[00:00:04] Speaker 03:
Our next case is number 182382, Apple Inc.

[00:00:09] Speaker 03:
versus Andrea Electronics Corporation.

[00:00:55] Speaker 04:
Mr. Cushman.

[00:00:57] Speaker 04:
May it please the court?

[00:00:58] Speaker 04:
Jeff Cushman for Appellant.

[00:01:02] Speaker 04:
This appeal concerns two decisions of the PTAB regarding the same patent.

[00:01:07] Speaker 04:
In those two proceedings, three or four claims are the only issues left to address.

[00:01:13] Speaker 04:
All of the other claims that were contested were found unpatentable and there haven't been no review of those decisions.

[00:01:19] Speaker 00:
Why did Apple break it into two separate

[00:01:22] Speaker 04:
So there were two different sets of prior art that showed why these claims were unpatentable and they were implicating different types of aspects of the processing systems and we just presented them in two different approaches.

[00:01:38] Speaker 01:
Is that a common approach in IPRs?

[00:01:40] Speaker 04:
It was.

[00:01:41] Speaker 04:
I think the board has more recently suggested that they should try to force everything into one petition for efficiency, but at the time these petitions were filed, I don't think that was an unusual practice.

[00:01:51] Speaker 04:
Thank you.

[00:01:53] Speaker 04:
So I'd like to first start with the 626 proceeding.

[00:01:56] Speaker 04:
And in that case, the board found that applet raised a new theory of unpatentability in its reply.

[00:02:02] Speaker 04:
And then it refused to consider part of the record that was before it.

[00:02:06] Speaker 04:
So the first question, whether the patentability question was different from what had been presented in the petition, is a question you reviewed de novo.

[00:02:16] Speaker 04:
The second issue is whether that decision to not consider it full record was an abuse of discretion.

[00:02:21] Speaker 04:
That was a link to its incorrect first decision.

[00:02:25] Speaker 04:
Apple didn't raise a new patentability theory in its reply.

[00:02:28] Speaker 04:
Its obviousness theory in the reply was the same one it presented in its petition.

[00:02:34] Speaker 04:
that it would have been obvious to run the noise floor algorithm that was in the Martin reference on the Hirsch apparatus.

[00:02:40] Speaker 01:
Why is setting W to 1 just an example and not the only condition under which the prior rendered the claims anticipated or obvious?

[00:02:53] Speaker 04:
So the ground we advanced, two answers to that.

[00:02:56] Speaker 04:
First, it's very clear in the record.

[00:02:58] Speaker 04:
We did not portray that as the obviousness ground.

[00:03:01] Speaker 04:
We actually said in the petition, this is an example.

[00:03:04] Speaker 04:
And I can give you the quote.

[00:03:06] Speaker 04:
At appendix 131 to 132, you see the challenge to claim four.

[00:03:10] Speaker 04:
What we say right there is a skilled person would have found it obvious to modify Hirsch's adaptive threshold calculation to use Martin's noise floor algorithm.

[00:03:19] Speaker 04:
And then when it got to the example that was used to illustrate how that algorithm worked, that example was from petitioner expert Dr. Hochwald.

[00:03:28] Speaker 04:
That was at Appendix 130.

[00:03:30] Speaker 04:
And Apple said, Dr. Hochwald illustrates Martin's algorithm with the following example.

[00:03:36] Speaker 04:
And from the record, it's clear that was an illustration.

[00:03:40] Speaker 04:
It was an example from our expert to show how the variables within that algorithm would change as it processes a portion of that sound signal.

[00:03:51] Speaker 04:
The ground, as I said, was run Martin's algorithm on Hirsch's apparatus.

[00:03:57] Speaker 01:
What page would you direct us to for knowing what the ground is?

[00:04:00] Speaker 04:
So I'd direct you to page appendix 131 to 132.

[00:04:03] Speaker 04:
This is the detailed explanation in the petition why claim four.

[00:04:08] Speaker 04:
was obvious.

[00:04:10] Speaker 04:
And then as you march through the rest of the preceding the section dealing with the combination of Hirsh and Martin, that's appendix 132 to 136, what you see is that Apple mapped the various requirements of those claims to the aspects of the algorithm that it presented.

[00:04:28] Speaker 04:
The algorithm as a whole is presented in the petition.

[00:04:32] Speaker 04:
It's not cribbed or otherwise constrained.

[00:04:35] Speaker 04:
There's no reference in the analysis of the obvious reasons in that section from 131 to 136.

[00:04:42] Speaker 01:
I did read your petition, and the only place where I saw the discussion of W equal 1 is on page A 130.

[00:04:50] Speaker 01:
Are there any other places that I missed?

[00:04:53] Speaker 04:
No.

[00:04:55] Speaker 04:
So the reason that's important, the way we framed our challenge, is that under this court's precedent in Parker Vision versus Qualcomm, when you're evaluating the patentability of a claim to an apparatus, you look to see whether the prior art apparatus has the capability of performing what the claims require.

[00:05:12] Speaker 04:
Here, our obviousness challenge was if you run Martin's algorithm on the apparatus that Hirsch had developed.

