3 papers
eess.AS2024
ChildAugment: Data Augmentation Methods for Zero-Resource Children's Speaker Verification
Vishwanath Pratap Singh, Md Sahidullah, Tomi Kinnunen
The accuracy of modern automatic speaker verification (ASV) systems, when trained exclusively on adult data, drops substantially when applied to children's speech. The scarcity of…
eess.AS2021
A Mixture of Expert Based Deep Neural Network for Improved ASR
Vishwanath Pratap Singh, Shakti P. Rath, Abhishek Pandey
This paper presents a novel deep learning architecture for acoustic model in the context of Automatic Speech Recognition (ASR), termed as MixNet. Besides the conventional layers, s…
eess.AS2021
A higher order Minkowski loss for improved prediction ability of acoustic model in ASR
Vishwanath Pratap Singh, Shakti P. Rath, Abhishek Pandey
Conventional automatic speech recognition (ASR) system uses second-order minkowski loss during inference time which is suboptimal as it incorporates only first order statistics in…