[00:00:52] Speaker 03:
We will hear argument next in number 181730, in Ray Baranoff.

[00:01:00] Speaker 03:
Mr. Cahill.

[00:01:05] Speaker 01:
May it please the court.

[00:01:07] Speaker 01:
My name is Ron Cahill, and I represent the appellants and patent applicants in this case, Dr. Demeter Baranoff et al.

[00:01:14] Speaker 01:
Dr. Baranoff's invention is not directed to an abstract idea under step one of Alice.

[00:01:19] Speaker 01:
because it provides an improvement to the operation of a medical device.

[00:01:23] Speaker 01:
And if we reach step two, Dr. Baranoff's invention is patentable because it's the software that's coupled to physiological sensors that measure certain data about the patient and then goes through an estimation routine to tell us things about the patient that the sensors could never tell us.

[00:01:45] Speaker 01:
In essence,

[00:01:46] Speaker 01:
the medical device, right, and the FDA regulates this as a medical device, can now measure parameters, measure, finger quotes, parameters that it could never measure before.

[00:01:57] Speaker 01:
And I'd like to talk about what some of those are.

[00:02:02] Speaker 01:
If we reach step two, Dr. Baranov's invention is patentable because its recited steps are unconventional and not known in the industry.

[00:02:11] Speaker 01:
Now, turning to claim one,

[00:02:17] Speaker 01:
One way to understand claim one is that it relates to information at three different levels.

[00:02:22] Speaker 01:
The most basic level is data.

[00:02:25] Speaker 01:
That data is acquired from physiological sensors.

[00:02:28] Speaker 01:
Now, that data is not the internal state variables of the claim.

[00:02:32] Speaker 01:
That's the next higher level of information.

[00:02:35] Speaker 01:
Rather, it's data associated with those variables.

[00:02:38] Speaker 03:
Is that what's covered by the M, the measure?

[00:02:41] Speaker 01:
Yes, that's exactly what.

[00:02:45] Speaker 01:
An example of data that's measured might be the output of a heart rate monitor.

[00:02:50] Speaker 01:
It gives you a signal that's representative of heart rate.

[00:02:53] Speaker 01:
It's a noisy signal.

[00:02:54] Speaker 01:
It's not perfect.

[00:02:55] Speaker 01:
It's not exact, but it's data, and the system can use that.

[00:03:00] Speaker 01:
Now, the next level up is the internal state variables.

[00:03:04] Speaker 01:
Now, the internal state variables are not the same thing as the data.

[00:03:08] Speaker 01:
They're not, even if, even if... What are they?

[00:03:11] Speaker 03:
Yes.

[00:03:11] Speaker 03:
What are they, if they're not that?

[00:03:13] Speaker 01:
They are estimated parameters.

[00:03:15] Speaker 01:
And so even if an internal state variable is heart rate, it may not be the same as what you measure.

[00:03:24] Speaker 01:
Because, as your honor noticed, in the math that gets done here, each internal state variable, its estimate is based on all of the measurements and on predictions from all of the other state variables.

[00:03:42] Speaker 01:
So when you see m as a function of t sub k there, the specification explains that that's all of the measurements.

[00:03:50] Speaker 01:
It's not just one.

[00:03:52] Speaker 01:
And so the state variables are based on all of the measurements.

[00:03:56] Speaker 03:
And the ISV sub s is a particular state variable, right?

[00:04:00] Speaker 01:
It's integrated across all of them.

[00:04:02] Speaker 03:
No, no, no, just the sub s before you get to the integral.

[00:04:05] Speaker 03:
I'm just looking at the right-hand term under the integral.

[00:04:07] Speaker 01:
So the sub s is replaced by

[00:04:11] Speaker 01:
ISV-1, ISV-2, ISV-3, ISV-4.

[00:04:14] Speaker 01:
So it's all of the internal state variables.

[00:04:17] Speaker 03:
Each term, each of the sub-essences is a particular internal state variable.

[00:04:26] Speaker 01:
Yes, that's correct.

[00:04:26] Speaker 03:
So you're saying you're looking at a particular internal state variable like heart health or something.

[00:04:35] Speaker 03:
What the probability of

[00:04:37] Speaker 03:
of that value is given all of the measurements on everything on at time t sub k correct okay and then you're multiplying by is it going to say that say that stay that way till tomorrow that's the first term under the until t sub k plus one that's what i mean by tomorrow exactly so yes and the way that works in the claims is there's first a prediction

[00:05:05] Speaker 01:
we predict what all of the state variables will be tomorrow.

