[00:00:00] Speaker 03:
Third argument for appeal this morning is docket number 25-1058, sentient censors versus Xilinx.

[00:00:09] Speaker 02:
Mr. Cochran.

[00:00:12] Speaker 02:
Thank you, Your Honors. May it please the Court, my name is Andrew Cochran. I represent Sentient Censors, LLC.

[00:00:19] Speaker 02:
We are here today because the Patent Trial and Appeal Board committed reversible error on its final written decision. In validating claims 1 through 5, 10 through 15, and 17 through 20 of the 177 patent, Mr. Blimble's innovative instrument controller patent. We have three issues that we'd like to discuss today. The first is the claim construction issue on the claim term independent processes. The second is the obviousness finding of the references like and what we call decorti.

[00:00:52] Speaker 02:
The first two issues are tied together. And then the third issue, which is a separate issue, is the obviousness finding of the power converter limitation by like and the reference steel.

[00:01:10] Speaker 02:
If there's no preference, I would like to start with the independent processes limitations.

[00:01:18] Speaker 02:
So the board construed independent processes to mean merely processes that run separately from one another. The construction was error because it replaced independent with separate, effectively rewriting the claim and stripping from it the very concept of independence.

[00:01:35] Speaker 00:
Do you contend the independent processes have to be independent from each other or independent from the processor? Yes. Your view of the right construction?

[00:01:44] Speaker 02:
Our position is actually both. It has to be independent processes or processes running in the FPGA that are independent of each other as well as independent of what's going on in the process.

[00:01:59] Speaker 03:
Didn't the board find both for both likes and accordion?

[00:02:05] Speaker 02:
So find both types of independence. So the board recognized that the specification supported both types of independence. Right.

[00:02:17] Speaker 03:
The claim construction ultimately was either type would satisfy independent processes.

[00:02:23] Speaker 02:
Correct.

[00:02:23] Speaker 03:
But then in its analysis of evaluating what likes would be with an FPGA or what Decorti discloses with its FPGA, didn't the board find both types? in both instances?

[00:02:40] Speaker 02:
Well, the board's analysis of what it contains for both constructions rests on what the board found was a waiver by our part on the construction of computational and without computational assistance.

[00:02:58] Speaker 03:
It rejected the idea that the process has to be a computational process, right?

[00:03:04] Speaker 02:
It did reject that idea. And the computational process really was meant by us to get to the concept of what independent really means within the context of this particular patent and its prosecution history.

[00:03:19] Speaker 03:
So just to really boil it down, isn't your dispute about whether the process has to be computational or not? No.

[00:03:27] Speaker 02:
I would not say that that is our boiled down dispute. Our boiled down dispute, I would say, is getting to what the actual meaning of independent would be in the context of this patent.

[00:03:40] Speaker 03:
Okay. Well, I understand that you have a computational argument. What's the other argument or other arguments aside from it needs to have the notion of computational process?

[00:03:55] Speaker 02:
Well, the other argument is that the – I'm actually not quite sure what you're getting at with other arguments.

[00:04:01] Speaker 03:
Well, what is the real defect in the claim construction?

[00:04:06] Speaker 02:
The defect in the claim construction that the board adopted is replacing independent, which is what's in the claim, with separate.

[00:04:14] Speaker 03:
Right, but that doesn't help me quite see what the problem is. I mean, I think you – have told us you want the claim construction to include this notion of computation, okay? What else do you need to be in the claim construction other than computation?

[00:04:33] Speaker 02:
Well, so let me get to it, Your Honor.

[00:04:46] Speaker 02:
So, again, computational was really – It is supported by the specification as we laid out in our brief, but really the thought behind the computational processes.

[00:05:01] Speaker 03:
You're getting into computational, and of course you're allowed to do that, but I'm just trying to get a top-level view of what are the things you want in the claim construction that are currently not in the claim construction. Number one is computation. Sure. Now I'm hunting for a number two because I've been telling you for the past couple of minutes I can't follow your brief as having and requesting something more above and beyond this computational notion. So what else?

[00:05:26] Speaker 02:
What I believe we need to be in the construction of independent is the concept of both independence in terms of processes occurring in the FPGA and processes occurring in the processor in parallel with each other.

