[00:00:06] Speaker 03:
Good morning, Your Honors.

[00:00:06] Speaker 03:
It may please the court.

[00:00:09] Speaker 03:
Jen Digital, formerly Symantec and Norton, appealed for errors from below that culminated in a $600 million final judgment that we ask you to set aside.

[00:00:19] Speaker 03:
Subject to Your Honor's questions, I'd like to focus this morning first on the 101 and infringement issues and the liability issues.

[00:00:25] Speaker 03:
And then if I have time remaining to turn to the foreign sales issue that was raised on appeal.

[00:00:33] Speaker 03:
The 101 issue, I think, is a particularly apt place to start in this case, because after seeing the red brief, Columbia does nothing or little to defend the district court's error.

[00:00:48] Speaker 03:
The district court in this case plainly found the so-called technological improvements that she relied upon to find that the claims were not abstract from the specification and not from the claims.

[00:01:02] Speaker 03:
This court's decisions in American Axel tell us the recitations and the specification that are not in the claims are irrelevant.

[00:01:10] Speaker 00:
OK, let's say that I'm not buying that argument.

[00:01:12] Speaker 00:
I think there's enough groundwork collected that starts the claim language to bring you to the spec.

[00:01:18] Speaker 00:
And I think we've got cases that allow use of the spec.

[00:01:21] Speaker 00:
So let's take that argument off the table for purposes of argument.

[00:01:24] Speaker 00:
What do you have left?

[00:01:25] Speaker 00:
I mean, Columbia argues that creating a model from two different computer models not only increases efficiency but improves reliability or security because the combined models make it harder for the attacker.

[00:01:38] Speaker 00:
Maybe I missed it, but I didn't see where in your briefing you addressed that point.

[00:01:45] Speaker 00:
And in any event, isn't that sufficient to improvement in computer functionality, just besides step one?

[00:01:52] Speaker 03:
It's not, Your Honor.

[00:01:53] Speaker 03:
So if we're turning to the specification, the specification on the models, and we do address this in our brief, the appendix page 287 tells us, and this is column 8, line 15 through 27 of the 322 patent.

[00:02:04] Speaker 03:
We always use the 322 as the exemplar.

[00:02:07] Speaker 03:
It says that the person building the claimed inventions can use, quote, any of a variety of suitable means for combining these models.

[00:02:17] Speaker 03:
That tells us not only is the notion of combining models known, but how to do it is known.

[00:02:22] Speaker 03:
There's no specific method for it.

[00:02:24] Speaker 03:
The claims don't tell you what the model is, how to build it, how to combine it.

[00:02:29] Speaker 03:
You're essentially taking a black box model, combining it with another black box model or an abstraction combined with an abstraction.

[00:02:36] Speaker 03:
And so the specification on the models in particular, we think, is failed to their cause.

[00:02:43] Speaker 03:
Then, if you look at the specification more broadly, it goes out of its way to tell you that everything in the claims is conventional, that it's using things that were known in the art.

[00:02:54] Speaker 03:
At appendix 285, column 3.

[00:02:57] Speaker 01:
That seems to be the heart of the issue, is what's conventional here, and whether there is anything in

[00:03:03] Speaker 01:
particular claims here, which are not conventional.

[00:03:08] Speaker 03:
Yes, Your Honor, I agree.

[00:03:09] Speaker 03:
And our view is strongly held that there's nothing in the claims that's non-conventional.

[00:03:14] Speaker 03:
The model, I think, is their champion argument on that.

[00:03:17] Speaker 03:
I just pointed you to specification language that tells you that you can use any suitable means for combining those models.

[00:03:25] Speaker 03:
If you go to the file history in the IPRs, the board

[00:03:30] Speaker 03:
expressly found that a prior reference called Agarwal taught combined models that were made at different times.

[00:03:36] Speaker 03:
The dispute and the breakdown came whether there was a combined model made from different computers.

[00:03:43] Speaker 01:
I think you'll hear different views on that issue, but when you look at the evidence from the final written decision... Well, some of the claims seem to talk about combined models, which are non-standard models.

[00:03:55] Speaker 01:
And those, but those claims are not asserted here, like in the 115 patent claim 9 and 322 claims 17 and 25.

[00:04:08] Speaker 03:
Yeah, those were invalidated, I think, in these years.

[00:04:14] Speaker 01:
Whether they are 101 eligible or not is not the issue here, but what I get from those claims

[00:04:22] Speaker 01:
is that when they wanted to claim a non-standard model, they knew how to do it, and they didn't do it in these particular claims that we have here.

[00:04:31] Speaker 03:
That's exactly our view, Your Honor, and that the specification further supports that they are telling the world, they're telling the public that

[00:04:39] Speaker 03:
This notion of combining models is a conventional one that can be done by, quote, any suitable means.

[00:04:46] Speaker 03:
And they use that language throughout the specification to point out that they're not using new techniques, no new improved model, no new improved.

[00:04:53] Speaker 00:
Well, what about in the combination?

[00:04:55] Speaker 00:
I mean, this claim survived the IPR, right?

[00:04:59] Speaker 00:
I don't go through the entire record of the IPR, but because there was no motivation to combine, suggesting there's some novelty.

[00:05:09] Speaker 00:
in what went on, and this is not exclusively conventional, right?

[00:05:15] Speaker 03:
So, two points to that, Your Honor.

[00:05:16] Speaker 03:
The board found that there was no motivation to combine, as Your Honor put it,

[00:05:22] Speaker 03:
to preempt what my friend is going to say.

[00:05:25] Speaker 03:
They said there was no motivation to combine that would teach the combined model made from two different computers.

[00:05:33] Speaker 03:
So they did say that language, but they also refer, they acknowledge the argument from Norton at the time and its expert that Agarwal taught combined models from multiple computers.

[00:05:44] Speaker 03:
They say there's just no motivation to combine the systems together to get the combined model together with the other.

[00:05:49] Speaker 03:
So from our perspective, that shows at least that it's conventional there.

[00:05:53] Speaker 03:
The specification tells us it's conventional.

[00:05:56] Speaker 03:
The law is clear.

[00:05:57] Speaker 03:
The SAP case and the Symantec case, for example, that we've cited in our briefs.

[00:06:03] Speaker 01:
What specification teaches it's conventional?

[00:06:07] Speaker 01:
What's the if here?

[00:06:08] Speaker 01:
Using two computers to create a single model?

[00:06:11] Speaker 03:
Just combining models in any way, Your Honor.

[00:06:13] Speaker 03:
The specification, again, it said

[00:06:18] Speaker 01:
There's an argument in the red brief, as I understand it, that the heart of this is efficiency.

[00:06:25] Speaker 01:
You can create a combined model more efficiently by the use of two computers or multiple computers rather than one computer, correct?

[00:06:34] Speaker 03:
The district court said that, too.

[00:06:35] Speaker 03:
Yes, that's the argument.

[00:06:36] Speaker 03:
That is completely contradicted by the specification.

[00:06:40] Speaker 01:
So the judge... What's contradicted?

[00:06:44] Speaker 03:
Let me point it out.

[00:06:45] Speaker 03:
So the judge said the improved efficiency

[00:06:48] Speaker 03:
This is at appendix page 21, was due to the use of distributed sensors, not models, sensors, that those sensors could be used to collectively build models more efficiently.