[00:05:10] Speaker 00:
And that prediction is based on what you've determined already?

[00:05:14] Speaker 01:
That's based on what's come before, yes.

[00:05:17] Speaker 00:
And what exactly has come before?

[00:05:19] Speaker 01:
What's come before is more data, more measurements, and prior estimations.

[00:05:25] Speaker 00:
And what are the prior estimations based on?

[00:05:28] Speaker 01:
Well, this is a recursive filter.

[00:05:31] Speaker 01:
So there's an initial estimation, and then at each time step,

[00:05:35] Speaker 01:
There's more data and more estimates.

[00:05:39] Speaker 01:
So we predict what the heart rate will be tomorrow.

[00:05:44] Speaker 01:
And then tomorrow, we check all the data and compare it to all of our predictions for the state variables.

[00:05:51] Speaker 01:
And based on that, there are posterior probability density functions that are built that tell you, that estimate for you, what the internal state variable is now.

[00:06:01] Speaker 01:
Now where this gets truly important is

[00:06:05] Speaker 01:
Internal state variables don't have to be the things that you measure.

[00:06:08] Speaker 01:
So, in fact, expressly in the claim, they're not.

[00:06:13] Speaker 01:
If you look to the specification, one of the key problems that's addressed on the way to developing the diagnoses that come at the end, the common element four of claim one, is the problem of hidden, hidden parameter estimates.

[00:06:31] Speaker 01:
So it turns out that a key parameter

[00:06:34] Speaker 01:
for this system is delivered oxygen.

[00:06:38] Speaker 01:
And you can see this example in the appendix at pages 59 to 63.

[00:06:44] Speaker 01:
It's something that you can't measure.

[00:06:46] Speaker 01:
It's something that your heart rate monitor and other physiological sensors can't tell you but claim one can because when it's an internal state variable that can be based on the data that you do measure and on other state variables.

[00:07:04] Speaker 01:
And so now, your medical device, it can measure, finger quotes again, parameters that it couldn't measure.

[00:07:13] Speaker 01:
And that's a key part of what this invention is.

[00:07:16] Speaker 01:
It's the problem that the inventor set out to solve when they set out to come up with a system that delivered objective diagnoses.

[00:07:27] Speaker 01:
And those are also the steps that the examiner found were not

[00:07:33] Speaker 01:
conventional or well-known in the industry.

[00:07:37] Speaker 01:
So this invention has an important technical effect, which we think takes it out of step one.

[00:07:44] Speaker 01:
But the very thing that gives it that technical effect, those are steps that are not conventional, right?

[00:07:55] Speaker 01:
Because the examiner said that it was the recited elements that were not conventional or well-known.

[00:08:04] Speaker 01:
And so these unconventional elements get you to this place through the first three elements of claim one.

[00:08:12] Speaker 02:
What did the examiner mean in the examiner's answer when the examiner said the only element of the current invention that may be considered a non-routine, unconventional, and not well-understood activity is applicants' specified algorithm?

[00:08:27] Speaker 02:
I'm looking at JA 898-89.

[00:08:34] Speaker 01:
So I think what he's talking about there are actually the elements of the claim.

[00:08:39] Speaker 02:
And so it sounds like he's saying everything in the claim, except for your recited algorithm, is routine, is conventional, and is well-understood activity.

[00:08:52] Speaker 02:
I mean, we can quibble over whether he's right or wrong, but that appears to be what he's saying in his examiner's answer.

[00:08:58] Speaker 01:
So the specifics of what he said on the record

[00:09:02] Speaker 01:
and I'm quoting this out of the solicitor's brief is, receiving data, processing the received data, and calculating using the formula a new set of data and subsequently identifying particular data are well understood routine and conventional activities.

[00:09:17] Speaker 01:
And we wouldn't disagree that once somebody has conceived and laid out all of these steps, that implementing on a general purpose computer can be done.

[00:09:26] Speaker 01:
A person of ordinary skill in the art can do that.

[00:09:29] Speaker 01:
But in fact,

[00:09:30] Speaker 01:
putting these steps together so that your medical device can measure, and I keep using my finger quotes, things that it can't measure.

