Improved far-field speech recognition using Joint Variational Autoencoder

no code implementations • 24 Apr 2022 • Shashi Kumar, Shakti P. Rath, Abhishek Pandey

Automatic Speech Recognition (ASR) systems suffer considerably when source speech is corrupted with noise or room impulse responses (RIR).

Automatic Speech Recognition Automatic Speech Recognition (ASR) +3

A Mixture of Expert Based Deep Neural Network for Improved ASR

no code implementations • 2 Dec 2021 • 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.

Automatic Speech Recognition Automatic Speech Recognition (ASR) +1

A higher order Minkowski loss for improved prediction ability of acoustic model in ASR

no code implementations • 2 Dec 2021 • Vishwanath Pratap Singh, Shakti P. Rath, Abhishek Pandey

In this paper we have proposed higher order minkowski loss (4th Order and 6th Order) during inference time, without any changes during training time.

Automatic Speech Recognition Automatic Speech Recognition (ASR) +1