Loading...

Detection of Quantum Phase Transition with Machine Learning

Emami Kopaei, Ali | 2020

806 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 53035 (04)
  4. University: Sharif University of Technology
  5. Department: Physics
  6. Advisor(s): Langari, Abdollah
  7. Abstract:
  8. Detecting phase transitions is one of the challenging problems in condensed matter physics. For systems, which show phase transitions, in which an order parameter smoothly becomes nonzero, identifying critical points needs finite-size scaling of very large systems. There also exist phase transitions in nature, that the order parameter is not precisely specified. Hence the detection of the phase transitions is a difficult task. Machine Learning methods are supposed to be powerful tools for investigating phase transition. In this thesis, we first introduce the structure of machine learning algorithms and describe the corresponding building blocks. We then introduce neural networks algorithms used in condensed matter physics and examine their applications in the detection of phase transition in various systems. Focusing on the Axial Next-Nearest-Neighbor Ising model in the high frustrated regime, we classify three distinct phases including paramagnetic, anti-phase and floating phases as an example of infinite-order phase transition. In this vein, we investigate the phase transition of the outlined system using different neural network architectures, such as feedforward and deep neural network, as well as learning by confusion,. At the end, we introduce a new neural network algorithm that is simultaneously trained through two (or more ) different types of physical input data, which will be able to comprehensively identify the rich phase diagram of ANNNI model using supervised learning
  9. Keywords:
  10. Machine Learning ; Artificial Neural Network ; Quantum Phase Transition ; Learning by Confusion ; Condensed Matter

 Digital Object List

 Bookmark

...see more