[00:09:39] Speaker 01:
That seems to be neither abstract nor is it conventional or routine.

[00:09:43] Speaker 02:
So can we talk about cases like, in your remaining time, electric power group and SAP, which seem to be about grabbing data, collecting data, maybe from sensors, and then analyzing that data, perhaps,

[00:10:01] Speaker 02:
running it through some formulas, algorithms, rules, and then outputting some kind of indication of something.

[00:10:11] Speaker 02:
Perhaps that the health status of an electric power grid or something to do with investments.

[00:10:18] Speaker 02:
And so what's going on here that differentiates what's going on here from those claims, which were at bottom about

[00:10:28] Speaker 02:
grabbing data, analyzing data, and then outputting a new type of data based on the analysis of the collected data.

[00:10:37] Speaker 01:
Well, in electric power, the claims did in fact call for collecting, analyzing, and displaying data in the context of power grid monitoring.

[00:10:46] Speaker 01:
Those claim elements, however, were said to be purely conventional, that the things that were measured were known measurements, that the analysis that was done was known analysis,

[00:10:57] Speaker 01:
and that the output was exactly the output that you'd expect.

[00:11:03] Speaker 01:
In essence, the invention was simply taking what people already did, what was already well-known and conventional in the industry.

[00:11:11] Speaker 03:
And SAP was not that.

[00:11:13] Speaker 01:
Putting it on a computer.

[00:11:16] Speaker 03:
SAP had a little twist to it, and that is... Maybe even a big twist, like a really unconventional algorithm.

[00:11:26] Speaker 01:
In SAP, there was a twist.

[00:11:31] Speaker 01:
And in SAP, we have selecting a sample space, generating a distribution function in a certain way, and that's where the twist is, and generating a plot of the distribution function.

[00:11:44] Speaker 01:
But again, essentially, these were all well-known steps in the art of financial analysis.

[00:11:51] Speaker 01:
And what changed was the particular distribution function that

[00:11:57] Speaker 01:
the inventors had recognized that the Gaussian distribution that had traditionally been used was maybe not accurate.

[00:12:04] Speaker 01:
And they came up with a resampled distribution.

[00:12:07] Speaker 01:
They essentially swapped out one distribution for another.

[00:12:10] Speaker 01:
And that was the, as I understand it, and Druner perhaps knows much more about this than I do, that was the twist.

[00:12:21] Speaker 01:
But the court found, and I think this is critical for our purposes here,

[00:12:27] Speaker 01:
that there were no factual allegations from which one could plausibly infer that those steps were inventive.

[00:12:35] Speaker 01:
That's what the court said.

[00:12:38] Speaker 01:
Here, it's different.

[00:12:39] Speaker 03:
I mean, this part I think I do remember.

[00:12:42] Speaker 03:
It wasn't because there were no facts on which one could conclude that it wasn't innovative, brilliant, clever, and useful, but just not on the right side of the line of where that

[00:12:57] Speaker 03:
advance has to take place.

[00:13:00] Speaker 01:
Well, isn't that the question of whether it's an abstract idea or not, as opposed to step two?

[00:13:06] Speaker 03:
Well, another way of putting it is that all of the innovativeness was in the abstract category.

[00:13:16] Speaker 03:
And why is this not that?

[00:13:20] Speaker 01:
Because this is a concrete series of steps that are sequenced in time

[00:13:25] Speaker 01:
pulling data from physiological sensors connected to a patient to then reach conclusions that have never been reached before.

[00:13:35] Speaker 03:
You are into your rebuttal time.

[00:13:36] Speaker 03:
Why don't you save that time?

[00:13:39] Speaker 03:
We can continue, but I think we're going to stick to the time on this one.

[00:13:42] Speaker 01:
I'll save it for rebuttal.

[00:13:43] Speaker 01:
Thank you.

[00:13:53] Speaker 04:
Good morning, Your Honors.

[00:13:54] Speaker 04:
May it please the court, Coke Stewart for Appellee, Director of the USPTO.

[00:13:59] Speaker 04:
What I think I heard this morning was a slight shift in argument on behalf of Baranoff.

[00:14:06] Speaker 04:
I think what the real focus of the argument was in the brief to this court was whether the claims describe an improvement to a computer.

[00:14:15] Speaker 04:
They didn't really talk a lot about medical devices or other technological improvements.

