[00:00:00] Speaker 02:
We'll hear argument next to number 20-1574, Qualcomm Incorporated versus Intel Corporation.

[00:00:08] Speaker 02:
Ms.

[00:00:08] Speaker 02:
Sweezy.

[00:00:11] Speaker 03:
Thank you, Your Honor.

[00:00:11] Speaker 03:
Good morning.

[00:00:12] Speaker 03:
May it please the court?

[00:00:14] Speaker 03:
This appeal rests on the construction of pull in claim 31.

[00:00:18] Speaker 03:
As the board recognized, our construction of pull is consistent with the intrinsic evidence, and yet the board adopted a different construction.

[00:00:29] Speaker 03:
essentially construing pool as merely meaning receive.

[00:00:34] Speaker 03:
The board gave two reasons from the intrinsic evidence for departing from our construction and that's where I'd like to start.

[00:00:41] Speaker 02:
First... Do you agree that if the data is queued up in the application that it would be sufficient for the thing to operate to say please give me or give the queued up data, is that sufficient?

[00:00:59] Speaker 03:
No, Your Honor.

[00:01:00] Speaker 02:
No?

[00:01:01] Speaker 02:
That would not.

[00:01:03] Speaker 02:
I know your position is that that's not within the claim language.

[00:01:08] Speaker 02:
But wouldn't that make it operable?

[00:01:14] Speaker 03:
It would not.

[00:01:16] Speaker 03:
Maybe I can answer it this way, Your Honor.

[00:01:17] Speaker 03:
That configuration and mechanism or operation would not satisfy pool as required in claim 31 and supported by the specification.

[00:01:28] Speaker 02:
OK, but that's not an answer to my question.

[00:01:31] Speaker 02:
My question is not whether that would come within your construction of full, but whether as a practical matter, the device would be operable if it said transmit the queued up data without identifying where it's stored.

[00:01:54] Speaker 03:
Thank you, Your Honor.

[00:01:55] Speaker 03:
I understand the question.

[00:01:56] Speaker 03:
In other embodiments or configurations, yes, we're not disputing that that is an inoperable way of the device being configured.

[00:02:06] Speaker 03:
And there are, in fact, other claims that would cover that sort of request and response.

[00:02:13] Speaker 02:
That would be operable.

[00:02:16] Speaker 03:
That would be operable.

[00:02:18] Speaker 03:
Yes, we don't dispute that.

[00:02:20] Speaker 03:
But it would not be operable in the context of a pull

[00:02:24] Speaker 03:
limitation of Claim 31 requires.

[00:02:26] Speaker 03:
It would be outside the scope of Claim 31.

[00:02:28] Speaker 02:
And why is that?

[00:02:32] Speaker 02:
Is that because of the references in the specification?

[00:02:38] Speaker 03:
It's both because of the claim language of 31 and the references in the specification.

[00:02:43] Speaker 03:
In Claim 31, and I think there are a couple of points to emphasize here, the claim language itself, which the board I think gave short shrift to,

[00:02:54] Speaker 03:
indicates and demonstrates that a pull is a function that is requiring the modem processor being in control and getting the data from the application processor.

[00:03:05] Speaker 03:
And I emphasize several points.

[00:03:08] Speaker 03:
First, of course, pull is used as a verb.

[00:03:10] Speaker 03:
It is an active action in Claim 31.

[00:03:14] Speaker 03:
The second point would be looking at the surrounding language of pull.

[00:03:18] Speaker 03:
So we're not just defining pull in isolation or in the abstract, but what is

[00:03:22] Speaker 03:
doing the polling and what is being pulled.

[00:03:25] Speaker 03:
In Claim 31, there are aspects that the application processor is doing, and specifically, it is configured to hold data, whereas the modem processor is configured to pull data.

[00:03:39] Speaker 03:
So that contrast reinforces that pull is something that the modem processor is controlling.

[00:03:45] Speaker 01:
And specifically... To follow up on Judge Dyck's question, though,

[00:03:49] Speaker 01:
Suppose that you have the modem processor demanding that all data that is currently being stored up in the application processor be shipped to the modem processor without reference to where that data may have come from.