[00:06:59] Speaker 03:
The specification where I say it contradicts it is Appendix 289.

[00:07:03] Speaker 03:
It's Column 12 beginning of Line 44.

[00:07:06] Speaker 03:
It says you can use any suitable sensors.

[00:07:09] Speaker 03:
It doesn't say you're using special sensors and the claims certainly don't say use distributed sensors in order to achieve more efficiency in combining models.

[00:07:19] Speaker 03:
So the claims don't say distributed sensors.

[00:07:21] Speaker 01:
My understanding of efficiency is that you can do it more quickly by using multiple computers rather than one computer.

[00:07:29] Speaker 03:
So SAP tells us even if it's novel, even if it's groundbreaking, even if it's revolutionary, that doesn't get you to eligibility.

[00:07:37] Speaker 03:
Symantec tells you even if it's non-obvious, it doesn't get you to eligibility.

[00:07:41] Speaker 00:
The oldest case from this court... We're not on the same page.

[00:07:45] Speaker 00:
I think Judge Dyke is trying to help you.

[00:07:47] Speaker 00:
if I'm understanding correctly, that if it's just efficiency, it's not enough.

[00:07:51] Speaker 00:
You use two things rather than one, and that grants you efficiency, and that's abstract.

[00:07:56] Speaker 03:
If I'm arguing with you, and you're trying to help me or honor, I apologize.

[00:07:59] Speaker 03:
I think the court's decision in Neola says that same thing, which is distributing computing in order to make it faster, more efficient, is not also something that's going to take you out of being abstract.

[00:08:10] Speaker 04:
How do you distinguish this case from Finging, number one?

[00:08:13] Speaker 04:
And then number two,

[00:08:15] Speaker 04:
So why am I wrong in looking at this case as not a case representing a technological solution to a technological problem?

[00:08:26] Speaker 04:
So this has only happened in the computer space area.

[00:08:29] Speaker 04:
It seems to me that a problem has been identified.

[00:08:34] Speaker 04:
The claims speak as to addressing that particular problem and the directitude to a specific solution.

[00:08:43] Speaker 03:
So in reverse order, Your Honor.

[00:08:45] Speaker 03:
Even Alice said that it was limited only to computers.

[00:08:48] Speaker 03:
The claims recited computers.

[00:08:50] Speaker 03:
It was going to be something that while a human might be able to do it, it would take too long.

[00:08:54] Speaker 03:
So that it recites computers as a way to make things more efficient, back to the efficiency point, is exactly what Alice and then many decisions in this court have said is abstract.

[00:09:04] Speaker 04:
In the computer world, there is a particular problem, right?

[00:09:07] Speaker 04:
This has to do with cyber security.

[00:09:10] Speaker 03:
It does.

[00:09:11] Speaker 03:
So did fair warning, sort of personal letters, so it's mandatory.

[00:09:13] Speaker 04:
The setting aside that all computers are conventional, there's still problems that arise in the computer space that pertain to that particular space.

[00:09:25] Speaker 04:
And I see this as one of those problems.

[00:09:30] Speaker 04:
Am I wrong on that?

[00:09:32] Speaker 03:
I think you are wrong, Your Honor.

[00:09:33] Speaker 03:
I think all you have here is the abstract concept of taking data from a program, comparing it to a model, and seeing if it's anomalous compared to that model.

[00:09:42] Speaker 03:
That's the organization or the classification of data.

[00:09:45] Speaker 04:
But it combines both models.

[00:09:46] Speaker 04:
There's a combination of both models.

[00:09:48] Speaker 04:
So there's actually a step that you missed there.

[00:09:50] Speaker 03:
Well, on the combined model, so fair warning is maybe my best case on this.

[00:09:55] Speaker 03:
If you go to the fair warning case, which we rely on heavily in our papers, they never mention by name.

[00:10:00] Speaker 03:
There's no response to fair warning, personal web, or semantic in their papers.

[00:10:05] Speaker 03:
They don't even refer to the cases.

[00:10:07] Speaker 03:
Fair warning has a combination too.

[00:10:09] Speaker 03:
It takes data.

[00:10:10] Speaker 03:
There are these audit logs that are being investigated and the data going into those audit logs is combined from multiple sources, which the specification and the opponent, the patent owner, excuse me,

[00:10:23] Speaker 03:
argued, showed, it made it more robust, made the data better, made it so you could learn more about your user and your patient to improve the quality of the computing.

[00:10:33] Speaker 03:
That was not enough, because all you're doing is combining abstractions at that point.

[00:10:38] Speaker 04:
You were going to address FinGen, aren't you?

[00:10:39] Speaker 03:
Yes, thank you.

[00:10:40] Speaker 03:
So FinGen's, I think, very supportive of our position.

[00:10:43] Speaker 03:
In FinGen, what the court found was a new kind, that was the quote, new kind of file that didn't exist before.

[00:10:51] Speaker 03:
It was a file that included something called a downloadable and a security profile.

[00:10:56] Speaker 03:
And not only did it not exist before, but the court found that it did things that weren't done before.

[00:11:02] Speaker 03:
That this new file that never existed before could identify suspicious code and detect dangerous or unwanted operations in a way that was unconventional.

[00:11:14] Speaker 03:
Conventionally, what it was saying was done was essentially what the claims in this patent do.

[00:11:19] Speaker 03:
which is compared to a model or a database and say, is this green light or red light?

[00:11:23] Speaker 03:
Is this a good or a bad action or operation?

[00:11:26] Speaker 03:
Is this access allowed or denied?

[00:11:29] Speaker 03:
That was conventional.

[00:11:31] Speaker 03:
What Finjen said was this is a whole new behavioral approach.

[00:11:35] Speaker 03:
We just don't have that fact pattern here.

[00:11:36] Speaker 03:
There's no new kind of file.

[00:11:38] Speaker 03:
The combined models are conventional.

[00:11:40] Speaker 03:
It's not doing something new that couldn't have been done before.

[00:11:43] Speaker 03:
And I don't think there's been any evidence to the contrary on any of those things.

[00:11:48] Speaker 03:
I think your honor is going to address step two yourselves as a matter of law.

[00:11:58] Speaker 03:
It's certainly in your purview.

[00:12:01] Speaker 00:
We were talking a lot in the other case about due process and sufficient notice.

[00:12:08] Speaker 00:
There was nothing that went down here.

[00:12:10] Speaker 00:
I mean, this case went down on a motion to dismiss, and then it just went away at trial.

[00:12:16] Speaker 00:
So how are we satisfied if the parties had an opportunity to make their arguments and someone had an opportunity to review those arguments?

[00:12:24] Speaker 03:
So I think where we sit on that is the, you're right, it went down on the motion, rule 12.

[00:12:30] Speaker 01:
Did the motion to dismiss cover both steps one and 10?

[00:12:34] Speaker 03:
The motion did.

[00:12:35] Speaker 03:
The opposition did.

[00:12:37] Speaker 03:
The judge found that as it was abstract, for the reasons I said, and so never reached step two, in a clarification order, the judge then said, I'm throwing this out as a matter of law, the whole thing.

[00:12:51] Speaker 03:
I think the parties addressed step two, the court didn't.

[00:12:55] Speaker 03:
Your honors in at least Free Stream Media and I think it was CardioNet have addressed step two, even when the district court didn't get to it, when the district court did what this one did, decided on step one.

