recurso: artículo
Estado:Versión aceptada para publicación
Fecha de publicación:2023
País:España
Institución:Universidad de León
Repositorio:BULERIA. Repositorio Institucional de la Universidad de León
OAI Identifier:oai:buleria.unileon.es:10612/23238
Acceso en línea:https://ieeexplore.ieee.org/document/10190103
https://hdl.handle.net/10612/23238
Access Level:acceso abierto
Palabra clave:Informática
Ingeniería de sistemas
Supervised learning
Learning-to-rank
Influence detection
Feature extraction
Darknet
Tor hidden services
3304.05 Sistemas de Reconocimiento de Caracteres
5701.04 Lingüística Informatizada
1203.04 Inteligencia Artificial
1209.03 Análisis de Datos
Descripción
Sumario:[EN] Automatic accent classification is an active research field concerning speech processing. It can be useful to identify a speaker's region of origin, which can be applied in police investigations carried out by Law Enforcement Agencies, as well as for the improvement of current speech recognition systems. This article presents a novel descriptor called Grad-Transfer, extracted using the Gradient-weighted Class Activation Mapping (Grad-CAM) method based on convolutional neural network (CNN) interpretability. Additionally, we propose a methodology for accent classification that implements Grad-Transfer, which is based on transferring the knowledge acquired by a CNN to a classical machine learning algorithm. The article works on two hypotheses: the coarse localization maps produced by Grad-CAM on spectrograms are able to highlight the regions of the spectrograms that are important for predicting accents, and Grad-Transfer descriptors computed from audios represent distinctive descriptions of the target accents. These hypotheses were demonstrated experimentally, clustering the generated Grad-Transfer descriptors according to the original accent of the audios using Birch and k -means algorithms. We carried out experiments on the Voice Cloning Toolkit dataset, seeing an increase of macro average accuracy, and unweighted average recall in the results obtained by a Gaussian Naive Bayes classifier up to 23.00%, and 23.58%, respectively, compared to a model trained with spectrograms. This demonstrates that Grad-Transfer is able to improve the performance of accent classification models and opens the door to new implementations in similar tasks.