BMO says it demonstrated quantum utility in a machine learning application on IBM systems

For the chief data officer at a regional bank

BMO said in a press release on Nov. 19 that it has demonstrated quantum utility in a machine learning application using IBM quantum computers, meaning the work was performed reliably at a scale beyond what exact traditional computing methods can practically reproduce.2

BMO said the work explores how quantum computing could help AI models identify which information matters most in complex datasets, potentially allowing them to find useful insights with less data than traditional approaches.2 The project focuses on feature selection, which BMO and its partners reworked by adapting the Proxy Quantum Games framework so that each feature of a dataset is a "player" scored on a quantum computer.12

According to BMO, a feature scores high when it moves together with the model's target variable and low when it is uninformative or repeats other unimportant features, and the scores rank the features.2

BMO said the method produces results of similar quality while requiring one to two orders of magnitude fewer samples than traditional feature selection techniques; the press release did not name the datasets or the traditional methods used for comparison.2

The work was developed with the Democritus University of Thrace on IBM quantum computers, including the IBM Quantum Nighthawk r2 processor, and was reviewed by IBM's quantum research team, BMO said.2

Kristin Milchanowski, Chief AI and Quantum Officer at BMO, said: "While we remain in the early stages of application development, these projects show how quantum computing can help solve business problems in new ways."2

Why it matters

The claim is parity with classical methods on less data, not better results, and it comes from BMO's own announcement. Banks weighing quantum pilots should treat it as an early signal on data efficiency, for example where labeled examples such as confirmed fraud cases are scarce, and not as proof of production value.

From the brief

Our take. BMO's quantum feature selection on IBM hardware matched classical methods using one to two orders of magnitude fewer data samples. Matching is not beating, so the budget case rests on data you lack, not accuracy you need.

Action. Data science: List fraud and risk models limited by scarce labeled samples; make them the required test case for any quantum pilot proposal.

What our records say

BMO runs its AI program from group level in Toronto, with the US bank as the group's growth priority, and aims to cut commercial underwriting time in half by 2028 with Aura, its first credit agent.1 Lumi Assistant gives Team BMO instant access to more than 8,000 policies and procedures in English and French.2 Aura drafts credit-assessment commentary that underwriters review.2

  1. BMO: AI programme summary (part 1)
  2. BMO: AI programme summary (part 2)

From our records

Sources

  1. ReportBMO demonstrates quantum computing milestone for AI and machine learning - FStech Fstech
  2. Primary sourceBMO Demonstrates Quantum Utility in Machine Learning Application Using IBM Quantum Computers Bmo