[00:05:40] Speaker 03:
Right.

[00:05:41] Speaker 02:
So that, in a nutshell, is what we need independent processes within the context of the claim to encompass.

[00:05:50] Speaker 00:
But didn't the PTAB make findings on both of those?

[00:05:54] Speaker 02:
The findings on what it contends is making findings out of both processes. There's basically three processes that I think that what we're looking for if we're going to boil down into – bare bones here, is three processes for independent processes in the claim to be truly independent of each other. The Board did not actually make a finding on if all three of those were present in the prior R under the construction.

[00:06:27] Speaker 02:
It made a finding that either could be met, but it did not actually make the finding that all were met.

[00:06:36] Speaker 00:
Could you point me to a specific page of the PTAB decision that you think kind of encapsulates the errors that you're thinking the Board made here? I want to just be able to see it.

[00:06:49] Speaker 02:
Sure. Give me a second to find it, Your Honor. Okay.

[00:07:18] Speaker 00:
And maybe I'll put this on your radar to think about after you answer the question I just posed to you. But are you contending there's a difference between finding that either is met and finding that both were met? Is that what your argument is? I'm trying to get an understanding of your argument.

[00:07:34] Speaker 02:
Sure, Your Honor. I do believe that there is a big distinction within the concept of this patent on whether either is met and whether both are met.

[00:07:43] Speaker 00:
So what is your argument? Just tell me what your argument is that in terms of the error that you're thinking the board made here.

[00:07:50] Speaker 02:
So our argument is that the instrument controller of this patent is this multifunctional processing unit that is able to process many different things all at the same time and all independently of each other. And the processor is doing its thing and what what processes might be going on in the processor at the time. And it has certain control over what it's doing. And then the FPGA is sitting somewhere else.

[00:08:23] Speaker 02:
And it is also configured to do independent processes, plural processes, in parallel with whatever is going on, with its own control and its own functional blocks.

[00:08:39] Speaker 02:
The parties both point to Figure 2, for example, of the patent as support for its own construction.

[00:08:47] Speaker 03:
So the board found that it could be either type of independent process that would meet the limitation of independent process's you think instead of either type, it has to be both types.

[00:08:59] Speaker 02:
That is correct, Your Honor.

[00:09:01] Speaker 03:
And that the processes have to be computational.

[00:09:04] Speaker 02:
We believe that the processes are computational. Now, I don't think it's necessarily critical to the construction that they're computational, because I do believe that computational was added into our construction, which has been consistent throughout these entire proceedings. was meant to get to the concept of what independence really means because of the further analysis of computational without computational assistance. That's what we were trying to get at, what really independent means in this context.

[00:09:36] Speaker 02:
So computational was added to the claim construction really to get at what the meaning of independent should be.

[00:09:43] Speaker 00:
But even setting aside your claim construction argument, you'd have to also win on your substantial...

[00:09:49] Speaker 00:
evidence challenge, correct?

[00:09:51] Speaker 02:
So, is he putting aside the claim construction?

[00:09:53] Speaker 00:
Yeah, I'm going to move on to another thing I'd like to talk about, which is whether or not there's sufficient substantial evidence support.

[00:10:00] Speaker 02:
So, if we're moving on to obviousness instead of claim construction, we believe that the board's claim construction is erroneous so that the obviousness finding is also erroneous. If we are given evidence a proper construction of what independent really means, then it's clear to us that the references used for the obviousness finding do not disclose these independent processes. So it really rests on what the claim construction is and whether it is correct or incorrect.

[00:10:34] Speaker 02:
My colleagues over here, Xilinx, believe that it should just be separate. Essentially, if you find that there is two processes that that are separate from each other, then that suddenly teaches independent processes. But there should be further analyses there. And that's where we have gotten into what Xilinx has continually abandoned and what the board agreed that we abandoned is the clock and control arguments. And like in Decorti, both references, there is a- We have never contended that our definition of independent, our claim construction of independent necessarily requires clocks and control.

[00:11:23] Speaker 03:
Didn't the board find, I can't remember, but that Decorti has its own internal clock?