[00:13:08] Speaker 03:
So you could, if you choose to, and we would certainly urge you to, because we don't think there's anything new that needs to happen down in the district court.

[00:13:15] Speaker 03:
to assess the intrinsic evidence here.

[00:13:17] Speaker 03:
It's a de novo question anyway.

[00:13:19] Speaker 03:
It's in front of your honors.

[00:13:20] Speaker 03:
You can resolve this case for that.

[00:13:21] Speaker 03:
So we would ask you to go ahead and do step two.

[00:13:24] Speaker 03:
You certainly could send it back if you chose to.

[00:13:27] Speaker 03:
Back to the specification, if I might go.

[00:13:32] Speaker 03:
I just wanted to point out that the specification tells us that everything here is conventional.

[00:13:38] Speaker 03:
Appendix 285, column 3, putting in line 16, it says you can use any suitable algorithm for anomaly detection.

[00:13:46] Speaker 03:
You can use the stem or the valve grind emulators that were known prior to our emulators.

[00:13:51] Speaker 03:
You can use, quote, any other suitable technique.

[00:13:54] Speaker 03:
And you see this for all of the different pieces and parts of the claims.

[00:13:57] Speaker 03:
And I don't have time to go through each of them, but I'd point out Appendix 287, Column 8 beginning at Line 15, Appendix 289, Column 11 beginning at Line 19, and Appendix 293, Column 19 beginning at Line 49.

[00:14:14] Speaker 03:
I'm almost out of time.

[00:14:14] Speaker 03:
I'm going to turn to the infringement issue unless Your Honor said questions for me further.

[00:14:17] Speaker 00:
I probably went to the enhancement issue.

[00:14:20] Speaker 03:
The enhancement issue?

[00:14:21] Speaker 03:
So, Your Honor, we would ask you to reverse on the enhancement issue because it's premised on two pieces of analysis that we think are both faulty.

[00:14:34] Speaker 03:
One is on willfulness.

[00:14:35] Speaker 03:
The second is on the so-called litigation misconduct.

[00:14:40] Speaker 03:
which I understand and I'll concede is a factual issue with a lot of... Well, you've got two pieces to your argument.

[00:14:46] Speaker 00:
One goes to the case we're about to hear next, which is assuming hypothetically we remove everything dealing with the dossier litigation in this case.

[00:14:56] Speaker 00:
We're still left with findings and a conclusion by the district court of the enhancement.

[00:15:03] Speaker 00:
And she suggests, without really saying clearly, that she's not considering

[00:15:08] Speaker 00:
the negative inference in reaching her conclusion.

[00:15:12] Speaker 03:
May I answer?

[00:15:13] Speaker 03:
Yeah.

[00:15:13] Speaker 03:
She says that she also says that the so-called litigation of conduct weighs as heavily towards enhancement as it possibly could.

[00:15:22] Speaker 03:
So she says both things.

[00:15:24] Speaker 03:
And we think that clearly when you read the analysis, you could see that the finding on litigation misconduct, which we strenuously disagree with,

[00:15:32] Speaker 03:
was very, very important to the way she did her decision.

[00:15:35] Speaker 00:
What about the rest of it?

[00:15:36] Speaker 00:
So the rest of it goes to willfulness.

[00:15:37] Speaker 00:
You'd get a remand if there's something left to it, and the judge will get a do-over, right, even if the egregiousness doesn't matter.

[00:15:44] Speaker 03:
I think if you decide it's a matter of law, there's no willfulness, then there's no need for a remand.

[00:15:48] Speaker 03:
OK, let's assume you don't.

[00:15:50] Speaker 03:
But then I think you're back to remand.

[00:15:52] Speaker 03:
Then I think that probably would go to a remand if you found that there's willfulness.

[00:15:56] Speaker 03:
Yes.

[00:15:57] Speaker 03:
But the litigation misconduct was way too heavy.

[00:15:59] Speaker 00:
So then there were other questions.

[00:16:00] Speaker 00:
Was there anything left other than if you take away the inference and the dossier dealing with the other litigation and the other patent?

[00:16:10] Speaker 00:
You challenged in your brief the closeness of the case, her argument that it was not close, and you challenged other pieces of her taking it apart.

[00:16:18] Speaker 00:
There's a number of lead factors that I think the parties have addressed both.

[00:16:21] Speaker 00:
But really, the closeness of the case, I mean, it seems to be kind of a legal question.

[00:16:24] Speaker 00:
Can you consider an enhancement of all the other stuff that went on beforehand and the stuff that you prevailed?

[00:16:34] Speaker 00:
answer to that, right?

[00:16:35] Speaker 03:
I think the way I would read the case law is that you look at the totality of the circumstances on the closest to the case.

[00:16:42] Speaker 03:
I think it's wrong, as Columbia says, that you should disregard the things that happened pre-trial.

[00:16:48] Speaker 00:
Well, how does that jive with the point you were making four minutes ago, which is we should take away everything that happened in the dossier stuff?

[00:16:54] Speaker 00:
That stuff was part of this litigation initially.

[00:16:58] Speaker 00:
So you can't have it both ways.

[00:17:00] Speaker 00:
Either we're supposed to include everything starting back five years ago, or we're not.

[00:17:05] Speaker 03:
I don't disagree.

[00:17:06] Speaker 03:
I'm not asking to have it both ways.

[00:17:07] Speaker 03:
But the dossier argument, the dossier so-called misconduct, went to an issue we want in front of the jury.

[00:17:14] Speaker 03:
It went to the question of fraudulent concealment.

[00:17:17] Speaker 03:
What they say Dossier was going to do was come to court and say, I think this patent that Norton has should have your name on it.

[00:17:26] Speaker 03:
That was a fraudulent concealment claim about Norton's patents, not about the patents of Sue.

[00:17:30] Speaker 00:
Fair enough.

[00:17:30] Speaker 00:
But if we're talking about the closeness of the case, we're dealing with the infringement findings of the jury, which were narrow.

[00:17:37] Speaker 00:
Why should other stuff that happened before necessarily, as a matter of law, be required to be considered by the district court in this circumstance?

[00:17:46] Speaker 03:
I think when you look at closeness of the case, you look at the case as a whole, not just the last three weeks of the case.

[00:17:52] Speaker 03:
And the case as a whole is one in which over and over again, Norton had significant success.

[00:17:58] Speaker 03:
Claim construction went its way.

[00:18:00] Speaker 03:
All six patents were found not to be infringed.

[00:18:03] Speaker 03:
Most of the claims were invalidated in IPR.

[00:18:06] Speaker 03:
Most of the claims, even of the patents in suit, were out of the case.

[00:18:10] Speaker 03:
And so if you're talking about closeness of the case, which decisions of this court have done,

[00:18:15] Speaker 03:
broadly, then you have to look at everything, not just the jury trial.

[00:18:21] Speaker 04:
Could you address the extraterritorial issues and foreign cells?

[00:18:26] Speaker 03:
Yes, Your Honor.

[00:18:28] Speaker 03:
So on the foreign sales point, maybe what I'll do to try to be as brief as I can is just go straight to Brumfield.

[00:18:35] Speaker 03:
So this court's decision in Brumfield, we think, is dispositive in establishing that the foreign sales in this case should not have been included in the calculation of the reasonable royalty.

[00:18:45] Speaker 00:
Well, they didn't define what proximate cause is.

