Amey Patil

Research Internship

Samsung•  May 2019 - July 2019

Investigations on Neural Speech to Speech Translation model + Study and exploration of various Sequence-to-Sequence models for Speech-to-Speech translation task + Generated target speech data used for the task of Speech-to-Speech translation from Multilingual Speech Translation (MuST-C) dataset using Neural Text-to-Speech generation + Implemented & trained an aforementioned Sequence-to-Sequence LSTM Encoder and Attention based Decoder model as a baseline on the new dataset + Proposed and Implemented new multi-task learning-driven module in Transformer architecture for End-to- End Speech-to-Speech Translation task