Investigation of Quantum Support Vector Machine for Classification in NISQ era

Speaker: Anekait Kariya
Quantum Machine Learning Researcher at Bikash's Quantum || Quantum Computing || BITS Goa.

Please join my meeting from your computer, tablet or smartphone.
https://global.gotomeeting.com/join/612278597

Linkedin: https://www.linkedin.com/in/anekait-kariya-a55ab114a/

Machine learning and artificial intelligence has been transforming our lives for several decades. These techniques rely heavily on data to be able to make accurate predictions, but today global data is in the order of zettabytes which has started to become a limit to be processed by classical computers.

Researchers from all over the world are turning to quantum computing for a solution, giving rise to the field of quantum machine learning (QML). Hope arising from the concept of quantum random access memory (QRAM) which provides exponential compression in representation of data.

Here, we first look at the concept of quantum data and QRAM, then quantum kernel classifier and its circuit. Later we dwell into the quantum support vector machine (QSVM) algorithm and its circuit. We compute the efficiency of the QSVM circuit implementation for two datasets; 6/9 and banknote.

We propose a general encoding procedure extending the QSVM circuit approach. We highlight the technical difficulties in applying the QSVM algorithm on current NISQ era devices. Lastly, we propose a new method, with shallow quantum circuits and enhanced efficiencies for both datasets.

QML is an emerging field and cutting edge research is happening as we talk. We conclude by exciting viewers about open questions and challenges ahead, to make the quantum hype a reality soon.