[00:05:20] Speaker 04:
then that apparatus doing the Martin algorithm will meet every requirement of claims 6 to 9.

[00:05:27] Speaker 04:
That was the ground we had in our petition.

[00:05:30] Speaker 04:
That was the ground we had in our reply.

[00:05:32] Speaker 04:
And that's why it was improper for the board to find that we had raised a new patentability theory in our reply.

[00:05:38] Speaker 04:
And it was also an abuse of discretion for them to then not consider the full record.

[00:05:42] Speaker 04:
And that abuse of discretion, I think, compels a remand.

[00:05:45] Speaker 03:
And you're not, as I understand it now, arguing that

[00:05:48] Speaker 03:
this particular illustration where W equals 1 is anticipatory?

[00:05:54] Speaker 04:
No.

[00:05:55] Speaker 04:
No.

[00:05:56] Speaker 04:
It's an obviousness ground, and the obviousness ground is based on running the algorithm as a whole in the apparatus.

[00:06:03] Speaker 03:
With multiple sub windows?

[00:06:05] Speaker 04:
Yes.

[00:06:06] Speaker 04:
In any configuration.

[00:06:07] Speaker 04:
It's a capability question.

[00:06:08] Speaker 04:
The algorithm where it runs, as it's depicted in the Martin algorithm, doesn't set the W number.

[00:06:15] Speaker 04:
It's an algorithm that you can set as you wish.

[00:06:17] Speaker 01:
I understood your reply to be saying that you have a twofold response.

[00:06:25] Speaker 01:
One was that you can set W to one, and it will still render the claims, I guess, obvious.

[00:06:31] Speaker 01:
And then alternatively, that you didn't intend to limit W one anyway.

[00:06:36] Speaker 01:
look at the example where w equals 4.

[00:06:39] Speaker 04:
So I think part of the confusion is the way that it was engaged by the patent owner in response.

[00:06:45] Speaker 04:
They attacked the illustration from Dr. Hochwald as being an unrealistic illustration of how Martin might work.

[00:06:53] Speaker 04:
The reason Dr. Hochwald used that illustration was just to help the board see what the variables, how they change.

[00:07:00] Speaker 04:
There are three variables.

[00:07:01] Speaker 04:
It's a future minimum, a current minimum, and there's a value of the sound.

[00:07:05] Speaker 04:
But the illustration was designed to show how the values of those variables change as it's processing a sample of sound.

[00:07:12] Speaker 04:
These samples usually involve thousands and thousands of data points.

[00:07:16] Speaker 04:
So he made an overly simplistic one that only had like 30 data points, just so you could see what the values are of those variables as they move through the data set.

[00:07:28] Speaker 04:
That challenge, or the attack of that as w equal 1 not being reasonable, was really aimed at his illustration.

[00:07:34] Speaker 04:
It wasn't about the challenge we presented for obviousness.

[00:07:37] Speaker 04:
And that's why some of the confusion has arisen.

[00:07:39] Speaker 04:
What you see in our reply is our explanation.

[00:07:42] Speaker 03:
They argued also, did they not, about the word w equals 4?

[00:07:47] Speaker 04:
Yes.

[00:07:47] Speaker 04:
And we actually, if you look at Dr. Hochwald's declaration, he goes through and he observes in Martin that w equal 4 example.

[00:07:56] Speaker 04:
But again, he's not using it to show obviousness of the system.

[00:08:00] Speaker 04:
He's using it to illustrate what those variables are doing as it's processing the sound sample.

[00:08:05] Speaker 04:
And as a consequence, when you see the challenge in the reply, what we did in the reply was respond to the arguments and criticisms that Andrea had lodged against our explanation, and including of the illustration.

[00:08:17] Speaker 04:
The reason that's not consequential is

[00:08:20] Speaker 04:
What we were showing in that illustration is that the requirements of the claims are met when you process data.

[00:08:26] Speaker 04:
And one thing to note, you'll see in the picture that's in the petition that came from Dr. Hockwell that it's a graph.

[00:08:32] Speaker 04:
There are two kinds of sound.

[00:08:34] Speaker 04:
One kind of sound has the magnitude consistently increases over the sample.

[00:08:39] Speaker 04:
That's called monotonically increasing sound in Martin.

[00:08:43] Speaker 04:
The other kind of sound varies.

[00:08:44] Speaker 04:
It goes up and down.

[00:08:45] Speaker 04:
That's not monotonically increasing.

[00:08:48] Speaker 04:
If you look at our petition, you'll see there's a block which says if it's not monotonically increasing, you go left on the left side of that equation.

[00:08:57] Speaker 04:
And that is the illustration that's being shown in the figure in Dr. Harbaugh's illustration.

[00:09:03] Speaker 04:
And it shows clearly that in that setting,

[00:09:06] Speaker 04:
the variable for the future minimum and the current minimum gets set to the lowest value no matter how many sub windows are used.

[00:09:13] Speaker 04:
That's just math.

[00:09:14] Speaker 04:
It just says find the lowest number within that block of samples.