[00:18:47] Speaker 00:
I mean, Brumfield left some issues open.

[00:18:50] Speaker 03:
It did.

[00:18:51] Speaker 00:
But there's a but for it.

[00:18:52] Speaker 03:
I'm sorry.

[00:18:53] Speaker 03:
I didn't mean to interrupt you.

[00:18:55] Speaker 03:
Go ahead.

[00:18:56] Speaker 03:
we believe that the facts in this case fail the Brumfield but for causation test not even if you agree with you that proximate cause is not fully vetted yet as far as what it means but it establishes a but for test that says did the for the domestic conduct

[00:19:14] Speaker 03:
was the value of that infringement increased by the foreign sales.

[00:19:21] Speaker 03:
And when you go through the facts, so if you look at the facts of Brumfield is supplying Western GECO, the master software in both cases was made, developed, updated, tested in the United States.

[00:19:35] Speaker 03:
The master software, there was a CRM claim in both patents from Brumfield in this case.

[00:19:42] Speaker 03:
Is that the CRM claim?

[00:19:44] Speaker 03:
There is a CRM claim in both, John.

[00:19:45] Speaker 03:
There's a CRM claim here, claim 11, and there was a CRM claim in Brumfield as well.

[00:19:49] Speaker 04:
And both the master disk was made in the US.

[00:19:52] Speaker 03:
Made in the US, developed, tested, all of that.

[00:19:54] Speaker 03:
And maybe most importantly is that master software was made before the patents issued in Brumfield and here.

[00:20:02] Speaker 03:
And that's really important to the approximate cause, excuse me, the but for cause question, because the overall software, the instructions that go into what they say in Fringe,

[00:20:12] Speaker 03:
were done in Sonar Bash, the accused products, two years before the patents started to issue at least.

[00:20:18] Speaker 03:
And so the same thing was true in Brumfield.

[00:20:21] Speaker 03:
And what that left was a question, well, what happens two years later when the patents issue that could not have happened overseas?

[00:20:30] Speaker 03:
Brumfield says, let's assume, just for sake of argument, that the manufacturer, the first initial development design, all that stuff had to happen in the U.S.

[00:20:38] Speaker 03:
You could make the same assumption here.

[00:20:41] Speaker 03:
Let's just say it had to happen.

[00:20:42] Speaker 03:
The question then becomes, well, what about two years later when the patents issue?

[00:20:46] Speaker 03:
Why couldn't it have been replicated overseas at that point in a way that would have been non-infringing?

[00:20:51] Speaker 03:
And there's no evidence in this case.

[00:20:53] Speaker 03:
There's no jury finding.

[00:20:54] Speaker 03:
There's nothing from the judge that says either the later enhancements to sonar bash when the patents came out.

[00:21:02] Speaker 03:
were themselves adding to the infringement were necessary, necessary to enable the foreign sales because you could have replicated it overseas.

[00:21:10] Speaker 03:
And there's nothing in the verdict, the final judgment, the judge's analysis that says that back to the Budfort question, that the value of that making of the software of the master was somehow enhanced by the foreign sales.

[00:21:25] Speaker 03:
So the making was excluded and affirmed for being excluded in Brumfield.

[00:21:31] Speaker 03:
It was allowed to happen and go to the jury in this case.

[00:21:34] Speaker 03:
That was legal error.

[00:21:36] Speaker 03:
There's a lot more I could say on Brumfield, Your Honor.

[00:21:38] Speaker 03:
I recognize I'm way over my time, though.

[00:21:40] Speaker 03:
Did you want it addressed further?

[00:21:42] Speaker 01:
I think we're out of time.

[00:21:43] Speaker 01:
We'll give you two minutes for a model.

[00:21:45] Speaker 01:
Thank you, Your Honor.

[00:21:54] Speaker 02:
Judge Dyke, and may it please the court.

[00:21:58] Speaker 02:
I first wanted a point of clarification.

[00:22:02] Speaker 02:
The four claims that were asserted were claims to 11 and 27 of the 322 and claim 2 of the 115.

[00:22:12] Speaker 02:
And all of those claims have the combined model limitation that says wherein the model is a combined model created from models created using different computers.

[00:22:24] Speaker 02:
So all of the claims have that limitation.

[00:22:27] Speaker 01:
Let's talk about

[00:22:28] Speaker 01:
whether these claims include things which are not conventional.

[00:22:33] Speaker 01:
It may be that the specification has things in it which are not conventional but I'm not quite understanding what's not conventional about the claims that are asserted here.

[00:22:45] Speaker 01:
And, you know, to some extent, you argue that it features seal activity, except that the claims don't require seal activity.

[00:22:54] Speaker 01:
What is it that's in the claims here that's not conventional?

[00:22:57] Speaker 01:
I mean, you would agree, I assume, that merely dividing a task that might have been performed by one computer among two or more computers is conventional.

[00:23:09] Speaker 01:
That's not something that renders it not abstract.

[00:23:13] Speaker 02:
I agree with your honor.

[00:23:14] Speaker 02:
But if I may, there is a lot more to these patent claims than just dividing a task among multiple computers.

[00:23:22] Speaker 02:
And if I could just spend like two or three minutes drilling down on the claimed invention, I'd like to persuade you that there is a lot more here than just the concept your honor mentioned.

[00:23:35] Speaker 02:
If we think back to when the professors came up with this invention, it's 22 years ago.

[00:23:41] Speaker 02:
I was graduating high school.

[00:23:43] Speaker 02:
And the Columbia professors gave computer systems the ability to detect new malware.

[00:23:50] Speaker 01:
We can't talk about the claims.

[00:23:53] Speaker 02:
Yeah, and if we look at the claim elements, there are really two parts to it, Judge Dyke.

[00:24:01] Speaker 02:
The first is the professors discovered that normal programs tend to access or call the operating systems.

[00:24:10] Speaker 01:
But you keep talking about what the professors invented.

[00:24:14] Speaker 01:
And it may be they invented something that's

[00:24:17] Speaker 01:
non-abstract.

[00:24:20] Speaker 01:
I'm talking about the claim language here.

[00:24:22] Speaker 01:
I mean, you have a complex specification which may well include things that are non-abstract.

[00:24:30] Speaker 01:
The question is, do these claims here, which claim two appears to be representative, incorporate that?

[00:24:39] Speaker 02:
Yes, they do, Your Honor.

[00:24:41] Speaker 02:
So when I'm talking about what the professors invented, I'm going to tie everything to something that's in the claim language.

[00:24:48] Speaker 02:
We can agree that the claim language includes function calls.

[00:24:53] Speaker 02:
In fact, the entire claim is built around function calls.

[00:24:58] Speaker 02:
And what the professors discovered is that the function calls made by normal programs have recurrent patterns.

[00:25:06] Speaker 02:
And to translate that into a technological solution, the claims talk about a model of function calls and comparing a function call made by a program.

[00:25:18] Speaker 01:
And modeling was conventional.

[00:25:20] Speaker 01:
That was well known, right?

[00:25:21] Speaker 02:
Not modeling function calls, Your Honor.

[00:25:24] Speaker 02:
So nobody was modeling function calls at the time.

[00:25:27] Speaker 02:
And the first half of the professor's invention here.

[00:25:32] Speaker 01:
Where do I find that?

