May 2021 – Aug 2022 · Punjab, India
MS Research Student
NMR Quantum Computing Group, IISER Mohali (PI: Prof. Kavita Dorai)
Overview
Neural networks for multipartite entanglement detection on a three-qubit NMR processor.
What I worked on
- Designed an ANN that detects genuine multipartite entanglement and assigns one of six SLOCC classes to three-qubit states: 128 real features (Re and Im of the 64 elements of ρ), reduced to 18 (14 Re, 4 Im) by the canonical form, then ranked by ANOVA within the 18.
- Validated on 30 experimental three-qubit states from an NMR processor: GME detection at 0.954 ± 0.046 accuracy with 4 features, SLOCC classification at 0.867 ± 0.090 with 6. Equal contribution with V. Gulati.
Problem, methods and outcomes
My MS research in Prof. Kavita Dorai's NMR Quantum Computing Group used machine learning to detect and classify entanglement. In the BS–MS thesis I trained neural networks, K-nearest neighbours, PCA + K-means and a GAN-based classifier on simulated two-, three- and four-qubit states, and tested a genetic-algorithm approach to NMR pulse design. The three-qubit work was developed with V. Gulati into the published paper: an ANN that detects genuine multipartite entanglement and assigns the SLOCC class from a reduced set of density-matrix elements, validated on 30 experimental states from the NMR processor.