[00:09:19] Speaker 04:
And that's what the reply explained, why it was an inconsequential issue for the ultimate question for that kind of sound.

[00:09:26] Speaker 04:
So I think in this case, the board just erred in finding that we raised a new patentability theory.

[00:09:32] Speaker 04:
We used the same mappings in the claims of the prior art to the claims in the petition as we did in the reply.

[00:09:41] Speaker 04:
We used the same part of the prior reference in both places.

[00:09:46] Speaker 04:
There's no difference in our patentability ground.

[00:09:48] Speaker 04:
And that was the error that you reviewed in NOVA that we believe should be reversed.

[00:09:53] Speaker 04:
How important is the issue periodically to you?

[00:09:59] Speaker 04:
For the 626 proceeding, it's not implicated because the Martin system does that at a regular interval.

[00:10:06] Speaker 04:
The periodicity issue is one issue that we believe can justify a reversal on the 627 proceeding.

[00:10:13] Speaker 04:
And that's because it's a claim destruction issue.

[00:10:16] Speaker 04:
And if you look at in that 626.

[00:10:19] Speaker 04:
Does it matter?

[00:10:21] Speaker 04:
It does matter.

[00:10:22] Speaker 04:
It doesn't matter for the 626 because it wasn't disputed in the 627 proceeding.

[00:10:26] Speaker 04:
So in the 627 proceeding, the reason why it doesn't matter is that under our construction, there's no dispute that the method of health will meet the requirements of the claims.

[00:10:38] Speaker 04:
They have not disputed under the broader construction that the claims are unpackable.

[00:10:43] Speaker 04:
So in the 627 proceeding,

[00:10:46] Speaker 04:
What I'd like to do is just talk through an illustration which shows what the meaning of periodically is as it's explained by the patent.

[00:10:56] Speaker 04:
And if you look in claim nine of the patent, the term that claim is referring to a future minimum.

[00:11:04] Speaker 04:
It says that the future minimum value is set to a current magnitude value periodically.

[00:11:09] Speaker 04:
That's claim nine's use of periodically.

[00:11:12] Speaker 04:
The future minimum is the value in this system where it's the lowest magnitude of the sample of the sound that's being analyzed.

[00:11:22] Speaker 04:
The current magnitude is the value of each data point in that sample.

[00:11:27] Speaker 04:
This process is illustrated in Figure 7 of the patent, which is at Appendix 68.

[00:11:32] Speaker 04:
And what Figure 7 shows

[00:11:35] Speaker 04:
Two things.

[00:11:36] Speaker 04:
First, it shows that if the value of the future minimum is greater than the current magnitude of the sample being evaluated, that's what's designated as Y current N. I'm sorry, but I'm looking at Figure 7, just to make sure I'm catching up with you.

[00:11:51] Speaker 04:
You'll see in the bottom middle left side, you'll see a line of logic going down where it's referring to the future minimum.

[00:11:58] Speaker 01:
OK.

[00:11:59] Speaker 04:
And you see in items 710 and 716, the processing.

[00:12:04] Speaker 04:
Those are events that are occurring not at a regular interval.

[00:12:07] Speaker 04:
Those are events that are being driven for each data sample being evaluated.

[00:12:11] Speaker 04:
So right there, it's showing that value, future minimums being set not at a regular interval, but from time to time, whenever the data condition is met.

[00:12:23] Speaker 04:
Then at the bottom of that figure, it takes a dog leg over, I guess, to the left.

[00:12:29] Speaker 04:
And the future minimum gets set at the five second interval again.

[00:12:35] Speaker 04:
That's what's referred to as the initiating of the value.

[00:12:38] Speaker 04:
So right within this figure seven are two ways of the value of the future minimum being set to a current magnitude as claimed nine is envisioning.

[00:12:49] Speaker 04:
One from time to time and the other one at a regular interval.

[00:12:53] Speaker 04:
And so that's why in our construction for periodically, we said that's the ordinary meaning of periodically, and it matches exactly what the claims specify.

[00:13:03] Speaker 03:
But what about the column eight, line 36, where it says the future and current minimum values are calculated continuously and initiated periodically, for example, every five seconds, as determined in step 724?

[00:13:19] Speaker 03:
That seems to contemplate that periodically means at regular intervals.

[00:13:25] Speaker 04:
Well, I think, in all fairness, it would contemplate as doing it both.

[00:13:29] Speaker 04:
So the continuous calculation, the continuous setting that's being illustrated in Figure 7 is showing that the values are going to change whenever the data conditions are true.

[00:13:43] Speaker 04:
Is the value lower than the one that's been stored?

[00:13:47] Speaker 04:
That's a continuous operation as it's laid out in Figure 7.

[00:13:52] Speaker 04:
And then the initiation is the thing that occurs at the end of that block of time, for example, five seconds.

[00:13:59] Speaker 04:
But I think it's also fair to look at the patent as a whole, because when you're engaging this question of what periodically means, you have to evaluate how it's periodically used in the patent specification as a whole and what construction accommodates all the different ways it can be understood.

