Learn how to deploy artificial intelligence to control and calibrate semiconductor quantum computing chips. Quantum computing is a fast-growing technology and semiconductor chips are one of the most promising platforms for quantum devices. The current bottleneck for scaling is the ability to control semiconductor computing chips quickly and efficiently.
Class Deals by MOOC List - Click here and see EdX's Active Discounts, Deals, and Promo Codes.
This course, aimed at students with experience equivalent to a master’s degree in physics, computer science or electrical engineering introduces hands-on machine learning examples for the application of machine learning in the field of semiconductor quantum devices. Examples include coarse tuning into the correct quantum dot regime, specific charge state tuning, fine tuning and unsupervised quantum dot data analysis.
After the completion of the course students will be able to:
- assess the suitability of machine learning for specific qubit tuning or control task and
- implement a machine learning prototype that is ready to be embedded into their experimental or theoretical quantum research and engineering workflow.
This course is part of the Quantum 301: Quantum Computing with Semiconductor Technology Professional Certificate.
What you'll learn
- To understand the utility of machine learning in tuning of semiconductor quantum devices
- To formulate various stages of tuning as a machine learning problem
- To develop and implement in Python a machine learning prototype for variety of semiconductor qubit tuning tasks
- To assess the suitability of machine learning in specific semiconductor quantum computing experimental workflows
Syllabus
Week 0: Introduction to the course and self-study of the prerequisites
Week 1: Supervised learning for quantum dot configuration tuning
- Review of neural networks
- Formulate configuration tuning as a neural network learning task
- Applicability for quantum experiments
- Coding demonstration: Supervised supervised neural network configuration classification
Week 2: Charge tuning with neural networks
- Introduction to charge tuning
- Tuning to specific charge states as supervised neural network with feedback loop
- Experimental charge tuning
- Coding demonstration: Charge charge state preparation using neural network with feedback loop
- Midterm exam (multiple choice)
Week 3: Unsupervised learning for analysis of quantum dot data
Introduction to unsupervised learning
Clustering methods for analysis of charge stability diagrams
Outlook and applicability to experimental systems
Coding demonstration: kernel-PCA clustering of charge stability data
Week 4: Fine-tuning with neural networks
- Introduction to fine-tuning
- Fine Fine-tuning as a Hamiltonian learning problem
- Experimental fine-tuning
- Coding demonstration: Hamiltonian learning for qubit characterization
Week 5: Conclusion and Recap
- Overview of the techniques and applications
- Outlook for artificial intelligence as a tool for control and calibration of quantum devices
- Final exam - multiple choice and optional project (video brief) with a forum for questions
Necessary prerequisites:
- Programming in Python
- Basic familiarity with quantum dots
Recommended prerequisites:
- Introductory knowledge of neural networks (we will provide reading material and review this concept at the beginning of the course, but some previous knowledge will better facilitate your learning).
- Basic familiarity with PyTorch (we will take time to explain the code in detail, but looking at the PyTorch package before the course starts will be very helpful).
