Quantum computing inventions before the European Patent Office, the US Patent and Trademark Office and elsewhere

The first invention is by Microsoft Technology Licensing, LLC and relates to a method of operating a quantum computing device for determining a solution to a combinatorial optimization problem. Determining a solution is done by finding the ground state of a Hamiltonian through a process of estimating an energy level of a quantum state of the qubits in a quantum computer. The quantum computer prepares “trial” quantum states and iteratively refines those states for a better approximation of the solution to the combinatorial optimization problem. Although an example of the quantum computer was presented in the application as a physical system having qubits embodied in physical form, the initial wording of the patent claims defined a patent scope arguably encompassing non-physical qubits – i.e., simulated, or virtual, qubits. This arguably broadened the scope of the invention from the physical towards the abstract in a manner consequential in Europe but, as we shall see, not in the USA. The Microsoft invention was pursued in patent applications before the USPTO as patent application number US16/286,337, and before the EPO as patent application number EP19710257.7.

The USPTO position was, in the past, rather favourable as regards patenting of abstract inventions. Of course, this changed with the US Supreme Court’s decision in Alice Corp. v. CLS Bank International, 573 U.S. 208 (2014), (abbreviated “Alice”). The Alice decision significantly expanded the definition of what constitutes an abstract idea and, consequently, greatly restricted what constitutes a patentable invention. In the case of the Microsoft invention, the USPTO did not question whether the invention as defined by the patent claims was abstract in a sense necessary for the Alice decision to bite. Instead, they considered a different form of abstraction, or more precisely, of extrapolation. In particular, the USPTO patent examiner pointed out that the specification of the patent application, although not the patent claims explicitly, stated that: This meant, in the examiner’s view, that the broadest reasonable interpretation of the patent claims encompassed a fault-tolerant quantum computer. The examiner concluded that the patent claims were not allowable because:

"…[the] time to create a large fault-tolerant quantum computer ... is more than a decade away,… [with the result that] …the applicant fails to enable the full scope of the claimed invention… [and therefore]…one of ordinary skill would not be able to make and use a fault tolerant quantum computer…"

Put in other words, the patent application claimed more than was delivered by the invention, in the examiner’s view, and was refused as a result. Microsoft appealed this refusal at the Patent Trial and Appeal Board (abbreviated “PTAB”). They argued that the Microsoft invention does not rely on the existence of fault-tolerant quantum computers, but is merely presented as being suitable for use by such computers in future, if and when they become a reality, saying:

“The law does not expect an application to disclose knowledge invented or developed after the filing date [of the patent application]. Such disclosure would be impossible.”

The PTAB agreed, and noted that to require that Microsoft should have disclosed, within their patent application, how to implement a fault-tolerant quantum computer which did not exist at that time, and whose future existence was purely speculative:

“…would be to impose an impossible burden on inventors and thus on the patent system.”

In light of this, the PTAB overturned the patent examiner’s rejection and the patent was granted.

The EPO is known to have a well-developed position towards abstract inventions, such as computer-implemented inventions and mathematical methods. A different position has not been set out in decisions specifically related to quantum computing inventions, and it is understood that the EPO will apply the same principles as used for assessing classical computer-implemented inventions to quantum computer-implemented inventions. In short, if it is possible to show that an invention in the form of a mathematical method nevertheless has a ‘technical character’ by virtue of its application to a particular field of technology, or by virtue of the way it is implemented, then this moves the invention from the abstract towards the physical – and towards the patentable – at least in principle.

In the case of the Microsoft invention, the EPO questioned whether the invention as defined by the patent claims was so abstract as to lack ‘technical character’. They acknowledged that the patent claim was directed to more than a mere generic computer implementation – namely, because it involved adiabatically evolving a quantum state – however, this alone was not enough to give the claimed invention the required ‘technical character’ because the claimed invention was not directed to a specific technical implementation of the mathematical methods used. The patent claims also failed to define any technical application of the method to a field of technology and this, the EPO argued, kept the patent claim within the realm of the abstract. The quantum state and its evolution, it seems, were deemed to be of abstract character, rather than ‘technical character’. By changing how the invention was defined by the patent claims Microsoft was ultimately able to convince the EPO to grant the patent. Microsoft achieved that by adding specific limitations to the patent claims requiring the use of a quantum processor comprising qubits, and a “readout device” for reading the qubits. This was sufficient to move the patent claim from the realm of the abstract and into the ‘technical’. Although the EPO did not identify whether it was the quantum processor or the “readout device”, or both, that added the necessary technical character, it was noted that quantum processors and qubits may be simulated in a classical computer as virtual entities made to behave according to the laws of quantum mechanics. This allows quantum circuits to be executed without ever using a physically real qubit. We note that the same cannot be said for a “readout device” for reading qubits. A “readout device” provides the essential bridge from the quantum realm to the classical world by inducing the so-called “collapse” of a quantum state necessary to convert quantum information into classical information. The laws of quantum mechanics are entirely silent on the dynamics of this assumed “collapse” process – it is a postulate of quantum mechanics rather than an observed fact. The result is that the outcome of a quantum state “collapse” is probabilistic and cannot be predicted with certainty. Therefore, it cannot be simulated meaningfully using computers and inherently requires a real “technical” object to intervene. It will be interesting to follow EPO practice in this matter to see if this distinction does indeed guide their future reasoning on the patentability of quantum technology.

