Unsupervised learning of topologically ordered phases of matter

Submission information

Submission Number: 11

Submission ID: 28

Submission UUID: d7868540-506d-4c5d-a866-e66b7f765f35

Submission URI: /form/project

Created: Fri, 08/30/2019 - 11:57

Completed: Fri, 08/30/2019 - 11:59

Changed: Thu, 11/18/2021 - 09:44

Remote IP address: 130.215.55.243

Submitted by: Northeast Cyberteam

Language: English

Is draft: No

Webform: Project

Project Title	Unsupervised learning of topologically ordered phases of matter
Program	Northeast
Project Leader	Chris Herdman
Email	cherdman@middlebury.edu
Mobile Phone
Work Phone	(802)443-5060
Mentor(s)	Adrian Del Maestro
Student-facilitator(s)	Gebremedhin Dagnew
Mentee(s)
Project Description	Identifying and distinguishing new phases of matter is a key challenge in condensed matter physics physics. Recent work has demonstrated that machine learning techniques can be used to identify phases of matter with a broken symmetry. Certain quantum phases of matter, such as spin liquids, have a topological order that doesn’t break conventional symmetries, and thus are harder to identify. However, supervised machine learning techniques have recently been successfully used to identify topologically ordered phases. This project seeks to extend this work to use unsupervised learning algorithms to identify topological phases of matter. In particular, this project will focus on applying dimensional reduction algorithms to study topologically ordered phases of matter. These algorithms will be implemented to take advantage of GPUs and deployed on an HPC cluster.
Project Deliverables
Project Deliverables
Student Research Computing Facilitator Profile
Mentee Research Computing Profile
Student Facilitator Programming Skill Level
Mentee Programming Skill Level
Project Institution	Middlebury College
Project Address
Anchor Institution	NE-University of Vermont
Preferred Start Date	08/26/2019
Start as soon as possible.	No
Project Urgency	Already behind3Start date is flexible
Expected Project Duration (in months)
Launch Presentation
Launch Presentation Date
Wrap Presentation
Wrap Presentation Date
Project Milestones	Milestone Title: unsupervised machine learning algorithms Milestone Description: The student compared the efficacy of using different unsupervised machine learning algorithms for distinguishing data that represented phases of matter with topological order from disordered data sets. Milestone Title: analysis Milestone Description: The student developed an error analysis method for determining the statistical uncertainty of unsupervised classification algorithms. Milestone Title: scaling analysis Milestone Description: The student performed a scaling analysis of unsupervised learning algorithms to determine how the efficacy of the algorithm scaled with the size of the system and data set.
Github Contributions
Planned Portal Contributions (if any)
Planned Publications (if any)
What will the student learn?	The student working on this project will learn to how to use unsupervised dimensional reduction algorithms, including principle component analysis, t-distributed stochastic neighbor embedding, and Uniform Manifold Approximation and Projection. Using open source packages that implement these algorithms, the student will learn how to run these algorithms on GPUs, and deploy them on an HPC cluster.
What will the mentee learn?
What will the Cyberteam program learn from this project?
HPC resources needed to complete this project?
Notes
What is the impact on the development of the principal discipline(s) of the project?	The machines learning algorithms studied in this project have a wide array of applications in many disciplines. Therefore the results may have relevance to many other applications of these methods.
What is the impact on other disciplines?
Is there an impact physical resources that form infrastructure?
Is there an impact on the development of human resources for research computing?	Through this project, the student developed substantial expertise in applying machine learning methods. In particular, the student became proficient in using the sci-kit-learn Python package. This expertise will be very valuable in future research computing experiences in the student’s career.
Is there an impact on institutional resources that form infrastructure?
Is there an impact on information resources that form infrastructure?
Is there an impact on technology transfer?
Is there an impact on society beyond science and technology?
Lessons Learned
Overall results	The overall result of this project was the development of a new unsupervised learning algorithm that was applied to study quantum phases of matter. The algorithm was applied to Monte Carlo simulation data which represented both systems in a so-called topologically ordered phase of matter, and disordered systems. This algorithm allows for these data sets to be distinguished without prior labels on the data. If applied to experimental data, such an algorithm may allow for new phases of matter to be identified and classified.