[00:11:29] Speaker 02:
It has a single clock in the processor that operates both the processor and the FPGA. And so it cannot run. So a process cannot run without a clock. So with the single clock, there must be one process running at one time. So right away, we don't get parallel processing, which is another claim element.

[00:11:51] Speaker 03:
But we also know that just the single clock would control... But the Cordy's FPGA has an internal system clock, and therefore runs processes independent of the processor. That's A57. So it's not just the Cordy's overall...

[00:12:09] Speaker 03:
MCM, it's the FPGA itself has a clock.

[00:12:13] Speaker 02:
So the FPGA in Decorti, we contend, does not have a clock.

[00:12:19] Speaker 02:
It is dependent upon the clock that is in the processor, which is why Decorti says that the FPGA is always intended to be subservient in a coprocessor to the multiprocessing capability. Your rebuttal, do you want to save it?

[00:12:35] Speaker 02:
Yes, Your Honor. I will save my rebuttal. Okay.

[00:12:49] Speaker 01:
Good afternoon, Your Honors. My name is Jeffrey Schneidman, and I represent the appellees, Xilinx, AMD, and ATI below.

[00:12:57] Speaker 01:
Let me just go right to the questions. Judge Chen, you asked, did the board find that both types of independent processes were taught by like and ducorti? The answer is yes. If you look at Appendix 23, let's see, excuse me, Appendix 25, ultimately the petitioner's arguments regarding independent processes applies to both interpretations, i.e. the relied upon FPGA processes both run separately from each other and separately from those on the processor.

[00:13:28] Speaker 01:
And then the details of that analysis are at Appendix 52 to 53 for Decorti.

[00:13:33] Speaker 03:
Wait, wait, slow down. Can you say that again?

[00:13:36] Speaker 01:
Yes. So Appendix 25.

[00:13:38] Speaker 03:
Right, no.

[00:13:40] UNKNOWN:
52, 53.

[00:13:40] Speaker 03:
52, 53. Okay.

[00:13:43] Speaker 03:
What are the quotes you want to highlight for us to make it clear that the Board, in fact, did make separate findings for that type 1 and type 2 of independent processes are being performed in Decority's system.

[00:14:01] Speaker 01:
That's correct, Your Honor. And similarly for like at appendixes 42.

[00:14:04] Speaker 03:
Where are the quotes?

[00:14:06] Speaker 01:
Oh, the quotes.

[00:14:07] Speaker 03:
Get some quotable quotes on the record.

[00:14:09] Speaker 01:
Sure.

[00:14:16] Speaker 01:
So, for example, at 52...

[00:14:21] Speaker 03:
How far down?

[00:14:21] Speaker 01:
Most of the way down. Petitioner argues persuasively that traffic routing is an independent process because it does not depend on other processes run on the DSP processor, i.e., it runs separately from the processor.

[00:14:35] Speaker 01:
Dr. Shanfield also relies on DeCourty's disclosure of FPGA preprocessing incoming data as teaching independent processing.

[00:14:42] Speaker 03:
Are these processes that are being performed by the FPGA?

[00:14:46] Speaker 01:
These are processes being performed by the FPGA.

[00:14:48] Speaker 03:
And here the Board is saying those processes are independent of the processor's processes.

[00:14:56] Speaker 01:
It's independent both with respect to what's running on the processor as well as what's running within the FPGA.

[00:15:04] Speaker 01:
If you see the top of Appendix 53, Your Honor, it says it runs separately from the processor. And then the next paragraph is petitioner further submits additional processes run on the FPGA run independently in parallel with the processor. These include and it gives the following.

[00:15:23] Speaker 03:
Okay, that's petitioner's arguments.

[00:15:25] Speaker 03:
Where does the Board say, yep, we agree with the petitioner?

[00:15:30] Speaker 01:
Yeah, so the Board credits throughout this discussion I'm sorry, Your Honor. I would say throughout this discussion, here it is crediting the findings.

[00:16:08] Speaker 01:
I'm flipping through the Board's findings.

[00:16:10] Speaker 03:
Could it be somewhere in the middle of 55? 55?

[00:16:14] Speaker 01:
Thank you, Your Honor.

