Blind Source Separation via Quantum Slime Mould Algorithm

By Hein Htut

Source separation, as the name suggests, separates the combined output of multiple sources into some representation of each individual source. Each audio source has its own unique overtone which we can use to identify them with, however, it also causes the massive overlapping frequencies that sum up to an incredibly noisy and messy wave representation of the overall mixed audio source. Even having sufficient data, it requires massive computational power to not only identify the wanted overtone as some unique pattern in the evolution of that audio source's frequencies over time but also separate that one waveform from the bigger overall waveform. Thus, this is where we can use quantum computing, specifically a Quantum CNN to speed up computation.

Research Poster

Research Presentation Slides

Presentation Video (FCPS Only)

Photo Gallery

     Photos are coming soon!

Updates