Amit Meghanani

Research Students, VoiceBase

I am a PhD student at UKRI CDT in Speech and Language Technologies under the supervision of Prof. Thomas Hain , University of Sheffield. I am also a member of SPandH and Machine Intelligence for Natural Interfaces (MINI) research group.
Currently, I am working on self-supervised learning for speech, disentangled speech representations, cost-effective self-supervised fine-tuning of self-supervised learning based speech models for task-specific speech representations. Acoustic word/sub-word embeddings.
Before joining my PhD, I worked as a senior data scientist at Publicis Sapient. Before that, I was a Project Associate at the EE department of Indian Institute of Science, Bangalore, working for Medical Intelligence and Language Engineering lab under the supervision of Prof. A. G. Ramakrishnan. I have finished my M.Tech Research from the same lab. During my masters, I have worked on various research topics like speaker recognition, speech recognition, natural language processing for social media , applied machine and deep learning for recognising cognitive impairments using clinical data.

Personal webpage: https://sites.google.com/view/amitmeghanani/home

Publications

2024

Amit Meghanani; Thomas Hain: SCORE: Self-supervised Correspondence Fine-Tuning for Improved Content Representations. In: IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), Forthcoming. (Type: Proceedings Article | )

Amit Meghanani; Thomas Hain : Improving Acoustic Word Embeddings through Correspondence Training of Self-supervised Speech Representations. In: Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (EACL 2024), Forthcoming. (Type: Proceedings Article | )

2023

Amit Meghanani; Thomas Hain: Deriving Translational Acoustic Sub-Word Embeddings. In: IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU), 2023. (Type: Proceedings Article | Links)