[00:04:13] Speaker 01:
Why wouldn't that be a form of pull?

[00:04:17] Speaker 03:
Judge Bryson, I can give you a couple of responses.

[00:04:20] Speaker 03:
And one is just, again, going back to that this is an active verb that the modem processor is engaging in.

[00:04:26] Speaker 03:
This isn't a full report.

[00:04:27] Speaker 01:
Right, but the modem processor, in my example, is actively pulling the data.

[00:04:32] Speaker 01:
That is to say it is demanding that the data be sent, right?

[00:04:37] Speaker 01:
So I don't see that that gets you anywhere.

[00:04:42] Speaker 03:
Let me give you a couple of responses, one based on the claims, and then a second on the specifications.

[00:04:47] Speaker 03:
So in the claims itself, there are other claims that do exactly, Judge Bryson, as you posed.

[00:04:53] Speaker 03:
For instance, asking or requesting information, and then that information is sent.

[00:04:59] Speaker 03:
And it really depends on which processor is doing the transferring of the data.

[00:05:04] Speaker 03:
Again, as a high-level context, the 490 patent is talking about data transfer across an interconnectivity bus

[00:05:11] Speaker 03:
And on the pull question is which processor is in control of that data transfer?

[00:05:18] Speaker 03:
Not simply asking for it and then the application processor is in control of sending it.

[00:05:23] Speaker 03:
But in the pull limitation of claim 31, it is the modem processor that is pulling the data.

[00:05:30] Speaker 03:
And claim 10 is one example that it is, there's a request and an invitation for

[00:05:37] Speaker 03:
the application processor to send something.

[00:05:40] Speaker 03:
And that would align with your hypothetical, Your Honor, but it would not align with the meaning of pull, the separate term used in Claim 31.

[00:05:48] Speaker 03:
And we know this because the specification, every time the specification discusses pull, and it does it in detail in two particular passages of the specification, it's always using pull in the context of the modem processor knowing where to get the data from.

[00:06:06] Speaker 03:
And the board understood this.

[00:06:08] Speaker 03:
I think this is an important point to underscore.

[00:06:10] Speaker 03:
The board repeatedly credited our understanding of the specification passages that run from column 9 through 10 and column 15 through 16 as the modem processor knowing, based on write pointers or other indicators, where to go and get that data.

[00:06:28] Speaker 03:
It needs to know where in the buffers or with the cues or wherever the application processor might be holding the data

[00:06:36] Speaker 03:
where the motor processor can go and get it and carry out its functionality of pulling the data across the bus back into the motor processor.

[00:06:45] Speaker 03:
So at Appendix 22, at Appendix 23 to 24, at Appendix 25, and at Appendix 27, the board is repeatedly discussing these passages at Columns 9 through 10 and 15 through 16 and crediting that understanding of those disclosures

[00:07:05] Speaker 03:
as pull needing to know the location of where the data is based on, for instance, the disclosures of the right pointers, the buffer ring, knowing where the application processor has put its data for the modem processor to pull it in the context of claim 31 and its pull requirement.

[00:07:25] Speaker 02:
But the problem that you have, it seems to me, is that the language of the claim doesn't say that.

[00:07:31] Speaker 02:
And even if one were to agree that there are some examples in the specification where the actual storage location is known, why is it that that should suggest that that's what the overall requirement is?

[00:07:49] Speaker 02:
And there's expert testimony that you don't need to know

[00:07:59] Speaker 02:
storage location and you have yourself agreed that you don't need to know the storage location in order to pull the data.

[00:08:07] Speaker 03:
And Judge Seig, let me give you a few responses and try to walk through the points you've raised.

[00:08:12] Speaker 03:
So we do believe that the claim language itself is requiring this both because of the word pull and in the context of it.

[00:08:19] Speaker 03:
The point you raised about the examples, of course there is always this fine line that needs to be balanced between importing limitations from the specification

[00:08:28] Speaker 03:
and using the specification to inform the meaning of the claim.

[00:08:32] Speaker 03:
And so while certainly a patentee doesn't have to spell out everything in the claim language, that's precisely what the detailed discussions in the patent specification are.

