Physics & Astronomy ETDs

Publication Date

Fall 12-1-2018


In recent years, quantum information processors (QIPs) have grown from one or two qubits to tens of qubits. As a result, characterizing QIPs – measuring how well they work, and how they fail – has become much more challenging. The obstacles to characterizing today’s QIPs will grow even more difficult as QIPs grow from tens of qubits to hundreds, and enter what has been called the “noisy, intermediate-scale quantum” (NISQ) era. This thesis develops methods based on advanced statistics and machine learning algorithms to address the difficulties of “quantum character- ization, validation, and verification” (QCVV) of NISQ processors. In the first part of this thesis, I use statistical model selection to develop techniques for choosing between several models for a QIPs behavior. In the second part, I deploy machine learning algorithms to develop a new QCVV technique and to do experiment design. These investigations help lay a foundation for extending QCVV to characterize the next generation of NISQ processors.

Degree Name


Level of Degree


Department Name

Physics & Astronomy

First Committee Member (Chair)

Robin Blume-Kohout

Second Committee Member

Carlton M. Caves

Third Committee Member

Francisco Elohim Becerra

Fourth Committee Member

Gabriel Huerta




quantum tomography, qcvv, machine learning, NISQ, QIP

Document Type