[00:25:35] Speaker 01:
Where do I find the idea that modeling function laws wasn't conventional?

[00:25:40] Speaker 02:
Well, so if the court looks at appendices A and B to the provisional patent application, for example, at appendix 4189, and then this is mentioned in the specification of the 322 at column 1, 26 to 33, at the time

[00:26:01] Speaker 02:
computer defenses used rules that were based on the attributes of known programs.

[00:26:07] Speaker 02:
And those didn't work when it came to a new type of attack.

[00:26:11] Speaker 02:
So what the professors did was look.

[00:26:15] Speaker 00:
Can you, I'm not going to dig through the appendices right now, but since you mentioned column one of 322, can you tell us what you're talking about in column one?

[00:26:22] Speaker 02:
Yes, Your Honor.

[00:26:24] Speaker 02:
So it's mentioned briefly here in column one of the 322 at lines 26 to 33.

[00:26:33] Speaker 02:
But it talks about the various types of computer attacks.

[00:26:37] Speaker 02:
And it says that many computers are protected by anti-virus software and firewalls.

[00:26:42] Speaker 02:
However, these preventative measures are not always adequate.

[00:26:47] Speaker 01:
And what that means, Your Honor, to a person who's- But where does it say modeling function calls

[00:26:55] Speaker 01:
was not conventional.

[00:26:58] Speaker 01:
And you talk about what the professors invented.

[00:27:03] Speaker 01:
That's very interesting.

[00:27:04] Speaker 01:
But I've got to find it in the claim as construed in the latest specification.

[00:27:11] Speaker 02:
I agree, Your Honor.

[00:27:12] Speaker 02:
But what I'm saying is the use of function calls specifically to identify malware, abnormal program executions, which is what the claim language says,

[00:27:25] Speaker 02:
Nobody had done that before.

[00:27:28] Speaker 01:
Is that your view, is that the one thing that's not conventional about a claim too?

[00:27:35] Speaker 02:
No, Your Honor.

[00:27:36] Speaker 02:
I think that as construed, the selective emulation, because the claim.

[00:27:43] Speaker 01:
Selectivity is not a claim requirement.

[00:27:45] Speaker 01:
It is within the claim, but it doesn't require selectivity.

[00:27:51] Speaker 02:
But the court assesses step one of the alics.

[00:27:54] Speaker 01:
Let's put that aside for a moment.

[00:27:55] Speaker 02:
Well, but the court includes in its assessment the claim constructions, because the constructions define the meets and bounds of the claim.

[00:28:03] Speaker 01:
The construction is not that this is limited to seal activity.

[00:28:07] Speaker 01:
The claim on its face is very clear that it's not limited to seal activity.

[00:28:13] Speaker 02:
But the claim construction requires selective execution, Your Honor.

[00:28:18] Speaker 02:
And the claim talks about executing at least part.

[00:28:22] Speaker 01:
Where does the claim construction require seal activity?

[00:28:25] Speaker 02:
The claim construction is software.

[00:28:27] Speaker 02:
Where?

[00:28:28] Speaker 01:
At the end.

[00:28:28] Speaker 01:
Where's the claim?

[00:28:30] Speaker 01:
What page?

[00:28:31] Speaker 02:
Of the court's claim construction order, Your Honor.

[00:28:34] Speaker 01:
What page?

[00:28:35] Speaker 02:
Apologies.

[00:28:48] Speaker 02:
Apologies, your honor.

[00:28:50] Speaker 02:
I don't have the citation right at hand.

[00:28:52] Speaker 01:
But the district court... You're supposed to come here.

[00:28:54] Speaker 01:
You make arguments.

[00:28:55] Speaker 01:
You're supposed to be able to point us to where the stuff you're relying on appears.

[00:28:58] Speaker 02:
I apologize, Judge Dyke.

[00:29:01] Speaker 02:
I have the construction memorized, which is software alone are in combination with hardware that permits the monitoring and selective execution of parts or all of a program.

[00:29:13] Speaker 01:
Parts or all.

[00:29:14] Speaker 01:
It includes all.

[00:29:15] Speaker 02:
But it has to be selective execution.

[00:29:18] Speaker 01:
OK.

[00:29:20] Speaker 01:
I don't want to argue with you about this, but it seems to be on the face of it.

[00:29:24] Speaker 01:
It clearly includes law.

[00:29:26] Speaker 02:
Well, no, but what was novel and what was not conventional, Judge Dyke, and this is described in Appendix C to the provisional in a fair amount of detail, what was not conventional

[00:29:39] Speaker 02:
was using the emulator to selectively execute parts or all of the program.

[00:29:45] Speaker 01:
Maybe so, but that's not a requirement of the claim.

[00:29:47] Speaker 01:
That's the problem.

[00:29:48] Speaker 01:
It says select or all, parts or all.

[00:29:53] Speaker 01:
But it has to be selectively.

[00:29:54] Speaker 01:
OK, but I don't want to argue with you about this.

[00:29:56] Speaker 01:
Let's pass over this here.

[00:29:58] Speaker 01:
And is there any other feature of this that's not conventional?

[00:30:03] Speaker 02:
I believe it's the combination, Judge Dyke, of selective emulation

[00:30:08] Speaker 02:
together with the combined model of function calls, which when those things were put together, they changed the way that a computer... Did the district court rely on the modeling of function calls as being not conventional?

[00:30:27] Speaker 02:
The district court relied on the model of function calls itself as being the key point.

[00:30:34] Speaker 01:
Okay, could you show me where it talks about that?

[00:30:38] Speaker 02:
Yes, Your Honor.

[00:30:40] Speaker 02:
In the court's 101 analysis, which is

[00:31:08] Speaker 02:
Here at Appendix 19 to 21, Your Honor, which is the court's decision denying Northeastern.

[00:31:16] Speaker 01:
Where is the specific language about modeling function falls not being conventional?

[00:31:27] Speaker 04:
What was that page number again, Counsel?

[00:31:30] Speaker 02:
I have Appendix 21 to 23, Judge Reyna.

[00:31:39] Speaker 02:
And apologies, the most specific language I would like to rely on is Appendix 24.

[00:31:46] Speaker 02:
And at the bottom of the page, Judge Dyke, the Court says, similar to the patent claims in FinGen, the 322 patent utilizes a new type of model that improves the efficiency of computer virus screening.

[00:32:00] Speaker 01:
But it doesn't refer to function calls.

[00:32:03] Speaker 02:
Well, but what the Court's talking about is the claims model of function calls, Your Honor.

[00:32:09] Speaker 01:
Where's this in your brief?

[00:32:11] Speaker 02:
In our brief?

[00:32:13] Speaker 01:
Yeah.

[00:32:13] Speaker 01:
Where does the brief talk about what's not conventional here is the modeling function call.

[00:32:20] Speaker 02:
Well, we address that.

[00:32:26] Speaker 02:
We talk about the model of function calls not being conventional at page 39 of the blue brief, the nine of the red brief, Your Honor.

[00:32:40] Speaker 02:
We say, as construed, the model must be created using the function calls actually made during program execution.

[00:32:56] Speaker 02:
and we point out it cannot be created based on a program's static code.

[00:33:00] Speaker 02:
We're talking about modeling function calls, which is a specific action that occurs in a... Okay, but where does it say that modeling a function call is something that's new?

[00:33:13] Speaker 02:
Is something that's new?