[00:16:15] Speaker 03:
Let me just take a look. The last two sentences above that one lower paragraph, here petitioners provided sufficient evidence of FPGAs and CPUs running as independent execution units and of FPGAs partitioning tasks to run separately as discussed above?

[00:16:34] Speaker 01:
Yes, Your Honor.

[00:16:35] Speaker 03:
That is one example of the finding.

[00:16:36] Speaker 01:
Thank you.

[00:16:40] Speaker 01:
The board similarly found for like as well that like similarly taught, and that's Appendix 42 to 45.

[00:16:47] Speaker 01:
But the point is, one reference teaches that the claim elements are met.

[00:16:54] Speaker 01:
The board also did find, Judge Chen, to your question, that the QWERTY did have its own clock. I think I can give you a better citation, I'm sorry, at Appendix 57, which is...

[00:17:10] Speaker 01:
The Xilinx 4D10, the FPGA used in the QWERTY, includes an internal system clock that allows processes to run on their own using its internal clock system.

[00:17:20] Speaker 01:
If the court has other questions, I'm happy to entertain. But other than that, I would.

[00:17:25] Speaker 03:
Okay. Thank you, Your Honor. Mr. Schneidman.

[00:17:34] Speaker 03:
Mr. Cochran, you still have some time left.

[00:17:40] Speaker 02:
Thank you. So I'll address what was just discussed on A25 and actually more specifically A52 and 53. These findings all rely on what we contend is the erroneous claim construction of independent being just simply separate, which we think cannot be. So the fact that

[00:18:09] Speaker 03:
Let's assume for the moment that these pages that were cited by opposing counsel demonstrate that the board made a fact finding that each of these two references teaches both types of independent processes. So the board didn't just apply one piece of that two-part claim construction, either this type of independent or that type of independent. It found both types of independents.

[00:18:40] Speaker 03:
Okay, so assuming that's true, now what's the problem?

[00:18:45] Speaker 02:
The problem is that both types of independence are not there, Your Honor, according to what our arguments are.

[00:18:53] Speaker 02:
Especially here, I say Appendix 52, when the Board credits Dr. Shanfield on the bus arbitration as a traffic cop, and it contends that the traffic cop, the bus arbitration, is an independent process. But the independent process, according to our understanding, or the traffic cop, I'm sorry, according to our understanding, is essentially just that, a traffic cop. It is not an independent process. It's just controlling what others... Is there a lack of substantial evidence support here? I'm sorry?

[00:19:23] Speaker 00:
Why is there a lack of substantial evidence support here?

[00:19:25] Speaker 02:
That's what I think your fundamental... Ultimately, it rests on the incorrect claim construction that we contend. That's, I think, really... We believe the claim construction is wrong and

[00:19:36] Speaker 03:
Therefore, the findings of fact… I assume for the moment here that the Corites FPGA is running a process of pre-processing incoming data and then a completely unrelated process to the process of pre-processing incoming data called this traffic routing thing.

[00:20:01] Speaker 03:
that those two processes are separate and unrelated to each other, not interrelated at all.

[00:20:09] Speaker 03:
What is wrong with saying that those are independent processes of each other?

[00:20:14] Speaker 02:
I would say according to our position, the two processes might be, there might be a process in the processor and one in the FPGA. Our position is there should be a third according to the claim language, the plural processes.

[00:20:28] Speaker 03:
But then the findings go further and say, okay, and then there's another process being run by the processor, the microprocessor, the CPU that's different from what's going on in the FPGA. And that process going on in the processor also is not interrelated in any shape or form with the two processes being run on the FPGA, which also happen to be independent of each other. So what's wrong with that?

[00:20:57] Speaker 02:
So I think that gets into what our arguments were upon clocks and control.

[00:21:01] Speaker 03:
Okay.

[00:21:03] Speaker 02:
And you have to do a further analysis of if it's truly independent according to the claim language. If there's control over a process, then it's not necessarily independent of other processes.

[00:21:19] Speaker 03:
You're out of time. Do you have any final thoughts?

[00:21:23] Speaker 03:
No, Your Honor, I'm out of time. Okay. Thank you very much. Appreciate it. Case is submitted.