Speaker Recognition at the Cocktail Party

A crowded, noisy place. Do you hear me?

Imagine that you are at a cocktail party with friends in a noisy room. How do I recognize them using just your hearing? The answer is given by solving the problem of recognizing speakers in a multi-speaker environment.

This project allows you to solve the stated problem. One of the main approaches the usage of a special regularization technique - mixup augmentation [1].

References:

1. Zhang H., Cisse M., Dauphin Y.N., Lopez-Paz D. “mixup: Beyond Empirical Risk Minimization”. ICLR 2018 Conference Blind Submission (2018). Link: https://arxiv.org/pdf/1710.09412.pdf