The second invention is by Google LLC and relates to a method of performing quantum Monte Carlo calculations using quantum computers. The Google invention claims a process for computing, by a classical computer, an accurate solution to the Schrödinger equation for a many-electron physical system such as a complex molecule. To do this, a trial wavefunction is formed using the quantum states of physical qubits in a quantum computer. The classical computer obtains data from the quantum computer representing the results of measurements of the trial wavefunction and uses those measurement results to calculate a solution to the Schrödinger equation at hand.

The European patent application initially contained claims that defined mathematical processes implemented by the classical computer after it had received the measurement data of the trial wavefunction from the quantum computer. The EPO objected that those claims merely defined a generic computer implementation of a mathematical method and lacked the necessary ‘technical character’. Google argued that a sufficient ‘technical effect’ arises from the claimed invention because it allows a hybrid quantum-classical computer to compute quantities required by the quantum Monte Carlo calculations (i.e., wavefunction properties) with a bounded number of “measurement” repetitions. In support of this argument, Google amended the patent claims to recite, amongst other things, the explicit step of:

“…performing, by the quantum computer, measurement operations on…the trial wavefunction…”

It is interesting to note that, in doing so, Google has chosen a claim amendment strategy very reminiscent of the amendment strategy adopted by Microsoft, as discussed above, to try to overcome an objection relating to non-patentable subject matter in a quantum computer invention. Furthermore, the operations of performing, by the quantum computer, measurement operations on the trial wavefunction and computing, by the classical computer, information using the data representing the results of the measurement operations, are examples of an indirect physical measurement of properties (i.e., the classical information) of a physical object (i.e., the trial wavefunction). This means, they argue, that those operations are ‘technical’ no matter what use is made of the results. Here a direct appeal is made to the reality of the wavefunction as being a physical object rather than merely an abstract mathematical tool. The unanswered question of the reality (or otherwise) of the quantum wavefunction as a physical object, and how this may influence IP for quantum technology, is discussed in detail in Mewburn Ellis’ Special Report – “Quantum and the Law”. The EPO is, at the time of writing, yet to respond to Google’s arguments, and we will monitor the progress of this case closely to see whether or not a strategy that succeeded for Microsoft will also succeed for Google.

At the time of writing, the USPTO have yet to issue their first Office Action on the US application, and we will monitor the progress of this case closely. However, a corresponding Australian patent application number AU2022301178 is pending for the Google invention.

The Australian IP office initially objected that the claims of the Australian patent application, which match those currently pending on the US application, did not define patentable subject matter. Google successfully responded to this objection with claim amendments and arguments that closely mirror those submitted on the co-pending European application, resulting in the grant of this patent.

The above examples show that many of the issues encountered for quantum computing inventions are neither especially quantum nor even new. Instead, the issues are common in related technology fields, such as computer-implemented inventions, and solutions which work in those related fields will also be applicable to quantum computing inventions. For example, keeping in mind that focussing on the physical implementation, in order to reduce the “abstract” quality may help with overcoming objections in Europe. As such, a robustly prepared patent application should mean good prospects for obtaining patent protection. In some cases, a quantum-specific discussion may arise and it may become necessary to convince a patent examiner of the merits of a quantum computing invention. The issue here may be that an examiner lacks sufficient expertise to form an understanding of the invention and to identify where the “physical” is separated from the “abstract”. Carefully crafted arguments supporting the invention and setting out the relevant context will be needed to explain the important quantum aspects of the technology and to convince the examiner of the patentability of the invention. Expert evidence supplied by researchers on particular technical points may, in future, become more important in this process. Eventually questions going to the very heart of quantum physics will become relevant as applicants attempt to secure increasingly ambitious protection. As we have seen above, amongst these questions is the matter of whether or not manipulation of a quantum wavefunction, or quantum state, is “abstract” (and therefore excluded from patentability) or “technical” (and so in principle patentable). It appears that at least some patent offices tend towards viewing wavefunctions (and quantum states generally) as abstract and therefore not patentable unless combined with hardware, such as a measurement device or operation. Presently it is unclear, however, whether this position is indeed final or whether or not it may be possible to shift at least some patent offices to adapt a different approach.