[00:08:42] Speaker 03:
And where pull is repeatedly and consistently used in the patent specification, as the board acknowledged, is always in the context of knowing the location.

[00:08:52] Speaker 03:
So it doesn't necessarily have to be particularly via a write pointer.

[00:08:56] Speaker 03:
Those are different implementations.

[00:08:58] Speaker 03:
But it's always knowing where to get the data in order for the modem processor to be in control of it.

[00:09:05] Speaker 03:
So for instance, at column 9, line 41, the patent specification specifically talks about this embodiment being where the modem processor is controlling the data, controlling the transfer of the data.

[00:09:23] Speaker 03:
There is, I think again, and this goes back to the board's mistake, while it credited the actual understanding of the particular discussions of pull at columns 9 through 10 and 15 through 16, it looked at two other portions of the specification to enlarge that meaning of pull.

[00:09:41] Speaker 03:
And those do not justify the board's departure from what the specification clearly explained a pull is, what pulling is.

[00:09:50] Speaker 03:
And I think those two reasons that the board relied on really exposed the error it made.

[00:09:54] Speaker 03:
So it first relied on the most generic of statements in most patents, the concluding passage that says, as a savings matter, other embodiments are covered.

[00:10:06] Speaker 03:
That passage doesn't use pull.

[00:10:08] Speaker 03:
And this court's case law has never allowed such a generic boilerplate passage

[00:10:13] Speaker 03:
to override the specific use of a term in the specification as the specification here does at columns 9 through 10 and 15 through 16.

[00:10:24] Speaker 03:
And certainly that would upend the ordinary claim construction inquiry of how the patentee chooses to disclose and describe a term that is expressly used in a claim.

[00:10:34] Speaker 03:
The second portion of the specification that the board relied on was one other single sentence.

[00:10:41] Speaker 03:
And this is a sentence in column 16 that the board just simply misread.

[00:10:47] Speaker 03:
It's a phrase that says there are other techniques, but of course we have to look at what the other techniques are referencing.

[00:10:54] Speaker 03:
And they're not referencing techniques for pulling.

[00:10:57] Speaker 03:
While pull is used in that sentence, the other techniques are describing ways to indicate data is available, a precursor step to the pulling.

[00:11:07] Speaker 03:
And so other techniques follow along the string of

[00:11:11] Speaker 03:
disclosures about setting doorbells and polling.

[00:11:15] Speaker 03:
And in column 16, I'll just finish this point if I may, Your Honor.

[00:11:19] Speaker 02:
Go ahead.

[00:11:22] Speaker 03:
Thank you.

[00:11:22] Speaker 03:
In column 16, right above that sentence as an issue, which is at 34 to 37, above that in lines 12 through 18,

[00:11:33] Speaker 03:
And again, at 18 through 30, the patent is making clear that those doorbell registers and polling are ways to indicate data is available, specifically line 27 at column 16.

[00:11:44] Speaker 03:
It's not informing the meaning of pull, unlike the other passages in the specification that are directly on point.

[00:11:51] Speaker 01:
Just a quick question.

[00:11:54] Speaker 01:
You did not move to amend claims 31 for the board, is that right?

[00:12:02] Speaker 03:
That's correct, Your Honor, and of course, that's not, you know, that's not a substitute for the ordinary and necessary claim construction inquiry, which is what the straightforward issue is in this appeal.

[00:12:16] Speaker 03:
I'll preserve the remainder of my time for rebuttal.

[00:12:18] Speaker 03:
Thank you.

[00:12:21] Speaker 02:
Okay, Mr. Saunders.

[00:12:24] Speaker 00:
Good morning, and may it please the Court.

[00:12:26] Speaker 00:
I just want to start by underscoring two points that came up.

[00:12:30] Speaker 00:
The idea that our construction and the board's construction of pull is passive receipt is a red herring.

[00:12:37] Speaker 00:
The board construed it to mean receiving data in response to a request for a data.

[00:12:43] Speaker 00:
So there is an action there being done by the modem processor.

[00:12:47] Speaker 00:
And then on the question about operability, the board credited the expert testimony from Dr. Lin.

