Is IBM’s Watson Still Patent Eligible?

By now most are likely already familiar with the unfortunate reality that the United States Court of Appeals for the Federal Circuit issued a non-decision in CLS Bank v. Alice Corporation on Friday, May 10, 2013. There were 10 judges who heard the case en banc, with 7 of the 10 finding that the method claims and computer-readable medium claims were not patent eligible. While there may be reasonable room for a difference of opinion relative to those claims, it was the system claims that specifically and clearly recited tangible structure that has thrown the patent law of software into such disarray. 5 Judges would have found that the systems claims were patent ineligible (Judges Lourie, Dyk, Prost, Reyna and Wallach), and 5 Judges would have found the systems claims were patentable subject matter (Chief Judge Rader, Judges Newman, Moore, Linn and O’Malley). For more see Federal Circuit Nightmare in CLS Bank and 5 CAFC Judges Say Computer Patentable, Not Software and Did the CAFC Ignore the Supreme Court in CLS Bank?

Today, however, I want to write about one of the more bizarre passages I have ever seen in any decision, and then pose an almost unthinkable question: Is IBM’s Watson still patent eligible in the view of Judges Lourie, Dyk, Prost, Reyna and Wallach?

First, let’s start with the passage. Judge Lourie, who was joined by Judges Dyk, Prost, Reyna and Wallach, actually wrote: “At its most basic, a computer is just a calculator capable of performing mental steps faster than a human could. Unless the claims require a computer to perform operations that are not merely accelerated calculations, a computer does not itself confer patent eligibility.”

Thus, Judge Lourie and those who joined him in this opinion are of the belief that it is not enough that a method be performed faster or more efficient. That is utterly astonishing and will have significant and almost certainly unintended consequences. Innovations relating to faster and more efficient processing have always been patentable. In fact, many methods that do nothing more than speed up the end result have been patented for hundreds of years, and many inventors are in the Inventors Hall of Fame for such inventions.

Judges Lourie, Dyk, Prost, Reyna and Wallach in their infinite wisdom think that doing something faster and more efficient than a human does not in and of itself mean the process is patent eligible. Astonishing is the only word that comes to mind, but that doesn’t begin to touch the depth of my feelings. They are throwing away hundreds of years of patent jurisprudence and they probably never even considered that was what this simple statement was doing.

Obvious But Irrelevant

Before proceeding, allow me to point out just how ridiculous the statement made by Judge Lourie really is. Software performs all kinds of tasks faster than a human could. Furthermore, in many contexts the fact that the software directs the computer to act quickly enough to juggle a multitude of complex calculations means the difference between the task being accomplished and the task NEVER being accomplished.

There is no way that a human could perform real-time GPS navigation of a missile, for example. So the fact that a human could do the calculations if they had infinite time over infinite lifetimes is a nice, yet obvious, observation. But even infinite time over infinite lifetimes wouldn’t allow for real-time GPS navigation of a missile.

Even infinite time over infinite lifetimes wouldn’t allow for any group of people to perform the number of operations that IBM’s Watson can perform in a single second.

What about Watson?

IBM’s Watson computing system broke new ground in 2011 when it defeated two celebrated human competitors on Jeopardy! But what is Watson? It is an artificially intelligent computer system that is capable of answering questions presented in natural language. It is, in essence, the modern day equivalent to the all knowing Star Trek computer. It is flat out ridiculous to be asking whether the Star Trek omnipotent computer could be patent eligible, but that is where we find ourselves because what makes the computer unique is the software that makes it possible for Watson to perform 80 trillion operations per second.

80 trillion operations per second is very impressive indeed, but if the question is whether a human could perform those operations albeit slower, then Watson must be patent ineligible in the view of Judges Lourie, Dyk, Prost, Reyna and Wallach.

But isn’t this really where it starts to get absolutely ridiculous? Watson can be used to save lives and help improve the process for making a medical diagnosis, which in fact is one area where IBM is deploying the Watson technology. None other than Sloan Kettering Cancer Center is deploying Watson in cooperation with IBM to assist oncologists.

Let’s put this into some perspective. The IBM website explains:

According to one expert, only 20 percent of the knowledge physicians use to make diagnosis and treatment decisions today is evidence based. The result? One in five diagnoses are incorrect or incomplete and nearly 1.5 million medication errors are made in the US every year. Given the growing complexity of medical decision making, how can health care providers address these problems?

Let me answer this… Doctors could leverage a super computer that can process 80 trillion operations per second so that they can make better, more informed decisions. Watson is very much faster than a human and it should be patent eligible precisely because it is faster than a human and provides greater accuracy in real time. What good is the right diagnosis after the patient has already died? The issues are not trivial despite what several of the learned Judges of the Federal Circuit may believe, and ignoring the power of software and trivializing it because it merely calculates faster than a human is insulting.

