Skip to content
Quantum.world

The Signal

BMO says quantum feature selection used fewer samples

BMO says its quantum feature-selection method produced results of similar quality to traditional techniques while requiring one to two orders of magnitude fewer samples. The bank says the application was developed with Democritus University of Thrace and run on IBM quantum computers.

BMO says quantum feature selection used fewer samples

Quantum game ranks machine-learning inputs

BMO says its quantum-computing approach to selecting features for machine-learning models produced results of similar quality to traditional techniques while requiring one to two orders of magnitude fewer samples. The project was developed with Democritus University of Thrace and run on IBM quantum computers, including the IBM Quantum Nighthawk r2 processor; IBM’s quantum research team reviewed the work.

Source newsroom.bmo.com

The application adapts the Proxy Quantum Games framework, turning feature selection from a classical statistical task into what BMO describes as a quantum strategic game. Each feature in a dataset acts as a “player.” Its score rises when it moves together with the model’s target, or outcome variable, and falls when it is uninformative or repeats other, unimportant features. The scores are then used to rank the features.

BMO says the feature-selection process runs on a quantum computer and yielded results of similar quality with substantially fewer samples than traditional feature-selection techniques.

Utility claim and application stage

BMO characterizes the demonstration as “quantum utility,” defining that as work performed reliably at a scale beyond what exact traditional computing methods can practically reproduce. The bank’s description makes this a claim about the scale and reliability of the work, alongside its reported comparison in sample use.

Kristin Milchanowski, BMO’s Chief AI and Quantum Officer and director of its Institute for Applied Artificial Intelligence & Quantum, said the bank’s application development remained in its early stages. The project explored whether quantum computing could help AI models identify important information in complex datasets and potentially find useful insights with less data.

BMO lists possible banking applications as more accurate client modeling, improved risk assessment and detection, and more personalized client experiences. These are potential uses of the capability, rather than applications the release says are already in operation.

Part of BMO’s quantum work

BMO joined the IBM Quantum Network in February 2025. The bank also cited partnerships with the Chicago Quantum Exchange and Quantum Industry Canada as part of its work on quantum applications in financial services.

BMO says it launched its Institute for Applied Artificial Intelligence & Quantum as an enterprise-wide center focused on responsible AI application and governance and on advancing quantum capabilities. The bank describes the feature-selection project as part of its broader strategy to apply quantum technologies to AI and machine learning.

Source

  1. BMO Financial Group

Related stories

Explore

More articles