I am a PhD student at the Colorado School of Mines interested on quantum information and machine learning.
My coding abilities involve simulation of closed and open Quantum Many-Body physics in quantum Ising Models and bosonic systems, using matrix product states, time-evolving block decimation (TEBD) and exact diagonalization, being aware of the limitations within each framework. Furthermore, I got exposure to modern complex network measures (such as, disparity, clustering coefficient, Pearson correlation coefficient, and density.) as quantifiers of Quantum Phase Transitions.
I am familiar with topics related to different kinds of classical noise involving stochastic processes modelled by random telegraph noise (RTN), colored RTN noise, and unbalanced RTN noise. These models are useful to study lattice defects or the relaxation processes under thermal fluctuations, leading to Non-Markovian dynamics. We applied those flavors of noise to continuous-time quantum walks and Bose-Hubbard models to explore their potential to create exotic behavior or complexity as they break generalized parity and time-reversal symmetries of certain quantum lattices setups like a quantum Ratchet ( i.e. an optical ring trap rotated off center), while keeping always track of the quantumness of the system by means of the purity, inverse participation
ratio and coherences as quantum measures.
Currently I am working on hybrid quantum-classical generative models with latent variables involving Variational Auto Encoders (VAEs), RBMs, DBMs and BMs. In the case of VAEs, the classical Deep learning is implemented through gated CNNs that learn efficient encodings of the original dataset. This compressed representation is suitable to be fed to the VAE’s latent space here formalized by a bipartite Boltzmann machine, and trained
via log-likelihood (LL) maximization. The LL gradients are approximated via sampling,
which requires computationally expensive MCMC chains, thus, given this classically challenging task of approximating Gibbs distributions, quantum annealers operating as samplers of classical or quantum thermal states replace this slow MCMC shemes. In particular, we use a non-conventional annealing protocol so-called Random Frequency Quantum Annealing (RFQA), a promising candidate to offer noise tolerant speed ups for optimization and sampling tasks.
Carla Mariela Quispe Flores
Applied Physics PhD Candidate
Undergraduate institution and degree with GPA: Mayor de San Andres University, Physics 3.8
Mines email address: carla.mariela729@gmail.com
Website (or social media page):
https://www.youtube.com/@carlamariela2763/playlists