I'm researching neural network quantum states and their applications in simulating quantum many-body systems. In particular, I currently focus on time evolution and improving our understanding of such machine learning approaches for quantum physics.

I specialized on quantum field theory (in condensed matter physics), algebraic topology, probability theory and machine learning. For my thesis project I implemented neural quantum states to investigate fracton models, three-dimensional highly entrangled stabilizer codes with extensive ground state degeneracy. Using a correlation-enhanced RBM, we found exact parametrizations of fracton ground states and mapped out a first-order phase transition in the checkerboard model [Machaczek, Pollet, Liu; Scipost Physics 2025].

I wrote my thesis on the tensorial-kernel support vector machine [Greitemann, Liu, Pollet; PRB 2019], in particular how to process results obtained from this algorithm to automatically detect and interpret phases in frustrated magnetic systems. This led to contributions towards [Rao, Liu, Machaczek, Pollet; PRR 2021].

I worked on bayesian neural networks to improve uncertainty estimation for medical image classification. In particular, we explored parametrized quantum circuits for weight generation to improve sample efficiency.

In the beginning I contributed towards [Rao, Liu, Machaczek, Pollet; PRR 2021] by analyzing results obtained from the tensorial kernel support vector machine. Then, I explored using topological data analysis to classify phases of classical spin systems.