[00:14:16] Speaker 04:
And that's where, when you look at the illustration in Figure 7,

[00:14:20] Speaker 04:
and all the other ways these values are shown being updated, it compels the finding that it's the broader meaning that it can occur from time to time or at a regular interval.

[00:14:32] Speaker 04:
Now again, as I said, this issue, if you find that it is, as we've described, it compels a reversal because there's no dispute about the patentability question under that broader construction.

[00:14:44] Speaker 04:
Just briefly, the last issue on a flag is also related to the 627 proceeding.

[00:14:50] Speaker 04:
And here, this is an issue that we, if you don't find for us on periodically, we do believe that this would also justify a remand on the 627 proceeding.

[00:15:02] Speaker 04:
What we said in our petition was that a skilled person would have found it obvious to adapt a part of the health system.

[00:15:08] Speaker 04:
This is a different reference.

[00:15:10] Speaker 04:
And that aspect is called the running estimator.

[00:15:13] Speaker 04:
And what we said to the petition was it would have been obvious to adjust that running estimator to set the noise floor at regular intervals based on the guidance in Martin.

[00:15:23] Speaker 04:
We didn't say that you should change when you should use that type of estimator.

[00:15:28] Speaker 04:
We said when it's running as the running estimator, you just adjust how that running estimator works.

[00:15:35] Speaker 04:
In this response, Andrea came back in the proceeding and argued that,

[00:15:42] Speaker 04:
Essentially another straw man argument, it said that it's Apple's position that we should call for changing to the preconditions that would dictate whether you use the running estimator or the stationary estimator.

[00:15:53] Speaker 04:
And that was not Apple's position.

[00:15:54] Speaker 04:
But then when you look in the board decision on the 627 proceeding, it actually embraces that argument from Andrea.

[00:16:04] Speaker 01:
Going back to periodically for a minute,

[00:16:07] Speaker 01:
Did you make that figure seven argument, walk through figure seven like that in your brief?

[00:16:14] Speaker 01:
I just don't recall seeing it.

[00:16:15] Speaker 04:
So in our brief, we did address the, I believe we addressed this figure seven issue.

[00:16:20] Speaker 04:
And I would have to double check that.

[00:16:23] Speaker 04:
Maybe I can respond to that in our rebuttal.

[00:16:27] Speaker 03:
Okay.

[00:16:28] Speaker 03:
We'll give you two minutes.

[00:16:29] Speaker 04:
Thank you very much, Your Honor.

[00:16:34] Speaker 03:
Mr. Bellinger.

[00:16:41] Speaker 02:
May it please the court, William Bellinger on behalf of Andrea.

[00:16:44] Speaker 02:
And I want to start my argument in reverse order, starting with the periodic argument that you just heard.

[00:16:51] Speaker 02:
And I'd like to point out what I think is a key deficiency in what counsel just directed you to with respect to Figure 7.

[00:16:58] Speaker 02:
So he's correct that in Claim 9, the patentees chose the word periodic to describe circumstances where the current magnitude value is

[00:17:08] Speaker 02:
Say excuse me the future minimum value is set to the current magnitude value periodically if you look just earlier to claim eight It says where the current minimum value is replaced with a current magnitude value Where so current minimum value is greater than said current magnitude value, and that's I guess consistent with the discussion in column eight lines

[00:17:32] Speaker 01:
38 through 40 where it's said to be different, calculating continuously versus initiating periodically.

[00:17:39] Speaker 02:
Exactly.

[00:17:39] Speaker 02:
So one example the patentee gives is if one value is greater than another, you replace them.

[00:17:45] Speaker 02:
The other example they give is you replace them at a fixed period of time.

[00:17:50] Speaker 02:
Both experts agree in the context of this patent periodically means a fixed interval of time.

[00:17:56] Speaker 02:
So there's no dispute that the construction that the board adopted was one that was supported by both Apple's expert and Andre's expert.

[00:18:05] Speaker 02:
But it does say, for example.

[00:18:08] Speaker 02:
The intrinsic record says, for example, five seconds.

[00:18:11] Speaker 02:
As the board found correctly, there was a dispute whether the example was five seconds or six seconds or 10 seconds or whether the example was a fixed period of time or from time to time.

[00:18:24] Speaker 02:
There's no evidence in the record that the for example meant from time to time.

[00:18:29] Speaker 02:
The example counsel gave is an alternate

[00:18:32] Speaker 02:
method, which is not a periodic update.

[00:18:35] Speaker 02:
The example doesn't preclude from time to time, does it?

[00:18:41] Speaker 02:
Respectfully, I would say the example, as understood by a person with an ordinary skill in the art, would preclude from time to time.

[00:18:47] Speaker 02:
Because they gave one example of an update from time to time

[00:18:51] Speaker 02:
They describe that as doing the update greater than.

[00:18:55] Speaker 02:
And then they give another example.

[00:18:57] Speaker 02:
The only example that they call the periodic update was every five seconds.

[00:19:01] Speaker 02:
The exemplary portion of that was the time.

[00:19:04] Speaker 02:
Obviously, doing it every five seconds wasn't limiting.

[00:19:06] Speaker 02:
It could be six seconds or 10 seconds.