[00:12:54] Speaker 00:
This is pages 23 to 24 of the appendix.

[00:12:57] Speaker 00:
that there was at least one other known interconnectivity bus that doesn't require an address or location to perform a poll.

[00:13:07] Speaker 00:
And his testimony talked about the example that came up of saying to the application processor, essentially, give me everything that you have queued.

[00:13:18] Speaker 00:
And, you know, turning to the language of the claims, there's no mention in them of accessing data based on a specified location

[00:13:27] Speaker 00:
no mention of right pointers, no context that indicates that a location is necessary.

[00:13:34] Speaker 00:
And when you put the claim 31 in the context of the patent as a whole and the other claims, it's clear that the thrust of the patent here is about saving power by reducing the number of transitions from low power to high power.

[00:13:57] Speaker 00:
says the problem in the prior art is that's not being controlled.

[00:14:00] Speaker 00:
Now our prior art shows otherwise, but what it's trying to do is group this.

[00:14:04] Speaker 00:
The data going one way, the data going the other way is being grouped during a single act of state.

[00:14:11] Speaker 00:
And so what you see in all the embodiments and throughout the claims are a series of iterations on this same idea that are focused on the timing

[00:14:22] Speaker 00:
and the trigger for the transfer of data.

[00:14:25] Speaker 00:
So claim one is a push after receipt.

[00:14:29] Speaker 00:
Claim five is a push after the expiration of the uplink timer.

[00:14:33] Speaker 00:
Sometimes you have a byte counter triggering.

[00:14:36] Speaker 00:
Sometimes you have a packet counter.

[00:14:38] Speaker 00:
Sometimes you have an express requirement that this be within a single wake state.

[00:14:43] Speaker 00:
And so the board's construction of pull to mean receiving data in response to a request

[00:14:50] Speaker 00:
fits very logically among all of these iterations that are focused on timing.

[00:14:57] Speaker 00:
And turning to the language and the specification at column 16 that the parties have disputed, which is the language about data can be pulled or pushed across the interconnectivity bus based on polling with an O, setting doorbell registers, or other technique.

[00:15:17] Speaker 00:
Qualcomm has tried to distinguish that language by saying this is about the initiation of the transfer as opposed to the execution of the transfer.

[00:15:28] Speaker 00:
But the fact that the specification is talking about the steps that Qualcomm describes as an initiation reinforces the distinction between the push and pull, reinforces that distinction is about whether the processor is receiving data in response to a request

[00:15:47] Speaker 00:
or whether the application processor is sending it over when it wants to send it.

[00:15:52] Speaker 00:
And so you have that broadening language.

[00:15:55] Speaker 00:
And overall, the attempt that's being made here is really just a classic example to read limitations from the specification into the claim.

[00:16:06] Speaker 00:
The figure four embodiment that goes on columns nine through 10

[00:16:13] Speaker 00:
is expressly exemplary.

[00:16:15] Speaker 00:
It begins by saying on column nine, figure four illustrates an exemplary power saving process.

[00:16:23] Speaker 00:
And in the paragraph that is discussing the use of right pointers, although there it's right pointers to show where the data is going to go from the modem processor to the application processor, it begins by saying, for example,

[00:16:37] Speaker 00:
And it's talking about a particular standard for the interconnectivity bus, the PCIe standard, that Dr. Baker, Qualcomm's expert, admitted at 2330 of the appendix is not required by the patent and that the board memorialized the concession by Qualcomm and its opinion to that effect on page 21.

[00:17:00] Speaker 00:
And for the figure 12 embodiment, you know, it is immediately followed by this broad or other technique

[00:17:07] Speaker 00:
language and all of this is in a section that is dedicated to descriptions of particular embodiments that has broad language at the beginning and the end saying that the intention is not to limit these embodiments.

[00:17:24] Speaker 00:
You know the board read this all also against the backdrop of the extrinsic evidence where it's fact findings are reviewed only for substantial evidence

[00:17:35] Speaker 00:
and said, you know, we have dictionary definitions here that are all consistent with receiving information after requesting it.