[00:33:15] Speaker 02:
Yeah.

[00:33:15] Speaker 02:
I believe that's what we were intending here on page 39, Judge Dyke.

[00:33:22] Speaker 01:
Okay.

[00:33:22] Speaker 01:
All right, go ahead.

[00:33:24] Speaker 02:
But I want to emphasize, what we're talking about in these patent claims is something that actually changed the way that the computer system worked.

[00:33:35] Speaker 02:
When you put this software on the computer, it was able to detect malicious programs that it could not detect before.

[00:33:44] Speaker 02:
If we look at precedent from this court, the Fingen case, and the SRI case, for example, and even Enfish,

[00:33:53] Speaker 02:
The claims here have more detail.

[00:33:56] Speaker 02:
They have more specificity about how to protect a computer than the eligible claims that we had in Finjin, SRI, and ENFISH.

[00:34:06] Speaker 01:
And I think that what we have did is different.

[00:34:08] Speaker 01:
I mean, each of these cases is different.

[00:34:11] Speaker 01:
Finjin created a new computer file, which was downloadable, which was attached.

[00:34:16] Speaker 01:
We don't have that here.

[00:34:17] Speaker 02:
It's different.

[00:34:20] Speaker 02:
In FinGen, the security profile that was created for the downloadable, as construed by the court, that was simply the identification of code that may have a malicious operation.

[00:34:35] Speaker 02:
And there was no description about how you do that.

[00:34:38] Speaker 02:
And if it was enough in FinGen to have the new security profile, it should be enough here to have a new model of function calls

[00:34:47] Speaker 02:
that when it's put on a computer, allows the computer to detect an abnormal execution that indicates an attack.

[00:34:54] Speaker 04:
Are the claims here an advancement in the technology in Fingin?

[00:35:01] Speaker 02:
Yes, Judge Raina.

[00:35:02] Speaker 02:
So what the Fingin technology did

[00:35:05] Speaker 02:
was it scanned static code.

[00:35:08] Speaker 02:
So you'd have a downloadable that came onto a computer system, like through email.

[00:35:13] Speaker 02:
And the Fingen claim scanned the static code.

[00:35:18] Speaker 02:
The claim specifically required an inspector.

[00:35:21] Speaker 02:
And then as construed by the court, that inspector identified code that may contain malicious operations and put it in what was called a security profile that was attached to the program.

[00:35:33] Speaker 02:
This is an advancement over FinGen, and the FinGen tech does not infringe these claims, because we're not using static code.

[00:35:41] Speaker 02:
The program actually has to execute on the computer, and that requires the emulator to protect the system.

[00:35:49] Speaker 02:
What was happening in FinGen, the program didn't execute, so it couldn't cause damage.

[00:35:56] Speaker 01:
Emulator is not new, right?

[00:35:58] Speaker 01:
Please have an emulator.

[00:35:59] Speaker 02:
The professor's emulator was new, as described in Appendix C to the provisional patent application.

[00:36:06] Speaker 01:
What they did was... But the use of emulators itself is not new, right?

[00:36:11] Speaker 01:
That was common.

[00:36:14] Speaker 02:
I will give you, Judge Dyke, that emulators were known.

[00:36:17] Speaker 02:
The professor's selective emulator that could do all the program or choose to do parts, nobody had ever done that before.

[00:36:28] Speaker 02:
If I could touch on step two briefly.

[00:36:31] Speaker 04:
Before you go, so if you practice the claim at an issue, are you in and of itself practicing conventionality?

[00:36:44] Speaker 04:
Is it conventional just because computers are involved?

[00:36:48] Speaker 04:
Is there something else we need to look at?

[00:36:51] Speaker 02:
Well, it's not conventional just because computers are involved.

[00:36:55] Speaker 04:
uh... you know as this court well i don't i don't think so either otherwise you know all code and all computer pans would be out the window so but our jurisprudence is that if you're using a conventional computer and practicing the claims uh... renders that computer into non-conventional computer or a special computer is that what we have here that is drayna so what we're talking about is software

[00:37:25] Speaker 02:
Just like in the McGrow case, just like in Enfish, just like in FinGen, just like in SRI, it's software.

[00:37:33] Speaker 02:
But when you download that software onto a computer, you change the way the system works.

[00:37:41] Speaker 02:
It is not simply performing a process or automating a process with a conventional computer, because the process in the claims doesn't exist outside of a computer system.

[00:37:53] Speaker 04:
So you're saying the model of function calls, that's what's not conventional here.

[00:38:00] Speaker 02:
I believe it is the combination, Judge Raina, of the model of function calls.

[00:38:08] Speaker 02:
It was not conventional to screen for malware with function calls.

[00:38:12] Speaker 02:
There was no computer code that reflected sequences of function calls to detect malware before the professors came up with their invention.

[00:38:23] Speaker 02:
And I believe that it is that, in combination with the selective emulation of all or part of the program, which Appendix C to the provisional says created enormous efficiency benefits when compared to a virtual machine or another sandbox.

[00:38:39] Speaker 02:
3,000% system improvement, Appendix C says, at Appendix 4220.

[00:38:45] Speaker 01:
So to get away from that selectivity feature, let's assume

[00:38:50] Speaker 01:
that we conclude, as I was suggesting earlier, that the claim isn't limited to selectivity.

[00:38:56] Speaker 01:
Let's say that's not part of the claim.

[00:39:00] Speaker 01:
Is this still non-abstract?

[00:39:03] Speaker 02:
Yes, absolutely, Judge Stike.

[00:39:05] Speaker 02:
The use of function calls on its own to detect malware was entirely new.

[00:39:15] Speaker 02:
If the court looks at the professor's testimony, for example, if you reach step two, and if the court looks at appendix 522.05 to 09 and 524.37 to 45, the professors explain why the use of function calls to detect malware was itself unconventional.

[00:39:37] Speaker 02:
Nobody was doing it.

[00:39:39] Speaker 02:
and then modeling the function calls.

[00:39:42] Speaker 00:
Beyond what the professors, anything in the spec, what's the best you have in the spec?

[00:39:49] Speaker 00:
You gave us claim one.

[00:39:51] Speaker 00:
I'm sorry, column one a few minutes ago.

[00:39:53] Speaker 00:
But I'm not that good with relying on what the professor said.

[00:39:59] Speaker 00:
I'd rather see it in the spec.

[00:40:01] Speaker 02:
Understood, Judge Gross.

[00:40:05] Speaker 02:
To be candid with you, the best intrinsic record citations I can give you are appendices A, B, and C to the provisional application, which unpack some of the points that are made more briefly.

[00:40:20] Speaker 01:
It's not in the spec.

[00:40:23] Speaker 01:
322 or the 115?

[00:40:25] Speaker 02:
But for step 101, the court can look to the intrinsic record.

[00:40:31] Speaker 01:
The answer.

[00:40:31] Speaker 01:
It's not in the spec for the 115 or the 322.

[00:40:34] Speaker 02:
Well, it is at column 126 to 33 saying that this invention was addressing inadequacies in antivirus software and firewalls.

[00:40:46] Speaker 01:
That's pretty general.

[00:40:47] Speaker 02:
Well, but if a person of skill in the art would understand what that meant, and they would look to the entirety of the intrinsic record, and appendices A, B, and C to the provisional, which were the professor's groundbreaking research papers,

[00:41:03] Speaker 02:
That provides a full unpacking of what it means for the invention to address the inadequacies of existing antibiotics.