[00:19:09] Speaker 03:
What about in Figure 7?

[00:19:11] Speaker 02:
Yes, Your Honor.

[00:19:12] Speaker 02:
So in Figure 7, as I noted, there are two types of updates.

[00:19:18] Speaker 02:
If you look at box.

[00:19:24] Speaker 02:
712 or 710, those are doing a comparison.

[00:19:28] Speaker 02:
So y sub n, that is the measured magnitude value, or the current magnitude value of the claims.

[00:19:35] Speaker 02:
And it's determining whether the future minimum is greater than that.

[00:19:39] Speaker 02:
So that's one type of update.

[00:19:40] Speaker 02:
You're doing a comparison to see whether one value is greater than another.

[00:19:46] Speaker 02:
And that's either yes or no.

[00:19:48] Speaker 02:
If you go down to 724, that's described by the patentee as being a periodic example.

[00:19:53] Speaker 02:
He points to 710 and 714.

[00:19:55] Speaker 02:
Yes, Your Honor.

[00:19:58] Speaker 02:
So if you follow the logic in 710, that's done consistent with claims 7 and 8, when one value is greater than another value.

[00:20:11] Speaker 02:
The only update that's done from time to time, excuse me, at a fixed interval of time, is 724 did five seconds lapse.

[00:20:19] Speaker 02:
So there's one example that's a periodic update every five seconds.

[00:20:23] Speaker 02:
That's what the patentee described as periodic.

[00:20:26] Speaker 02:
There's another example where it's done based on a comparison of whether one value is greater than another.

[00:20:32] Speaker 02:
And I would direct you to the board's decision.

[00:20:37] Speaker 02:
This is an appendix.

[00:20:38] Speaker 02:
32 and 33.

[00:20:39] Speaker 02:
They started as is probably.

[00:20:41] Speaker 02:
I'm sorry, what was that saying?

[00:20:43] Speaker 02:
Appendix 32 and 33.

[00:20:44] Speaker 02:
Thank you.

[00:20:45] Speaker 02:
This is where the board evaluated the claim construction issue.

[00:20:48] Speaker 02:
They probably started with the intrinsic record.

[00:20:51] Speaker 02:
And then they looked to the extrinsic record, including the dictionary definition that Apple places primary reliance on, as well as the testimony of both experts.

[00:21:02] Speaker 02:
And as is well established in this court's precedent,

[00:21:07] Speaker 02:
General dictionary definitions are not binding.

[00:21:11] Speaker 02:
The question is, what would one of ordinary skill and the art understand?

[00:21:15] Speaker 02:
Both Apple's expert and Andre's expert agreed that in this context, signal processing, that general dictionary definition did not apply.

[00:21:22] Speaker 02:
The proper definition was a fixed interval of time.

[00:21:26] Speaker 01:
Are you going to turn to the other issue, the question of whether it's rebuttal or a new argument?

[00:21:30] Speaker 01:
Yes, Your Honor.

[00:21:31] Speaker 01:
And just what I wanted to do.

[00:21:33] Speaker 01:
I wanted to ask you one question about that, which is just that did you argue that this was a new argument?

[00:21:40] Speaker 01:
Before the board came to that conclusion?

[00:21:43] Speaker 02:
The board actually came to that conclusion on their own at the oral hearing.

[00:21:47] Speaker 02:
And then it was discussed at the oral hearing.

[00:21:50] Speaker 02:
And I apologize, but just so I don't lose this,

[00:21:52] Speaker 02:
Council suggested there is some admission that the prior art anticipates if you reverse claim construction.

[00:21:58] Speaker 02:
That's actually not correct.

[00:22:00] Speaker 02:
The board evaluated anticipation, found a difference between the reference based on their construction, and so didn't reach Andre's other distinctions between the reference and claim six, seven, and nine.

[00:22:11] Speaker 02:
So that's just a subtlety.

[00:22:12] Speaker 02:
I think the board's construction was correct, and if you affirm the construction,

[00:22:16] Speaker 02:
you would affirm the finding, no anticipation.

[00:22:19] Speaker 02:
But I think it's incorrect.

[00:22:20] Speaker 02:
Council suggested that there's an admission of anticipation, even if you reverse.

[00:22:26] Speaker 03:
And you're having a response on that.

[00:22:32] Speaker 03:
argue the obviousness question where W equals four, right?

[00:22:38] Speaker 02:
Yes.

[00:22:39] Speaker 02:
And just to point out why I think the board is correct, twofold.

[00:22:43] Speaker 02:
One, counsel suggested that this is a de novo review as to whether it's a new argument, respectfully disagree.

[00:22:49] Speaker 02:
All of this court's precedent suggests that you review that for an abusive discretion.

[00:22:53] Speaker 03:
Why isn't it appropriate for them

[00:22:56] Speaker 03:
file a reply responding to your argument that you made in the patent on a response.

[00:23:01] Speaker 02:
It's certainly permitted by the board's rules to file a reply responding to our argument, so long as they're advancing the same theory of patentability that was advanced in the petition.

[00:23:12] Speaker 02:
And respectfully, the board evaluated it.