[00:14:18] Speaker 04:
And I think based on the case law that's been decided by this court,

[00:14:23] Speaker 04:
This type of discussion is really not something that's improving the computer.

[00:14:27] Speaker 04:
If you look at the specification, and this kind of falls under the government's inquiry as to whether the claims are adding significantly more, the specification talks about the use of conventional hardware.

[00:14:47] Speaker 04:
And that's conventional hardware in terms of the sensors,

[00:14:50] Speaker 03:
which are described... So medical devices today, many of them, will have extremely sophisticated processing.

[00:15:01] Speaker 03:
Let's just call those computers.

[00:15:02] Speaker 03:
So the part of the medical device that's being improved is the computer.

[00:15:06] Speaker 03:
It's still a medical device and you're getting health information that you wouldn't get.

[00:15:10] Speaker 03:
So why is that not eligible?

[00:15:14] Speaker 04:
Well, following, you know, the two-step framework, first we look to see whether there's been an abstract idea.

[00:15:20] Speaker 04:
And here, I think the examiner found and the board properly affirmed that what's being discussed may be very novel.

[00:15:29] Speaker 04:
The algorithm may be novel.

[00:15:31] Speaker 04:
It may not have been used before.

[00:15:32] Speaker 04:
It may not have been used before in the health care industry.

[00:15:35] Speaker 04:
But that's very similar to other cases, including cases before the Supreme Court, like Parker v. Fluke.

[00:15:41] Speaker 04:
So we can't look to the novelty of the formula to evade the Section 101 inquiry.

[00:15:48] Speaker 04:
And when we look to step two and we see, you know, is there something more happening here?

[00:15:52] Speaker 04:
Or is it just, you know, take this idea and put it on a computer?

[00:15:56] Speaker 04:
That's where we have, you know, our conventional medical equipment.

[00:16:00] Speaker 04:
We have our conventional computer hardware, which is recited as just processor memory, network interface, et cetera.

[00:16:08] Speaker 04:
And, you know, what's really happening here, because these claims have been rejected as obvious before the algorithm was added, is the novelty of the algorithm.

[00:16:18] Speaker 04:
And unfortunately, I think our hands are just tied that the Supreme Court and this court has said over and over again that novel algorithms do not take claims outside of the scope of patent, does not make claims eligible.

[00:16:35] Speaker 04:
So I hear their argument, and it may be that this algorithm is doing something very special.

[00:16:42] Speaker 04:
But the case law that we have in front of us

[00:16:46] Speaker 04:
doesn't allow that to make those claims not eligible.

[00:16:50] Speaker 02:
What if hypothetically someone invented a medical device that measures heart rate but outputs on a display on this medical device the likelihood of some kind of arrhythmia condition?

[00:17:10] Speaker 02:
with that device, and then there's a lot of processing power going on inside this device using a series of mathematical operations to convert the data collected on the heart rate to produce some kind of diagnosis about arrhythmia.

[00:17:33] Speaker 02:
Would that kind of medical device be patent eligible?

[00:17:38] Speaker 04:
I think that would still be in a bit of a gray area, Your Honor.

[00:17:43] Speaker 04:
The benefit that your described claim has that these claims don't have is that level of specificity.

[00:17:49] Speaker 04:
I mean, for example, in Vonda, you're treating a very specific problem.

[00:17:53] Speaker 04:
I mean, the likelihood that someone's going to have a heart attack.

[00:17:56] Speaker 04:
You're measuring a specific kind of data in a specific way, someone's heart rate or the oxygen in their blood.

[00:18:04] Speaker 04:
The more specificity added to the claims

[00:18:06] Speaker 04:
Increases the likelihood that they're going to be found eligible and that's one of the major problems that we have with these claims is their breath How do these claims rise or fall in light of the agency's new examination guidelines on section 101?

[00:18:26] Speaker 04:
we think the the outcome would have been the same under the new guidelines because in fact

[00:18:32] Speaker 04:
It happens that while the new guidelines start with three broad categories of abstract ideas, at least two of those categories were discussed in the examiner's office actions and answered in the board's decision, which are mathematics and methods of organizing activities.

[00:18:50] Speaker 04:
So we think the framework applied might have been slightly different.

[00:18:54] Speaker 04:
We would have started with broader categories, and then we would have moved down from there, but the outcome would have been the same.

[00:18:59] Speaker 03:
The methods of organizing activity, what does that mean?