[00:17:45] Speaker 00:
The board, you know, Qualcomm has pointed to some language, some examples within those definitions involving an internet context and the board made a factual finding

[00:17:56] Speaker 00:
page 23 of its opinion, that the examples weren't limited to that context and that's not the relevant context for the patent here.

[00:18:04] Speaker 00:
The board, as I said, credited the testimony of our expert Dr. Lin about other known techniques that would not require address information.

[00:18:14] Speaker 00:
And then it addressed the issue Qualcomm had raised about testimony from Apple's experts in the ITC.

[00:18:24] Speaker 00:
and made factual findings and said it wasn't persuaded that there was a conflict with that testimony.

[00:18:30] Speaker 00:
It put it in the context of the other proceeding.

[00:18:33] Speaker 00:
It pointed out some of the limited information it had before it about that testimony.

[00:18:38] Speaker 00:
And if the panel has further questions on that, I can address the specific statements.

[00:18:44] Speaker 00:
But, you know, all of this from start to finish is claim language that isn't restrictive

[00:18:51] Speaker 00:
specification that is just exemplary embodiments without language indicating an intent to restrict the meaning and all supported by factual findings on the extrinsic evidence and the dictionary definitions that back the broad understanding of the term poll in this context.

[00:19:13] Speaker 00:
I'm happy to answer any questions the panel may have.

[00:19:17] Speaker 01:
Mr. Sanders, this is Judge Bryson.

[00:19:20] Speaker 01:
curious to follow up on my question to Ms.

[00:19:25] Speaker 01:
Sweezy.

[00:19:27] Speaker 01:
What significance, if any, should we give to the fact that when this issue arose as a dispute as to the breadth of claim 31, there was no motion to amend?

[00:19:43] Speaker 00:
I think the significance is it goes hand in hand with the board's application of the BRI standard.

[00:19:50] Speaker 00:
which is part of the, and this is the case that still is governed by that standard, and part of the justification there is we do that and we'll read this and really resolve that ambiguity in favor of press because the patent owner has that opportunity.

[00:20:08] Speaker 00:
Now we think that, as we argued in the alternative,

[00:20:13] Speaker 00:
even if the patent owner amended the claim here to adopt its definition of pull, it still is going to run into problems with our prior art.

[00:20:22] Speaker 00:
But the fact that there isn't that attempt to amend just reinforces the use of the BRI standard here and resolving any of this ambiguity in our favor.

[00:20:34] Speaker 00:
And that's really what the board did.

[00:20:36] Speaker 00:
Even before the board got to the extrinsic evidence, it looked at the intrinsic evidence and said,

[00:20:43] Speaker 00:
Both, you know, both constructions could be reasonable under this.

[00:20:48] Speaker 00:
And so under the BRI standard, that favors Intel without going to the extrinsic evidence.

[00:20:53] Speaker 00:
And then the extrinsic evidence was on top of that and was further support for our position.

[00:21:00] Speaker 02:
But there would be reasons that a patentee wouldn't want to amend, right?

[00:21:04] Speaker 02:
Because that would create potentially intervening rights.

[00:21:10] Speaker 00:
Sure, there are reasons that they will want to argue the claim construction and wouldn't want to amend.

[00:21:17] Speaker 01:
In light of that, we have never said, have we, that the failure to seek to amend is somehow an indication of weakness in the position of the patent owner because of the point that Judge Dyck has made.

[00:21:36] Speaker 01:
We frequently see the statement that in the petitioner's briefs that, well, there was no effort to amend.

[00:21:46] Speaker 01:
And I've always wondered, well, you know, should we attach any significance to that or in light of the fact that the patent owner would be at least potentially confined by any amendment, does that have any weight at all?

[00:22:05] Speaker 00:
Well, I will remember this convert.

[00:22:08] Speaker 01:
Yeah, go ahead.

[00:22:09] Speaker 00:
We'll remember this because in the parallel IPRs, Intel's where it's appealing Intel standing is being challenged.

[00:22:17] Speaker 00:
So I would like to think that they were

[00:22:20] Speaker 00:
We're thinking of whether there would be further suits against Intel, but that'll be a question for another day and another appeal.

[00:22:27] Speaker 00:
And we by no means are urging the panel to rely on this, the fact of the lack of amendment.

