Polypilot product mascot

Introducing PolyPilot:

Our AI-Powered Mentorship Program

Learn More
Go to Polygence Scholars page
Ram Sivaraman's cover illustration
Polygence Scholar2022
Ram Sivaraman's profile

Ram Sivaraman

LASA High SchoolClass of 2025Austin, Texas



  • "Which noise-adaptive qubit mapping method is best for minimal error in a cryptographic hashing function?" with mentor Yash (Nov. 1, 2022)

Project Portfolio

Which noise-adaptive qubit mapping method is best for minimal error in a cryptographic hashing function?

Started Apr. 16, 2022

Abstract or project description

Cryptographic hash functions take some input and uniquely map it to a protected output. The speciality of these functions is that they are not easily reversible so that it is practically impossible to get back the input given the output. There are several methods already in place for creating a secure cryptographic hash function. I would like to focus on the using MFCCs and NSSCs from the human voice to create a hash function. MFCCs and NSSCs are coefficients that succinctly represent the envelope of the human voice by looking at energies from different parts of the vocal tract. This voice data is processed through a BiGAN network that convolves this data and produces a unique vector. This vector represents the unique cryptographic key. To improve the performance, especially since the GAN can have several parameters, I would like to implement a quantum GAN instead of a classical GAN. However, this assumes an ideal case where there is no noise between quantum gates. This project researches and compares how different noise-adaptive qubit mapping methods minimize error due to unwanted environmental or qubit interactions in this quantum GAN, allowing us to produce a highly secure cryptographic hash function.