[00:19:03] Speaker 04:
Well, to answer Judge Chen's question, the guidelines take abstract ideas and put them into three broad categories.

[00:19:11] Speaker 04:
And the methods of organizing activities here would be collecting information, processing information, what doctors do with that information.

[00:19:21] Speaker 04:
It would also probably be, in some ways, a mental process as well.

[00:19:28] Speaker 02:
So does Electric Power Group and SAP individually and combined control the outcome of this case?

[00:19:39] Speaker 04:
Well, I would say they're certainly persuasive and helpful to the office.

[00:19:46] Speaker 04:
We focused on Parker v. Fluke because we felt like that was the clearest guidance from the Supreme Court on

[00:19:54] Speaker 04:
you know, the use of mathematical algorithms and trying to make predictions.

[00:19:59] Speaker 04:
But we definitely agree that, you know, to the extent that electric power and SAP talk about gathering data, processing data, you know, making predictions based on that data that's certainly analogous to this case and would support the office's decision.

[00:20:19] Speaker 04:
If there are no other questions, we rest on the other arguments in our brief.

[00:20:22] Speaker 04:
Thank you.

[00:20:31] Speaker 01:
First, I think breadth is not a 101 issue.

[00:20:34] Speaker 01:
One can have broad or narrow claims that are abstract or not abstract.

[00:20:38] Speaker 03:
In a kind of separate line of cases, we've talked about breadth being a 101 issue, some of them summarized in SAP.

[00:20:53] Speaker 03:
The more general you get, the more close to character

[00:20:58] Speaker 03:
you know, that is actually has become one of the ways we have identified something as abstract.

[00:21:05] Speaker 01:
Well, I think we would agree that the more abstract something becomes, the more abstract it is.

[00:21:09] Speaker 03:
Well, the word abstract has a number of different meanings.

[00:21:11] Speaker 03:
Generality happens to be one of them and intangibility another.

[00:21:16] Speaker 01:
If I could, I'd like to mention, I'd like to go back to the cases a little bit.

[00:21:25] Speaker 01:
You know, we talked a little bit about electric power,

[00:21:27] Speaker 01:
Critically, the court said there that there were one of the ways that the court distinguished the claims there as being not subject matter eligible is the court said that there were no new techniques for analyzing the data, that there were no new algorithms claimed.

[00:21:42] Speaker 01:
I think that's important.

[00:21:44] Speaker 02:
And... Are you saying if the electric power group claim recited a mathematical formula, that that would have passed muster under Section 101?

[00:21:54] Speaker 01:
If it had done a new analysis

[00:21:57] Speaker 01:
that had gone beyond the abstract idea that had, that was not conventional and well-known, then yes, it would then pass Alice Step 2.

[00:22:11] Speaker 02:
When I think about the Bilsky case, the independent claim was just about hedging risk.

[00:22:19] Speaker 02:
You have two sets of contract transactions at two separate prices.

[00:22:24] Speaker 02:
And then there was another claim that recited a formula for calculating what the price would be to be set for making the hedge.

[00:22:37] Speaker 02:
The Supreme Court killed both those claims.

[00:22:42] Speaker 02:
Are you saying that if that second claim recited a formula that was really novel, extra novel,

[00:22:53] Speaker 02:
very innovative, that would have been enough?

[00:22:56] Speaker 01:
If it went beyond the abstract idea and was not conventional or well known, then yes.

[00:23:01] Speaker 02:
So math, innovative math can be a basis for patentability?

[00:23:08] Speaker 01:
It can certainly help.

[00:23:10] Speaker 01:
We're talking about a method, and we have steps in our method, and if those steps go beyond the abstract idea, we're in step two, and those steps are

[00:23:23] Speaker 01:
not well-known or conventional or routine, then yes.

[00:23:27] Speaker 01:
And I think to get there, you have to look at what the abstract idea is.

[00:23:31] Speaker 01:
And here, right, the abstract idea from the board was predicting patient health risks and diagnoses at a broader level to a mathematical formula or relationship.

[00:23:43] Speaker 01:
I think claim one goes well beyond that and addresses this problem of estimating hidden parameters, which is a stepping stone to be able to

[00:23:53] Speaker 01:
objectively diagnose a patient.

[00:23:56] Speaker 01:
I think that goes beyond the abstract idea that's been stated and the steps themselves are not conventional or routine.