[00:22:33] Speaker 00:
As I said, I think it goes hand in hand with broadest reasonable interpretation, but as we argued, you know, under the Phillips standard, all the indications are pointing in our direction on these claims as not being limited.

[00:22:49] Speaker 00:
Okay.

[00:22:51] Speaker 02:
Okay.

[00:22:52] Speaker 02:
Ms.

[00:22:52] Speaker 02:
Sweezy.

[00:22:54] Speaker 03:
Thank you, Your Honor.

[00:22:55] Speaker 03:
I'm going to try to run quickly through about five points, and I think I'll just start where we ended, which is the question, as everyone agrees, is claim construction not consequences or choices to be made based on amending the claims?

[00:23:07] Speaker 03:
And it is the Phillips Inquiry.

[00:23:08] Speaker 03:
So first, there was a point that polling could be read as receiving.

[00:23:13] Speaker 03:
I would just direct the Court to other claims, such as Claims 1, 29, and 10,

[00:23:19] Speaker 03:
which are sufficient for receiving, but pull demands something more as used in Claim 31.

[00:23:25] Speaker 03:
The second point was that there were references to other buses that may not require pull as Claim 31 does.

[00:23:32] Speaker 03:
That is a reference to Exhibit 1021.

[00:23:35] Speaker 03:
The board explicitly said it was not relying on that exhibit.

[00:23:39] Speaker 03:
This is at Appendix 27 to construe pull.

[00:23:42] Speaker 03:
In fact, that reference never describes a pull.

[00:23:45] Speaker 03:
It's simply a different way of operating, but the question here is not what any and all buses might do, but what a pull requires.

[00:23:54] Speaker 03:
Again, it's expressly required by the claim.

[00:23:56] Speaker 03:
Third, there was a discussion about embodiments.

[00:23:59] Speaker 03:
I'd like to emphasize, of course, there are 31 claims in this patent, and while other claims might be sufficient for focusing on timing or triggers, claim 31 has this additional element expressly recited of the pull function.

[00:24:15] Speaker 03:
Fourth, there was a reference to column nine that begins one sentence by saying, for example, this court's case law is clear that when you're evaluating what might be an example, you have to look at the context.

[00:24:28] Speaker 03:
And the person of ordinary skill in the art reading that passage would recognize it is an example of the particular bus mentioned there, the PCIe bus.

[00:24:36] Speaker 03:
It's not pulling in everything in that column or the patent.

[00:24:41] Speaker 03:
particularly when there is no other use of pull in the patent than one that requires a specified location as the board recognized in crediting our reading of the patent specification.

[00:24:52] Speaker 03:
And my final point is the reference to extrinsic evidence.

[00:24:55] Speaker 03:
The board itself at Appendix 22 recognized and we agree that the intrinsic evidence answers this.

[00:25:02] Speaker 03:
It is sufficient.

[00:25:03] Speaker 03:
The extrinsic evidence could be informative only to the extent it's relying on

[00:25:08] Speaker 03:
context similar to the 490 patents.

[00:25:11] Speaker 03:
The extrinsic evidence, to the extent it has any relevance, is looking at addresses.

[00:25:17] Speaker 03:
But, of course, we always have to read it in light of the 490 patent, which is seeking a construction of pull based on one processor, grabbing data over high data transfers over the interconnectivity bus, using ring buffers, and needing to know where to grab that data, where the location is.

[00:25:35] Speaker 03:
That is the meaning consistent with the detailed discussions of the patent specification.

[00:25:40] Speaker 03:
And the board's reliance on two passages that just frankly do not address POLL and cannot be used to enlarge out the meaning of POLL, whether it's under Phillips or the broadest reasonable construction, is grounds for reversing the board and at minimum remanding but in our view because POLL does not have sufficient evidence reversing the board's judgment.

[00:26:02] Speaker 03:
Thank you, Your Honor.

[00:26:04] Speaker 02:
Thank you, Ms.

[00:26:04] Speaker 02:
Sweezy.

[00:26:05] Speaker 02:
Thank you, Mr. Saunders.

[00:26:06] Speaker 02:
The case is submitted.