Kipu Quantum adds Rimay machine learning tool to IBM Quantum Functions

Kipu Quantum is extending its capabilities within the IBM Quantum ecosystem with the launch of Rimay, a quantum feature extraction tool now available to users on the IBM Quantum Flex and Premium Plans. Rather than replacing existing machine learning models, Rimay adds a quantum step, encoding data into a Hamiltonian and running a counterdiabatic evolution on IBM’s quantum computers to generate enhanced features.

“Putting Rimay directly into the IBM Quantum Functions Catalog is an important step in making quantum machine learning easier to use,” says Dr. Tobias Grab, Managing Director of Kipu Quantum. This launch builds on Kipu’s existing Iskay Quantum Optimizer in the IBM Quantum Functions Catalog, expanding their offerings from optimization into quantum machine learning.

Rimay Integrates Quantum Feature Extraction into IBM Quantum Functions

This addition allows data science teams to enhance existing machine learning pipelines without requiring a complete overhaul of their current infrastructure or expertise in quantum computing. Rimay is specifically designed for tabular datasets where conventional machine learning performance has reached a plateau, offering a method to uncover previously hidden correlations within the data. The launch extends Kipu Quantum’s footprint on the IBM Quantum Platform, building upon the existing Iskay Quantum Optimizer already available to users, the company says.

Rather than functioning as a replacement for established machine learning models, Rimay operates as an intermediary step, enriching data before it is processed by classical algorithms. The process begins with encoding training data into a Hamiltonian, a mathematical representation that maps each feature to a single qubit. A counterdiabatic evolution is then executed on IBM’s quantum computers, generating quantum-transformed features that capture subtle relationships within the dataset.

These transformed features are then fed into existing classical models, either independently or alongside the original data, potentially improving predictive accuracy and operational efficiency. Recent trials of Rimay with a Fortune 100 company demonstrated a performance improvement in a customer-intent prediction model using the latest IBM Quantum Nighthawk processor, according to the company. The quantum-enhanced workflow integrated smoothly with the customer’s existing machine learning stack, suggesting a pathway to significant cost savings.

To facilitate adoption, Kipu Quantum has also released a credit-risk classification tutorial as part of IBM Quantum documentation, guiding users through data preparation, quantum feature map fitting and integration with conventional machine learning workflows. “Companies should not have to rebuild their machine learning stack or become quantum experts to find out whether quantum computing can improve a real business problem. With Rimay, the models stay unchanged, the stack stays unchanged, but the data becomes richer. For AI teams, that means improving what existing models can learn from the same underlying data, rather than simply adding more compute or complexity.” Scott Crowder, Vice President, IBM Quantum Adoption and Business Development, added, “We appreciate the progress Kipu Quantum is making to bring quantum computing into practical machine learning workflows. The addition of Rimay to the IBM Quantum Functions Catalog gives data scientists an accessible way to explore quantum feature extraction that complements existing classical models, and helps them investigate new correlations in their data to solve real business challenges.”

More like this

Quantum HardwareIQM tests quantum machine learning on its Emerald and Garnet chips

Machine LearningQwen3.5-2B GSM8K accuracy increased from 21.6% to 69.4% using Tree-Ising.

CybersecurityQuantum data from ORCA Computing’s PT-2 chip refines malware detection

Quantum Machine LearningResearchers Survey Quantum Machine Learning for Medical Imaging