Some Illustrative Watson Claims

Before proceeding to look at claims let me be clear. Watson is patented and it SHOULD be patented and any Judge that would ever contemplate finding Watson patent ineligible should simultaneously resign or face impeachment for incompetence. I am going through this exercise step by step to demonstrate just how absurd the outcome of CLS Bank actually is.

Yet, the fact remains that Judges Lourie, Dyk, Prost, Reyna and Wallach have said that the fact that a process can be done faster than a human is of no relevance to the patent eligibility inquiry. Thus, we should probe that to see if it holds water when put to the test, just as the Supreme Court will do if the case is accepted by the Court.

Watson has many patents, but one patent that I am aware of is U.S. Patent No. 8,275,803. Claims 1 through 19 all cover a computer-implemented method. Claim 1, which is illustrative, recites:

A computer-implemented method of generating answers to questions based on a corpus of data, said method comprising:

receiving an input query;

performing a query context analysis upon said input query to break down said input query into query terms, said query terms comprising searchable components;

conducting a first search in said corpus of data using one of more of said searchable components to obtain documents potentially including candidate answers, wherein all documents potentially including candidate answers are stored in a data storage device;

analyzing all of said documents and each document’s metadata, in a candidate answer generation module, to generate a set of candidate answers;

conducting a second search in said corpus of data using said candidate answers and said searchable components of said query terms to obtain one or more supporting passages, wherein said supporting passages have at least one of said candidate answers and at least one of said searchable components of said query terms;

scoring said candidate answers using said supporting passages, wherein said scoring is carried out by a plurality of parallel implemented scoring modules, each scoring module producing a candidate score;

selecting one or more query answers based on said candidate score;

generating a query response based on said one or more query answers for delivery to a user,

wherein each said plurality of parallel implemented scoring modules for scoring all candidate answers using the said supporting passages automatically conducts, in parallel, one or more analyses each producing a candidate score, wherein one candidate score comprises a term match score obtained by implementing executable instructions for counting the number of terms in said supporting passage and determining if said number matches a number of terms in a candidate answer; and,

wherein a further score comprises a textual alignment score obtained by said processor implementing executable instructions for determining if placement of words in said supporting passages are in alignment with placement of words of said candidate answers; and,

wherein a further score comprises a deeper analysis score obtained by said processor implementing executable instructions for determining the meaning of the supporting passages and input queries by analyzing lexical or semantic relations.

(emphasis added to draw attention to structural, tangible components).

There is very little structure recited in this claim. If the Federal Circuit is correct about the method claims of CLS Bank v. Alice Corp., then I just don’t see how this claim is patent eligible today. Maybe the meager references to tangible components would be enough for Chief Judge Rader and Judge Moore, but if Judges Lourie, Dyk, Prost, Reyna and Wallach were to apply consistent logic and reasoning to these claims they would have to say it isn’t patent eligible because it just does things faster than a human could, which is not enough.

What about the system claims from the Watson ‘803 patent? Claim 20 recites:

A system for generating answers to questions based on a corpus of data comprising:

a memory storage device;

a processor in communication with said memory storage device and configured to perform a method comprising:

receiving an input query;

performing a query context analysis upon said input query to break down said input query into query terms, said query terms comprising searchable components;

conducting a first search in said corpus of data using one of more of said searchable components to obtain documents potentially including candidate answers, wherein all documents potentially including candidate answers are stored in a data storage device;

analyzing all of said documents and each document’s metadata, in a candidate answer generation module, to generate a set of candidate answers;

conducting a second search in said corpus of data using said candidate answers and said searchable components of said query terms to obtain one or more supporting passages, wherein said supporting passages have at least one of said candidate answers and at least one of said searchable components of said query terms;

scoring said candidate answers using said supporting passages, wherein said scoring is carried out by a plurality of parallel implemented scoring modules, each scoring module producing a candidate score;

selecting one or more query answers based on said candidate score;

generating a query response based on said one or more query answers for delivery to a user,

wherein each said plurality of parallel implemented scoring modules for scoring all candidate answers using the said supporting passages automatically conducts, in parallel, one or more analyses each producing a candidate score, wherein one candidate score comprises a term match score obtained by implementing executable instructions for counting the number of terms in said supporting passage and determining if said number matches a number of terms in a candidate answer; and,

wherein a further score comprises a textual alignment score obtained by said processor implementing executable instructions for determining if placement of words in said supporting passages are in alignment with placement of words of said candidate answers; and,

wherein a further score comprises a deeper analysis score obtained by said processor implementing executable instructions for determining the meaning of the supporting passages and input queries by analyzing lexical or semantic relations.

(emphasis added to draw attention to structural, tangible components).

If the system claims in CLS Bank v. Alice Corp. are not patent eligible then I don’t see how these claims can be patent eligible either.

Conclusion

Have we really reached the point where truly astonishing innovations, innovations once thought to be impossible, are not patent eligible? Talk about jumping the shark!