[00:23:14] Speaker 03:
Well, why did you address it in the response if you didn't think that the petition raised this issue?

[00:23:20] Speaker 02:
We addressed in our response the argument made in the petition.

[00:23:25] Speaker 02:
to clarify, if you go to, again, looking at the board's decision on this issue, this is appendix pages 13 to 15, the exchange was as follows.

[00:23:39] Speaker 02:
The Apple and the petition chose to modify or, as Council mentioned, simplify the algorithm and the prior art to try to make it fit into an obvious analysis.

[00:23:51] Speaker 02:
As the board properly found, the entire purpose of the

[00:23:55] Speaker 02:
the Martin reference was to determine whether a signal was increasing quickly or not.

[00:24:00] Speaker 02:
And as the board found and as Apple's expert admitted, you can't figure out whether a signal is increasing rapidly or monotonically as the description is without having at least two samples.

[00:24:14] Speaker 02:
So as the board found and as our expert critiqued their obviousness analysis, it was an improper modification and combination of the two references to assume that

[00:24:24] Speaker 02:
You would start with a reference that was focused on determining whether a signal was increasing rapidly.

[00:24:30] Speaker 02:
Assume that you would take only a single sample, therefore you would be, that's the W equals one.

[00:24:35] Speaker 02:
So you'd be unable to tell whether the signal was increasing rapidly or not.

[00:24:39] Speaker 02:
And then import that algorithm into the base reference, the health reference.

[00:24:44] Speaker 03:
Why is that important?

[00:24:45] Speaker 03:
Okay, but so what you're saying is that it's a bad example, but that it is listed in the petition as illustrative, as an example, right?

[00:24:54] Speaker 02:
And respectfully, no, it's listed in the petition as the requisite element by element analysis comparing the references to the claims.

[00:25:03] Speaker 02:
So in their petition, they chose an element by element analysis

[00:25:07] Speaker 02:
which depended on this proposed modification of one reference substitution into the other reference.

[00:25:14] Speaker 02:
We argued, and the board agreed with this argument.

[00:25:16] Speaker 01:
They disagree with you on that, right?

[00:25:18] Speaker 01:
They disagree with you on that.

[00:25:20] Speaker 01:
They say that at page A 130, Dr. Hochwald illustrates Martin's algorithm with the following example.

[00:25:28] Speaker 01:
And then they take the position that it was not limited to w equal 1 later in

[00:25:33] Speaker 01:
what would probably be termed more of a claim chart analysis, when they go claim by claim and explain why the prior discloses each of the claim limitations, right?

[00:25:43] Speaker 02:
And I think the board found and they correctly found that that claim by claim analysis did only assume that W equal one.

[00:25:49] Speaker 02:
So that's in the board's decision in appendix 14.

[00:25:51] Speaker 01:
Let me ask you something else.

[00:25:52] Speaker 01:
I mean, we have case law.

[00:25:54] Speaker 01:
Do you agree with our case law that says an IRPR may introduce new evidence after the petition stage if the evidence is a legitimate reply?

[00:26:03] Speaker 01:
to evidence introduced by the patent owner.

[00:26:05] Speaker 01:
You agree with that, right?

[00:26:07] Speaker 01:
Yes.

[00:26:07] Speaker 01:
Your position is that you don't think it's a legitimate reply.

[00:26:10] Speaker 02:
Correct.

[00:26:10] Speaker 02:
Because we believe the board properly found it was advancing a new and different theory of unpatability, just as this court has also found in the intelligent biosystems case, in the Erosa case, and recently in the Frymaster case.

[00:26:23] Speaker 01:
It's new because you think when they set W equal to 1,

[00:26:27] Speaker 01:
that they were saying it's only in this circumstance that this reference renders these claims obvious?

[00:26:33] Speaker 02:
Correct.

[00:26:33] Speaker 02:
That was the only detailed analysis they gave.

[00:26:36] Speaker 02:
They gave a detailed element-by-element analysis where they assumed that W equals 1.

[00:26:41] Speaker 01:
And that's how they committed it.

[00:26:42] Speaker 01:
But they do say it's for the purpose of simplicity, don't they?

[00:26:45] Speaker 01:
Maybe it's in the expert's declaration.

[00:26:48] Speaker 01:
Correct.

[00:26:49] Speaker 01:
And in the expert's declaration.

[00:26:49] Speaker 01:
I'm just trying to understand.

[00:26:50] Speaker 01:
When somebody uses an example, when will they be stuck to it?

[00:26:53] Speaker 01:
When they say it's illustrative.

[00:26:55] Speaker 01:
When that's the only example that they compare element by element.

[00:26:58] Speaker 01:
So in other words, when somebody uses an example to compare element by element, they better sure be careful and make sure they include more examples.

[00:27:05] Speaker 02:
If they want to rely on a different example, a different portion of reference as their element by element analysis.

[00:27:10] Speaker 02:
Why is it a different portion of the reference?

[00:27:13] Speaker 02:
Because instead of assuming that you effectively remove that portion of the algorithm, you resolve it.

[00:27:20] Speaker 01:
But it's not a different column and line number of the reference, right?

