Keys to QML Patent Success
In this co-published article, Laura Compton of Haseltine Lake Kempner takes a practical look at how to formulate claims and write applications for quantum machine learning inventions given EPO patent eligibility requirements.
In quantum machine learning (QML), classical machine learning algorithms, or their expensive routines, are usually tailored to run on a quantum computing device. QML uses quantum resources to improve the execution time and/or performance of classical machine learning algorithms.
Aspects of QML that may be patentable include using a quantum computing device to more efficiently run all or part of a classical machine learning algorithm (e.g., using a quantum computer to more efficiently compute classical distances for nearest neighbor, kernel, and clustering methods), or to run a model itself (e.g. reformulating a stochastic model into a quantum system). Other related aspects include reformulating an optimization problem so that it can be solved using a quantum computing device.
Another aspect of QML that may be patentable includes improvements to existing QML algorithms or models (e.g., an improvement that reduces the depth of quantum circuitry needed to run the algorithm or model, and/or uses less complex gates , and/or avoids the repetition of certain subroutines of the algorithm). Some improvements may be specific to the problem to be solved itself (for example, modifying the operations applied to a quantum computing device so that a more limited space of potential solutions to an optimization problem is then traversed by the device).
Inventions relating to these aspects will be considered patentable subject matter at the EPO when the quantum computing device forms an integral part of the invention.
For such inventions, the independent claims are likely to refer to the quantum computing device and the manner in which the algorithm has been adapted to be implemented thereon. Dependent claims, if not the independent claim itself, must:
- specify the initial state of the qubits of the quantum computing device;
- the variables represented by this initial state;
- how the qubits are manipulated according to the algorithm;
- the output obtained by measurement; and
- what the output represents.
Given the EPO’s “technical” requirements, it is recommended to have a dependent claim that specifies how the output of the quantum computing device and/or the output of the machine learning model is then used in a technical process.
Where the invention relates to more general QML methods or improvements to these methods (which could be applied to a wide range of problems in a wide range of fields), it is also recommended to provide a number of different use cases which demonstrate how the invention can be applied to various practical problems in the dependent claims, or description,.
Quantum computing in general, as well as QML, is a complex and rapidly evolving field. As such, drafting applications that meet the EPO’s requirements for sufficiency and clarity can be challenging. Therefore, when writing patent specifications, it is recommended to include a full mathematical description of the quantum implementation of the algorithm or model, as well as how each operation applied to qubits relates to the algorithm or model implemented (for example, describe how a series of operations applied to qubits are representative of an objective function to be minimized).
For inventions that relate to the improvement of existing algorithms or QML models, a detailed description of how the modifications to the quantum circuit achieve the improvement should be included. As in any rapidly evolving field where there is a lack of universally accepted terminologies, for applications relating to quantum computing in general, the terms used in the claims of the application should be defined in the description.
Finally, experimental data may be particularly useful in terms of demonstrating an improvement in speed or accuracy over the state of the art and may be useful in supporting inventive step arguments in a prosecution. later. One should also consider the configuration of the technical problem that the invention solves in terms of why conventional processes suffer from drawbacks that make them commercially or technically unviable (e.g., too slow for real-time deployment).
In summary, the above points can be used to help draft QML inventions that can be submitted to the EPO and can be used to provide the applicant with the best possible chance of obtaining a commercially useful patent.
Previous articles by Haseltine Lake Kempner authors in this series can be accessed here:
How to Obtain AI Patents in Europe
AI Patent Drafting for Success at the EPO – Eligibility and Formulation of Claims
Writing AI patent applications for success at the EPO – writing the full specification
Technology trends – why patent your hidden AI?
Google and Samsung top list of AI-related patent filers at EPO
EPO and UKIPO approaches to AI and patentable subject matter
How the revised EPO guidelines affect the treatment of AI inventions
Monetize data, machine learning’s most valuable asset