UROP Proceedings 2020-21

School of Science Department of Physics 68 Machine Learning Aided Detection of the Quantum State in an Atomic Quantum Simulator Supervisor: JO Gyu Boong / PHYS Student: MAK Ting Hin / PHYS-IRE Course: UROP4100, Spring Classifying topological quantum phases have generically been a key challenge in quantum many-body physics due to the absence of local order parameter. Here, we report a successful classification of experimental images with different symmetry-protected phases by training a deep convolutional neural network (CNN). The whole phase diagram mapped out with the trained CNN gives consistent phase transition points with previous results extracted by the conventional method. Visualizing the filters and postconvolutional results shows that spin-imbalance information is important for the neural network to make the classification. Our work points out the potential of using machine learning techniques to physically analyze complex systems and large experimental datasets. Electromagnetic and Acoustic Metasurfaces Supervisor: LI Jensen Tsan Hang / PHYS Student: YANG Sandra Sang Xiao / PHYS Course: UROP1100, Fall Optical properties of metal nanomaterials can be tuned using different metals and structures. The conduction electrons of the metal can be excited under visible and near-IR light. When the electrons are collectively excited, they form coherent oscillation. This is known as the localized surface plasmon resonance. The frequency of the resonances varies with its shape, dimensions, and the surrounding dielectric background. Since the resonance frequencies are very sensitive to the dielectric environment and are easy to detect, it can be used for sensing in the nanoscale. This report will discuss the plasmon resonant frequencies of a thin film of modified Sierpinski triangle. Electromagnetic and Acoustic Metasurfaces Supervisor: LI Jensen Tsan Hang / PHYS Student: HO Tin Seng Manfred / PHYS-IRE Course: UROP1100, Summer Solving the reverse scattering problem involving elastic waves means to be able to tell the density distribution and geometry of a material from its scattered waves. This report demonstrates an experimental approach where the scattering of elastic waves on a sample metamaterial can be accurately mapped and analyzed, subject to only a few experimental errors. This can serve as a benchmark for other methods of analysis of the problem, drawing comparisons between theoretical algorithms, machine learning techniques and experimental data.