[00:27:23] Speaker 02:
It's a completely different reading of the reference.

[00:27:26] Speaker 01:
In your view, it would be a different embodiment, depending on how you set w. Yes.

[00:27:31] Speaker 02:
And why that's important, just simply, Your Honor, is if you assume a way, if w equals 1,

[00:27:38] Speaker 02:
the entire window, there's only one sub-window.

[00:27:41] Speaker 02:
And so you can't tell whether it's increasing or not.

[00:27:42] Speaker 03:
If this was so clear, how come you didn't raise it?

[00:27:46] Speaker 03:
Excuse me.

[00:27:46] Speaker 03:
If this was so clear that this was a new argument, how come you didn't raise it?

[00:27:53] Speaker 02:
So and I think you're reviewing the correctness of the board's decision.

[00:27:58] Speaker 02:
Speak up a little, please.

[00:27:59] Speaker 02:
Sure.

[00:27:59] Speaker 03:
We're reviewing the correctness of the board's decision.

[00:28:01] Speaker 03:
No, no.

[00:28:02] Speaker 03:
But that doesn't answer my question.

[00:28:03] Speaker 03:
If it's so clear that this was a new argument in the reply, why didn't you argue that and only make the argument after the board raised it?

[00:28:12] Speaker 02:
So that's just the basic.

[00:28:16] Speaker 02:
So the process is we put in a patent owner's response.

[00:28:18] Speaker 02:
Our next opportunity to address the issue was at the oral hearing.

[00:28:22] Speaker 02:
We went second.

[00:28:23] Speaker 03:
We could have moved to strike the reply.

[00:28:24] Speaker 03:
There are plenty of opportunities.

[00:28:26] Speaker 03:
You had to say that it was a new argument.

[00:28:28] Speaker 03:
You didn't say it until the board said it, right?

[00:28:31] Speaker 02:
And respectfully, the flow of the argument was petitioner went first.

[00:28:36] Speaker 02:
And so petitioner was making his argument.

[00:28:38] Speaker 02:
The administrative judge,

[00:28:40] Speaker 02:
address this.

[00:28:41] Speaker 02:
We had proposed demonstratives, which made this argument, but there's not a separate briefing schedule.

[00:28:49] Speaker 02:
We got the patent owner's response.

[00:28:50] Speaker 02:
They put in the reply.

[00:28:51] Speaker 02:
We went to oral hearing.

[00:28:52] Speaker 02:
The rules didn't provide for a sir reply.

[00:28:56] Speaker 03:
They do now, but at the time... Well, they provided for a motion to strike if it's a new argument, right?

[00:29:02] Speaker 02:
I believe they did provide for a motion to strike, but they also provided the opportunity at the oral hearing.

[00:29:06] Speaker 02:
It was just the sequence was petitioner went first, we went second, and the judge raised this issue first.

[00:29:14] Speaker 02:
But back to what I think is the core of the issue, it's unquestionable if you look at the board's decision, they did actually evaluate

[00:29:22] Speaker 02:
the argument made in the petition.

[00:29:24] Speaker 02:
So they give a detailed analysis.

[00:29:26] Speaker 02:
This is on appendix pages 13 through 15, where they say they considered Dr. Hopwald's declaration, they considered the petition, and they considered the detailed element-by-element analysis contained in the petition.

[00:29:40] Speaker 02:
And they credited the testimony of Dr.

[00:29:45] Speaker 02:
Douglas over the testimony of Dr. Hochwald for this proposed combination.

[00:29:49] Speaker 00:
Can you tell me where you're reading from?

[00:29:51] Speaker 00:
How is that page 14?

[00:29:53] Speaker 02:
Appendix page 14 to 15.

[00:29:54] Speaker 02:
And their conclusion at the top of the first carryover paragraph at the top of page appendix 15

[00:30:00] Speaker 02:
They say, petitioner fails to provide persuasive reasoning as to why one skilled in the art would have modified Hirsch's teaching in a manner that is contrary to the express disclosure of Mark.

[00:30:12] Speaker 02:
So they weighed the expert.

[00:30:13] Speaker 01:
That was with W limited to one, right?

[00:30:16] Speaker 02:
That was given what was presented in the petition.

[00:30:19] Speaker 02:
Got it.

[00:30:19] Speaker 02:
And so I think that's really the fundamental question.

[00:30:24] Speaker 02:
If there were multiple, if petitioner is correct, that they clearly articulated in their petition

[00:30:30] Speaker 02:
Multiple examples, w equals one was one example, but they had others.

[00:30:35] Speaker 02:
It's clear that the board gave them the opportunity to explain that and evaluate that.

[00:30:38] Speaker 02:
So the board evaluated the entire petition, the rationale given by their expert, and concluded that they hadn't met the standard of showing unpatentability.

[00:30:46] Speaker 02:
So the question is, did they get another chance to come up with a new theory of unpatentability in the reply?

[00:30:53] Speaker 02:
The board did consider the portions of the reply that supported the original theory of the petition.

[00:30:59] Speaker 02:
What the board said they would not consider is what they deemed a new theory, changing the way the reference was being read, and changing the proposed modification.

