We accordingly “held that simply implementing a mathematical principle on a physical machine, namely a computer, [i]s not a patentable application of that principle.” (Alice v. CLS Bank).
An apparatus that generates the signal is of course a machine. (In re Nuijten)
In a previous IPWatchdog article I have argued that a programmed computer is a machine, not an abstract idea. What about computers performing mathematical formulas? Do these computers not act merely as calculators in those cases, and does SCOTUS not actually have a point there?
The use of a mathematical formula in a claim can trigger a 35 USC 101 rejection or invalidation of the patented claim. However, the role of mathematics in physics and in inventions has changed dramatically over the last 50 years. To many legal professionals mathematics merely consists of formulas, with no further distinction. Fortunately, many USPTO Examiners are well trained in technology and often appropriately allow mathematically modeled inventions.
An issued patent claim with a mathematical formula remains subject to potential invalidation, certainly after Alice. The fundamental issue of the digital age is that mathematical formulas are now an essential part of devices in the digital economy. These formulas are part of executable structures that establish a physical device. Thus, the interpretation of mathematical formulas as an abstract description of reality is an outdated leftover of the industrial revolution that itself struggled with the epistemological significance of mathematics. The history of science is full of examples of doubt about the “existence” of concepts that were expressed in mathematical form. And to many people, even of academic education, mathematics and its relation to science and engineering remains somewhat opaque, to say the least.
The Courts are correct about computers, at least in part, sometimes being mere calculating machines. That is, one can feed almost any expression into a computer and provide or initiate data and generate an outcome that can be displayed on a screen. One may call that a conventional use of a computer. I would not call it an abstract idea, but more like an obvious and limited use of a computer.
However, physical reality has fundamentally changed with the use of computers. In many cases the physical reality nowadays is determined by a mathematical model implemented on a computer and making the programmed computer the actual physical functional device that replaces another physical device or does something “physical” that is otherwise not possible.
In a narrow sense, the origin of this physical transformation of data lies in the Shannon sampling theorem and the Nyquist rate criterion, which show that any analog signal is fully characterized by a limited number of discrete samples of the signal under certain conditions. This enables the quantification of the physical world. By digitizing the discrete samples, for instance in binary words which each represents a number, a physical model is now merely a set of discrete operations on numbers. (Actually, there are no numbers in a computer, which are abstract ideas, but there are signals or material states which can be represented by numbers.)
These numbers/signals can be converted back into a physical analog signal by using a device like a digital/analog (D/A) converter. Computers are programmed to operate on the numbers to perform a defined task, such as filtering or demodulation, by performing calculations in a certain order and at a certain speed. Cellphones, DVD players, telephone sets, 3D printers, cameras, MRI machines and the like are in essence computers with an A/D and D/A converter and/or transducers.
Take electronic filters as an example. There actually are different types of math that describe the performance or behavior of a filter. A first description is what is called a transfer function that illustrates the frequency behavior of a filter. For instance a transfer function of a low-pass filter shows that for low signal frequencies there is limited or no attenuation of a signal and for high frequencies there is significant attenuation. Unfortunately, the transfer function cannot be implemented as such to create a filter. One may calculate what the attenuation of the filter is for a certain signal frequency, but the transfer function formula by itself does not filter a signal. It is like E=mc2, a formula often recited in patent related discussions. Calculation of a value of E for a certain input does not release the calculated energy.
There is another type of mathematics in electric filters, which applies time-delays, clock-signals, multiplications and additions of the earlier mentioned numbers (which are actually signal samples) that generates output numbers. These output numbers, when processed by a digital/analog (D/A) converter, translate into signals that are filtered signals.
This provides the absolutely astonishing fact that the computer with its implemented mathematical model replaces an electrical filter that previously was built from wires, inductor coils, capacitors, op-amps and resistors, the ultra physical stuff. There is no doubt that a digital filter is as real as any other electronic filter. But …, the heart of the digital filter is a computer that basically performs calculations in accordance with a mathematical algorithm. We can argue that the processor does something “significantly more”, namely the filtering. But it changes nothing to the reality of the filter being implemented by basically a calculator.
One can make a similar analysis of medical images such as 3D MRI images. These 3D images do not exist as such, but are merely sets of data. They are only recognized as true images when they are displayed on a 2D output medium such as a computer screen by for instance rendering algorithms, which are calculators. However, no one would characterize MRI machines as merely abstract ideas. Other mathematical executable models exist, also outside pure electronics, such as in medicine, virtual reality, 3D printing, control systems and decision support.
This is the fundamental nature of the digital age. Data and mathematical algorithms are no longer an abstract description of reality, but are the reality in this digital age. Many people, including Congress and the Courts, seem to look at mathematics as a difficult to understand abstract language that has little to do with physical reality. It is an old phenomenon from which the “practical men” in Britain already suffered when opposing Maxwell’s theory in the late 19th century.
This does not mean that USPTO Examiners are naive about computerized machines; many are clearly not. But they are not receiving the legal guidance and support that corresponds with what many have learned in college. It seems that a true understanding what computers actually do with mathematics is missing at the Courts and in Congress and that “applying math” by a computer is incorrectly believed to be “doing what only humans would do.” Undeniable, this opinion fits well into the political and economic interests of some stakeholders and they do not actively dispute this outdated notion. But the time of the computer being a mere calculator has gone for over more than five decades now.
It is perhaps a scary thought, but increasingly, if you can model something on a computer (which is an art in itself), it is reality. However, cases litigated in courts and before the PTAB show that the “abstract idea” allegation can be applied as a strong weapon against patent owners and in my mind denies the current state of technology and the economy. I believe that the courts, with some exceptions that jump out, are not well prepared for the above type of inventions and rely on outdated interpretations that force digital age inventions into the straight jacket of the industrial revolution of mechanical machines. The requirement that the “abstract idea” has to do “significantly more” seems arbitrary and relies on what a (non-expert) reviewer thinks is conventional or abstract. It is time for Congress to end this outdated and unscientific practice and bring the patent system into the 21st century on this aspect.
The computer executable mathematical model is at the core of the digital economy. In a great number of inventions there is no longer a distinction between a mathematical principle and a physical principle, and because of the computer they are the same.
Thus the courts have it actually backwards. Mathematics in the human mind is abstract, but a programmed computer enables the physical execution of a mathematical model and thus transforms the mathematical model into a physical device or machine. Modern control theory relies entirely on that approach.
I believe the current “abstract idea” standard to be outdated, detrimental to novel digital inventions and not supported by scientific facts in many cases. Congress should change it.