[00:41:11] Speaker 00:
And where is that cited in red?

[00:41:13] Speaker 00:
Did you cite in your red brief in connection with one-on-one?

[00:41:16] Speaker 00:
Do you cite these provisional appendices?

[00:41:19] Speaker 02:
Yes, we do.

[00:41:21] Speaker 02:
So Judge Prost, in the statement of the case, when we are describing the invention,

[00:41:37] Speaker 02:
For example, on page seven of the registry, the paragraph on selective emulation, we specifically say as described in Appendix C to their provisional, and we cite appendix.

[00:41:49] Speaker 01:
That's a different point.

[00:41:51] Speaker 01:
That's a different point.

[00:41:53] Speaker 02:
And we cite on page five, Judge Dyke, where we're talking about the detecting new attack modes.

[00:42:02] Speaker 02:
which continues on to page 6.

[00:42:05] Speaker 02:
We specifically say, as the professors described in Appendix A to their provisional patent application.

[00:42:12] Speaker 02:
And we cite to Appendix 4188 to 96.

[00:42:15] Speaker 00:
But then you have a whole section that deals with 101 separately, right?

[00:42:20] Speaker 00:
In your argument section, page 33 to 41, almost 10 pages.

[00:42:25] Speaker 00:
Is that referenced back?

[00:42:27] Speaker 00:
It is.

[00:42:28] Speaker 01:
It is, Judge Frost.

[00:42:29] Speaker 01:
Where does the 101 discussion refer to these appendices?

[00:42:33] Speaker 02:
But we supersite Judge Dyke back to our description of the invention, just so that we didn't have to repeat everything that we had said about the invention.

[00:42:44] Speaker 02:
It's not in the 101 discussion.

[00:42:46] Speaker 02:
No, it is, Judge Dyke.

[00:42:47] Speaker 01:
Specific references to these appendices are not in the 101 discussion.

[00:42:52] Speaker 02:
Judge Dyke, I believe that some of the appendix page numbers are cited in the 101 discussion, but we do supersite

[00:43:01] Speaker 02:
in 101 back to the description of the invention.

[00:43:04] Speaker 01:
Where does the description of the invention talk about the function?

[00:43:11] Speaker 02:
So the page number that I just gave the court, which is page six, for example,

[00:43:24] Speaker 02:
specifically talks about the benefits of modeling function calls made by executing programs in the ordinary course, and then comparing a function call made by an unknown program to the model.

[00:43:37] Speaker 01:
I mean.

[00:43:39] Speaker 01:
But again, it doesn't say that this is not convention.

[00:43:42] Speaker 01:
It doesn't tell us that this was something new.

[00:43:47] Speaker 02:
Judge Steich, if you get to step two,

[00:43:51] Speaker 02:
The professors have testified now that this was not conventional.

[00:43:55] Speaker 02:
The appendices to the provisional patent application say it was not conventional.

[00:44:00] Speaker 02:
And ultimately, after Berkheimer, if there is a fact fight about whether- Could you give us an example from one of the appendices?

[00:44:09] Speaker 01:
Where did we find that?

[00:44:10] Speaker 02:
Yes.

[00:44:11] Speaker 02:
So page 4189 of the appendix, Judge Dyke.

[00:44:14] Speaker 01:
Which volume?

[00:44:17] Speaker 02:
I think that is volume two, Judge Dyke.

[00:44:39] Speaker 02:
And Appendix 4189 is the first.

[00:44:42] Speaker 02:
We're on the page.

[00:44:45] Speaker 02:
So if you look at the introduction, Judge Seidt, it says this talks about the deficiencies in existing computer defenses.

[00:45:00] Speaker 02:
And it says.

[00:45:00] Speaker 01:
This doesn't talk about modeling a function call speed.

[00:45:05] Speaker 02:
Well, this talks about modeling all anomalies from running programs being nipped.

[00:45:11] Speaker 01:
Where is the talk about modeling function policies nipped?

[00:45:14] Speaker 02:
It says here.

[00:45:19] Speaker 04:
Where are you reading from, counsel?

[00:45:21] Speaker 02:
Reading from page 4189, which is the first page of appendix A to the provisional patent application.

[00:45:32] Speaker 02:
And it talks about the fact that we describe a new approach.

[00:45:38] Speaker 01:
Where?

[00:45:39] Speaker 01:
Where are you reading from?

[00:45:39] Speaker 02:
I'm reading from the top of column two, Judge Dyke, on appendix 4189.

[00:45:45] Speaker 02:
It says, we describe a new approach based on anomaly detection using a method that trains on normal data and looks for anomalous behavior that deviates from the normal mode.

[00:46:00] Speaker 02:
This method can better identify unknown attacks.

[00:46:03] Speaker 02:
And then the paper's talking about access to the Windows Registry, which is- What you read doesn't talk about function calls.

[00:46:10] Speaker 02:
Well, but Windows Registry access is a function call.

[00:46:15] Speaker 02:
This paper talks about why that was new, Judge Dye.

[00:46:19] Speaker 01:
OK, why don't you take a couple of minutes to talk about the other issues that were addressed.

[00:46:24] Speaker 00:
Can I just make one comment, which is it's very challenging for us and sort of frustrating.

[00:46:29] Speaker 00:
I mean, you've led with efficiency, and then you seem to have given that up.

[00:46:33] Speaker 00:
And now we're pivoting to modeling of function calls, which wasn't really an argument fully developed, at least in the argument section of 101, was it?

[00:46:45] Speaker 02:
Well, just to, and I apologize for the frustration, of course.

[00:46:50] Speaker 02:
I'm not giving up efficiency.

[00:46:52] Speaker 00:
I think that the- OK, well, let's turn to the function calls.

[00:46:56] Speaker 00:
I mean, in fairness to you, I did find one sentence in red that says the invention accomplished something the conventional methods could not.

[00:47:04] Speaker 00:
And it cites to back to five and six.

[00:47:06] Speaker 05:
Correct.

[00:47:07] Speaker 00:
So in fairness to you, there is some incorporation of that in the argument section.

[00:47:11] Speaker 00:
But it's hardly a fully developed argument.

[00:47:15] Speaker 02:
Well, but I guess, Judge Prost, we didn't want to repeat everything that we had said in the description of the invention.

[00:47:23] Speaker 01:
But you're saying now this is the heart of your argument, and it's not developed in the brief.

[00:47:27] Speaker 02:
Well, no.

[00:47:28] Speaker 02:
Judge Stike, I think that what makes the patent claims eligible, all that I'm trying to say is it's more than efficiency.

[00:47:36] Speaker 02:
It's the entirety of claims 211 and 27 of the 322, starting from the very first line of the claim language.

[00:47:45] Speaker 02:
which looks at a function call that is made by an executing program.

[00:47:50] Speaker 02:
It is a specific action that occurs only in a computer system, and this invention changed the way that a computer system responded to that action.

[00:48:02] Speaker 02:
I am focusing on step one, and my only point, I'm not trying to deviate from the arguments we made, is to say there is more than efficiency in the claim.

[00:48:15] Speaker 02:
I would propose to talk about the global revenue issue, unless the court, I'm way over time too, so please let me know if I need to.

