• Machine Learning

• Deep Learning

• Natural Language Processing (NLP)

• Speech Processing

• Evolutionary (Genetic) Algorithms

• Coding Theory and its applications

Evolutionary Generative Models for Whisper-to-Speech Conversion

This project presents a significant extension of GAN-based model in the framework of Evolutionary Genetic Algorithm for the cross-domain whisper-to-speech conversion task.

Guide: Prof. (Dr.) Stephanie Forrest

Recommendation System on Netflix Dataset

Using existing technique of sequence-to-sequence modeling (GRU cell) and GAN framework, proposed new training paradigm for generating rating of specific user based on his/her temporal history of rating.

Guide: Prof. (Dr.) Guoliang Xue

Effectiveness of Generative Adversarial Networks (GANs) in Medical Domain

Proposed novel MMSE DiscoGAN (Deep Learning Architecture) for Non Audible Murmur(NAM)-to-Whisper and Whisper-to-Speech Conversion. The piece of work in Whisper-to-Speech conversion using MMSE DiscoGAN was selected as research paper in MLSLP 2018.

Guide: Prof. (Dr.) Hemant A. Patil

Implementation of DNA encoding using run-length and GC-content constraint

Simulated the DNA encoding scheme which satisfy both run-length and GC-constraint using python language which map binary sequence in Z4 domain and use this mapping for encoding binary string into DNA.

Guide: Prof. (Dr.) Manish K. Gupta

Speech Enhancement (SE)

Motivated by the promising results of GANs in a variety of image processing tasks, I explore the potential of conditional GANs (cGANs) for SE. In particular, I make use of the image processing framework to learn a mapping from the spectrogram of noisy speech to an enhanced counterpart.

Guide: Prof. (Dr.) Hemant A. Patil