AI-Powered Tool for Quantum Materials
This project empowered UBC Blusson Quantum Matter Institute (UBC Blusson QMI) to develop a centralized, user-friendly tool utilizing the power of Amazon Web Services (AWS), AI, and quantum computing for material discovery and other quantum applications. The tool is an interactive, AI-powered platform that allows researchers to navigate and engage with quantum materials science resources more effectively, fostering discovery, and supporting advancements, in the field.
Laboratory for Atomic Imaging Research (LAIR)
Quantum matter research is fundamental, not only to advancing our understanding of exotic material physics, but also, as the backbone of emerging quantum technologies.
Quantum computing, meanwhile, is expected to be profoundly disruptive, with materials simulation identified as one of its most significant early applications.
Despite the inherent connections between these two fields, there is surprisingly little cross-pollination of research between them.
As new AI generative tools emerge, there is a growing opportunity to bridge this divide and accelerate progress by integrating these related yet distinct areas. While universities and corporations have started to provide resources, many existing tools are designed for advanced users, and research data remains fragmented across various platforms. Developing a centralized, user-friendly tool could address these challenges; fostering greater innovation, collaboration and discovery within UBC Blusson QMI and beyond.
Quantum definition: Quantum physics is a branch of science that explains how particles at the atomic and subatomic level behave. These tiny particles act in ways determined by probabilities and chances.
Approach
The UBC Blusson QMI collaborated with the University of British Columbia Cloud Innovation Centre (UBC CIC) to develop an interactive application prototype that leverages generative artificial intelligence (Gen AI) and natural language processing (NLP) to simplify access to QMI’s research in quantum matter. At its core, it is a large language model (LLM) that integrates a database of materials from QMI—such as research outputs, datasets, and publications produced by the institute. By enabling users to query this database in a conversational manner, the LLM transforms complex quantum research into accessible and clearer information
Users can explore quantum materials by asking questions and receiving detailed, tailored answers. This conversational tool supports a wide range of queries, from specific concepts to broader topics, helping users deepen their understanding of quantum research. By offering personalized insights, it encourages discovery and fosters greater engagement with QMI’s research, making quantum materials more accessible to researchers.
Screenshots of UI
General User View
Researchers and those wanting to interact with Quantum AI can log in and start a new chat relating to one of the topics.
Users can ask questions relating to the content provided and the solution will provide an accurate and helpful response. The user can ask follow up questions and continue the conversation.
Administrator View
Administrators can manage the conversation topics as well as create new ones.
After selecting “Edit”, they can adjust the title of the topic, the prompt, and the associated files.
Administrators can also view analytics relating to user engagement with the topics.
Administrators can view further insights into each topic.
Architecture Diagram
To learn more about the architecture diagram, visit the Architecture Deep Dive on our GitHub repository.
Technical Details
The solution employs a serverless architecture to provide a scalable, efficient, and user-friendly platform while reducing the complexity of infrastructure management. Hosted on AWS Amplify, the application offers a seamless interface for both users and administrators, connecting securely to the backend via Amazon API Gateway.
Administrators can easily upload topics and supporting materials, such as documents or datasets, which are securely stored in Amazon S3 using pre-signed URLs to ensure safe file transfers. These uploads are processed by AWS Lambda functions that extract, segment, and transform the text into vector representations using Amazon Bedrock’s embedding capabilities. The resulting vectors are stored in Amazon RDS for PostgreSQL, with pg_vector enabling fast and efficient data retrieval.
Users interact by selecting a topic and submitting queries. The platform utilizes a retrieval-augmented generation (RAG) approach, combining the user’s query with relevant stored data to generate accurate and contextually rich responses. This process is powered by the Llama-3.1-70B LLM hosted on Amazon Bedrock, ensuring high-quality feedback tailored to user inputs.
To enhance the user experience further, the application tracks key engagement metrics such as session durations and message activity. These insights are visualized for administrators, helping identify popular topics and guiding the addition of new resources to meet user needs effectively.
Checkout the solution on GitHub.
Video
Acknowledgements
This project was created in collaboration with the Stewart Blusson Quantum Matter Institute.
Image from the UBC Blusson Quantum Matter Institute.
About the University of British Columbia Cloud Innovation Centre (UBC CIC)
The UBC CIC is a public-private collaboration between UBC and Amazon Web Services (AWS). A CIC identifies digital transformation challenges, the problems or opportunities that matter to the community, and provides subject matter expertise and CIC leadership.
Using Amazon’s innovation methodology, dedicated UBC and AWS CIC staff work with students, staff and faculty, as well as community, government or not-for-profit organizations to define challenges, to engage with subject matter experts, to identify a solution, and to build a Proof of Concept (PoC). Through co-op and work-integrated learning, students also have an opportunity to learn new skills which they will later be able to apply in the workforce.