[00:48:24] Speaker 02:
It's our time.

[00:48:27] Speaker 02:
Go ahead.

[00:48:27] Speaker 02:
So on the global revenue issue, I want to start with where my friend left off on this issue, which is the idea that all of the infringement occurred before patent issuance.

[00:48:45] Speaker 02:
The record in this case is that Norton made a bash, which is the infringing software.

[00:48:53] Speaker 02:
They made bash version 7.0 through 11.5 on servers in the United States entirely after patent issuance.

[00:49:04] Speaker 02:
They had an early version, bash 6.0, that they made before the patent issued.

[00:49:09] Speaker 02:
But the facts here are different from Brumfield because almost all of the acts of infringement occurred after patent issuance.

[00:49:18] Speaker 02:
And so we have from 2013 to 2023, Norton making and storing on servers in the United States thousands of copies of its bash software that had no purpose but enabling sales to customers abroad.

[00:49:36] Speaker 02:
Norton admits at page five of its blue brief that it kept these copies in the United States.

[00:49:43] Speaker 02:
And this included the foreign-only masters.

[00:49:46] Speaker 02:
And critically, this is the critical fact in this case, Norton continuously used these foreign-only copies on US servers to enable distribution of its software to customers abroad.

[00:50:00] Speaker 02:
The jury heard unrebutted 30b6 testimony on these issues.

[00:50:06] Speaker 02:
And the jury heard that that infringement, the making and using of the invention in the United States, enabled $9.4 billion in revenue from customers abroad.

[00:50:17] Speaker 02:
And for that infringement, the jury awarded only $94 million.

[00:50:22] Speaker 02:
That was consistent with the running royalty that Norton offered to pay

[00:50:27] Speaker 02:
on global revenue when Norton tried to license this invention from Columbia in 2005, and that's appendix site 603-90-93.

[00:50:40] Speaker 02:
Section 284 entitles Columbia to no less than a reasonable royalty for all of Norton's infringement, and the jury was allowed to take this revenue into account.

[00:50:53] Speaker 04:
The counsel on the other side that argues some sort of temporal aspect should be considered here as the foreign sales.

[00:51:04] Speaker 02:
Well, we disagree with that, Judge Raina.

[00:51:06] Speaker 02:
What happened in Brumfield that my friend referred to

[00:51:11] Speaker 02:
the only US copy of the infringing, of the would-be infringing software was made before patent issuance, and it was shipped abroad.

[00:51:20] Speaker 02:
And all subsequent copying and use occurred only abroad.

[00:51:25] Speaker 02:
And this court said that wasn't infringing.

[00:51:29] Speaker 02:
You didn't have any active infringement after the patent issue.

[00:51:33] Speaker 02:
What we have here is

[00:51:36] Speaker 02:
significant infringement, thousands of instances on US servers that were made and constantly used all after the Columbia patents issue.

[00:51:46] Speaker 02:
That makes the temporal point entirely irrelevant.

[00:51:51] Speaker 01:
Yeah, but what's the nexus between that infringement and the foreign sale?

[00:51:56] Speaker 02:
Well, because whenever Norton created a new version of the BAF software, Judge Dyke, they created what they called a geo-locked version.

[00:52:06] Speaker 02:
And this is described at 655-66.

[00:52:11] Speaker 02:
And the geo-locked version that was made and stored in the United States was accessible by and openable by only customers in the intended foreign country.

[00:52:23] Speaker 02:
And that copy was also translated into the relevant foreign language at 655-54.

[00:52:30] Speaker 02:
And all of those foreign-only copies, Judge Dyke,

[00:52:35] Speaker 02:
They were stored on servers in the United States.

[00:52:39] Speaker 02:
And they enabled the distribution to customers abroad.

[00:52:42] Speaker 02:
And they were locked only for a foreign country.

[00:52:45] Speaker 02:
And so the nexus between the domestic making and use of the patented invention and the revenue from the customer abroad could not have been more direct.

[00:52:55] Speaker 00:
And you think that Bromfield then covers you?

[00:52:58] Speaker 02:
Yes.

[00:52:59] Speaker 02:
If proximate cause is satisfied by anything, Judge Prost, it would be satisfied when the infringer specifically intends for a domestic act of infringement to cause a foreign sale, which is what we have here.

[00:53:15] Speaker 01:
OK.

[00:53:15] Speaker 01:
I think we're out of time.

[00:53:16] Speaker 01:
Thank you.

[00:53:17] Speaker 01:
Thank you.

[00:53:18] Speaker 01:
Mr. Loomis, you've got two minutes.

[00:53:33] Speaker 03:
Thank you.

[00:53:34] Speaker 03:
On the argument that modeling function calls is new and non-conventional, I do think that's a new argument, but it's easily dispatched.

[00:53:43] Speaker 03:
Claim one, for example, a host of claims were invalidated by the PTAB and the IPRs.

[00:53:49] Speaker 03:
Claim one are good examples.

[00:53:51] Speaker 03:
Both 115 and 322 have comparing function calls, two models,

[00:53:57] Speaker 03:
Just like the claims that went to jury treatment.

[00:54:01] Speaker 01:
But we're not dealing here with obviousness.

[00:54:04] Speaker 01:
We're dealing with 101.

[00:54:08] Speaker 01:
And what's the evidence here that modeling function calls, one, was the invention, which is the first time I've heard that today.

[00:54:19] Speaker 01:
And second of all, what's the basis for saying that it's convention?

[00:54:25] Speaker 03:
So my basis for saying it's conventional, I never heard it before either.

[00:54:29] Speaker 03:
So I was scrambling to answer that question, your honor.

[00:54:32] Speaker 03:
I think that we know it was in the prior art because we know the IPRs found it to be invalidated in other claims, articulated in the same exact way, shows us it's not new and it's conventional.

[00:54:45] Speaker 03:
There's nothing in the claims that say new types of function calls or new models of function calls.

[00:54:50] Speaker 03:
It just says a model of function calls.

[00:54:53] Speaker 03:
That's an abstraction.

[00:54:54] Speaker 03:
Even if it were new, back to SAP and Symantec, that would go to 102, not 101.

[00:55:00] Speaker 03:
You can have an abstract new concept.

[00:55:04] Speaker 03:
And so at best what they have, I'd suggest, is that if the function calls were new.

[00:55:09] Speaker 03:
We don't think they're new.

[00:55:10] Speaker 03:
We do think they're conventional because of what the patent board found.

[00:55:15] Speaker 03:
Same is true for the selective emulation.

[00:55:18] Speaker 03:
I wanted to actually read from the

[00:55:22] Speaker 03:
claim, the specification itself.

[00:55:25] Speaker 03:
So, counsel said that selective emulation by the emulator was new and non-conventional, but if you go to Appendix 285, beginning at Column 3, Line 28, it's referring to the priority emulators, including STEM, which there's no dispute predates these patents, and it says that STEM, which is described below, quote, and which permits the selective execution

[00:55:52] Speaker 03:
of certain parts or all of a program inside an instruction level emulator using the Valgrind emulator.

[00:56:00] Speaker 03:
So, I'm sorry, I think I tripped over into the next discussion of the Valgrind emulator.

[00:56:05] Speaker 03:
It's describing STEM as doing exactly what he just told you was non-conventional and new in the prior art.

[00:56:11] Speaker 03:
So it's contradicted by their own specification.