[00:31:08] Speaker 02:
And keep in mind, this is an obviousness combination.

[00:31:10] Speaker 02:
This isn't a question of anticipation.

[00:31:13] Speaker 02:
And so they were proposing to modify these two references to meet the claim.

[00:31:18] Speaker 02:
And so the proposed modification... Why do you think it makes a difference?

[00:31:20] Speaker 01:
I wasn't... I was following you until you said, keep in mind this is obviousness and not anticipation.

[00:31:26] Speaker 01:
Why should it matter?

[00:31:28] Speaker 02:
Because why it should matter is they proposed a way to modify the references in their petition.

[00:31:36] Speaker 02:
Essentially, what they're arguing now is they should get a chance in reply to propose a different way to modify the references.

[00:31:41] Speaker 01:
Even if it was anticipation, under your theory, it would be new, because there would be a different way to read the reference to anticipate the claims.

[00:31:48] Speaker 01:
I don't see the difference.

[00:31:49] Speaker 01:
That's what I'm trying to say.

[00:31:52] Speaker 01:
Unless you have some other reason to add, maybe I'm missing something still.

[00:31:57] Speaker 02:
Sure.

[00:31:58] Speaker 02:
I think it's a matter of they looked at the two references that were proposed to be combined.

[00:32:03] Speaker 02:
They looked at the rationale provided in the petition.

[00:32:05] Speaker 02:
And they concluded that there was additional material not contemplated by the patent office's rules in the reply.

[00:32:12] Speaker 02:
That's what they didn't consider.

[00:32:13] Speaker 02:
But it's clear from the record, they did consider the entirety of the petition.

[00:32:16] Speaker 02:
And that's what the rules provided, that Apple had to give their basis in the petition.

[00:32:21] Speaker 01:
What about the fact that the APAA says that the board has an obligation to consider rebuttal arguments?

[00:32:27] Speaker 02:
if it's merely rebuttal as opposed to raising new issues.

[00:32:31] Speaker 02:
And so that, respectfully, that's the standard of review.

[00:32:34] Speaker 02:
Was there a rational basis to conclude this was new?

[00:32:37] Speaker 02:
I would submit that there was, and therefore, you should defer.

[00:32:41] Speaker 03:
OK.

[00:32:42] Speaker 03:
Thank you, Mr. Bellinger.

[00:32:43] Speaker 03:
Thank you.

[00:32:59] Speaker 04:
I'll try to be efficient here.

[00:33:02] Speaker 04:
I just want to touch on a few things We've heard first just to the site in our brief it said page 51 to 53 of our opening brief where we discuss flame 9 and

[00:33:14] Speaker 04:
and include the discussion, and there's a reference to Figure 7's operation.

[00:33:18] Speaker 04:
Second issue I heard him mention was that both experts agree about the meaning of reasonable or periodically.

[00:33:27] Speaker 04:
And I think that's actually a disputed issue.

[00:33:29] Speaker 04:
The testimony that they were relying on involves our expert saying in response to a deposition question,

[00:33:37] Speaker 04:
What is the meaning of periodically?

[00:33:39] Speaker 04:
And he says, well, outside the context of the patent.

[00:33:43] Speaker 04:
And then he says, inside the context of the patent.

[00:33:45] Speaker 04:
And he goes into a discussion.

[00:33:47] Speaker 04:
The bottom line on this testimony is it's extrinsic evidence.

[00:33:51] Speaker 04:
And if you find, as we've explained, that the intrinsic evidence compels a meaning of periodically, this isn't going to change it.

[00:33:59] Speaker 04:
You can't have contrary expert testimony override what the intrinsic evidence demonstrates the meaning of a term is.

[00:34:06] Speaker 04:
Next thing I want to mention is that we said that they didn't dispute obviousness or anticipation if you find for our construction periodically.

[00:34:14] Speaker 04:
And the basis for that is Appendix 2042-84.

[00:34:22] Speaker 04:
and also 2494.

[00:34:22] Speaker 04:
They didn't raise a different outcome in the case in their briefs before you, whether or not the return periodically changes.

[00:34:32] Speaker 04:
And under our construction, there's no argument that they've advanced in this appeal.

[00:34:38] Speaker 04:
The question of how you review the new theory of patentability was definitively addressed by this court in evasive.

[00:34:44] Speaker 04:
And this is at 841 at 3rd, 966 at 970.

[00:34:49] Speaker 04:
says clearly whether a ground the board relied on was new, requiring a new opportunity to respond is a question of law subject to NOVA review.

[00:34:59] Speaker 04:
And then finally, one thing to keep in mind is that the board found a motivation to combine Hirsch with Martin in its decision.

[00:35:07] Speaker 04:
And that hasn't been disputed.

[00:35:08] Speaker 04:
That's Appendix 13, which goes to our ultimate point.

[00:35:11] Speaker 04:
If you see a motivation to run the Martin algorithm on the Hirsch apparatus, you yield an apparatus which, under Parker vision,

[00:35:18] Speaker 04:
renders unpatentable these claims.