(348h) Quantum Integer Programming: An Open-Access Practical Guide and Toolkit for Novel Computational Paradigms for Optimization

This presentation will cover the learnings and experiences that we had during the design and teaching of the Quantum Integer Programming (QuIP) course at Carnegie Mellon University (CMU). This course was taught for the first time during the Fall of 2020, aimed to graduate students in Engineering, Operations Research, and Science. Given the strong presence of Process Systems Engineering (PSE) at the Chemical Engineering Department at CMU, many of the students taking the course were Chemical Engineers by training. Students taking this course were enrolled in different university programs, including Chemical Engineering, Computational Biology, the Business School, Physics, and Information Science. Although the course's main topics were related to Computer Science, Operations Research, and Physics, having a Chemical Engineer as one of the lecturers provided a ChemE perspective motivated by applications and examples stemming from this field.

To our knowledge, this course is the first of its kind to provide a practical approach to Quantum Computing by solving challenging optimization problems in Engineering, Business, and Science. The unique approach of this course was requiring no prior knowledge in Quantum Mechanics from the students. This allowed students who are experts in their areas, such as Engineering or Sciences, to start using unconventional computational resources to solve relevant problems in their discipline. Another one of this courses' particularities is the access it provided to hands-on experience using Quantum Computers and quantum-inspired software on specialized hardware. Through a common ground based on discrete optimization, students learned these concepts while completing an applications-related group project. We designed fully remote lectures, which could be accessed asynchronously, through the publication of lecture notes[1], online videos, and ready-to-use interactive codes (Python on Google Colab, with APIs to access the various hardware and solvers).

The motivation for designing this course was the collaboration between the Tepper School of Business at CMU, The Universities Space Research Association (USRA), and the Center of Advanced Process Decision Making at the Chemical Engineering department at CMU on exploring the use of Quantum Computing tools to solve challenging combinatorial optimization problems. To teach a practical class on Quantum Computing without requiring students to have a background in Quantum Mechanics or Theoretical Computer Science led to making Integer Programming (IP) the primary language to pose optimization problems and then solve them using unconventional computing methods. IP has had a considerable reception in Chemical Engineering education and practice, being integrated into curricula around the country as part of optimization-related courses and used heavily in different industries such as Oil & Gas.

This course's development was supported by the Universities Space Research Association (USRA) and Amazon Braket, providing the students and lecturers access to the quantum computers and the software tools required to interact with them. Furthermore, the USRA used the material developed for this course as training material for the Airforce Research Laboratory (AFRL). The material developed for the CMU course has already been used by awardees of the Feynman Quantum Academy Fellowship at USRA (20 students in the 2020-2021 year). The central focus of the QuIP syllabus on combinatorial optimization makes it relevant to a broad range of disciplines, allowing for the required diversification of the quantum-ready workforce.

In serendipitous synergy with the access opportunities demanded by the COVID pandemic, we designed all the material to be accessed remotely from the start. This included video recordings of the lectures, interactive slides, lecture notes[1], and Jupyter notebooks that could be run without any installation through Google Colab and open-access software. This material, first developed for the CMU community, was subsequently made available openly at https://bernalde.github.io/QuIP. This approach to education, together with the considerable interest in Quantum Computing, made it possible for the course to be almost immediately replicated in other institutions. For example, with only two weeks of delay with the original CMU course, the same material was taught in the Indian Institute of Technology at Madras (IIT-Madras) at its Electrical Engineering Department. Moreover, interested students from other universities and professionals in the industry have used the course individually, indicating that the material is also suitable for self-study.

We designed this course to train the future engineering workforce, one with practical knowledge of Quantum Computing. Through an approach towards the solution of practically relevant applications, this course provides a toolkit for addressing optimization problems by formulating them as IPs and then solving them using either classical or unconventional computational methods. This framework allows for the students to become “quantum-ready” effectively. Using lectures from instructors of different disciplines such as Engineering, Business, and Physics, together with lectures from invited external experts, the students obtained the skills and tools to understand the current state and limitations of this emerging field and its potential breakthroughs in the near future.

The students' evaluation was primarily a team-based final project, together with a series of individual low-stakes quick quizzes to ensure their continued progress along the course. As a final project, each group proposed an application related to its members’ background, which could be posed as a discrete optimization problem. The students tackled these problems with both classical and quantum computing resources through different problem formulations as IPs. Having students from different backgrounds in the course led to a myriad of project topics, including subatomic particle tracking, job-shop scheduling, cyclopeptide identification, bin-packing, image classification, message decoding, and quantum state tomography. By aiming to address problems in their respective field, these students significantly advanced towards being quantum-ready practitioners.

The Tepper Business School at CMU has already highlighted this course as part of its newsletter https://www.cmu.edu/tepper/news/stories/2020/october/quantum-computing-c....

Moreover, through the students' good reception of the course, other departments at the university have shown interest in it. At the Chemical Engineering Department, the Center of Advanced Process Decision Making (CAPD) has made this course material available for its students and industrial partners. Besides, this course has already motivated invited lectures by other departments and institutes at CMU, such as the Institute of Software Research and the Physics Department. It is being considered to be cross-listed with other Engineering Departments at the University.

[1] D. BERNAL, S. TAYUR AND D. VENTURELLI. 2021. Quantum Integer Programming (QuIP): Lecture Notes. arXiv: 2012.11382v2.