UNITED STATES PATENT AND TRADEMARK OFFICE
`____________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________
`
`APPLE INC.,
`Petitioner,
`
`v.
`
`CALIFORNIA INSTITUTE OF TECHNOLOGY,
`Patent Owner.
`____________
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`____________
`
`Record of Oral Hearing
` Held: May 8, 2018
`____________
`
`
`
`Before KEN B. BARRETT, TREVOR M. JEFFERSON and
`JOHN A. HUDALLA, Administrative Patent Judges.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`____________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________
`
`APPLE INC.,
`Petitioner,
`
`v.
`
`CALIFORNIA INSTITUTE OF TECHNOLOGY,
`Patent Owner.
`____________
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`____________
`
`Record of Oral Hearing
` Held: May 8, 2018
`____________
`
`
`
`Before KEN B. BARRETT, TREVOR M. JEFFERSON and
`JOHN A. HUDALLA, Administrative Patent Judges.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`
`APPEARANCES:
`
`ON BEHALF OF THE PETITIONER:
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`RICHARD GOLDBERG, ESQUIRE
`MICHAEL H. SMITH, ESQUIRE
`WilmerHale
`Wilmer Cutler Pickering Hale and Dorr, LLP
`1875 Pennsylvania Avenue, N.W.
`Washington, D.C. 20006
`202-663-6055
`
`
`
`ON BEHALF OF THE PATENT OWNER:
`
`
`
`
`
`
`
`
`
`ALSO PRESENT:
`
`
`
`MICHAEL ROSATO, ESQUIRE
`QUINCY LU, ESQUIRE
`MATTHEW ARGENTINI, ESQUIRE
`Wilson Sonsini Goodrich & Rosati, LLP
`701 Fifth Avenue
`Seattle, Washington 98115
`206-883-3529
`
`
`
`JAMES M. DOWD, ESQUIRE
`Wilmer Cutler Pickering Hale and Dorr, LLP
`350 South Grand Avenue, Suite 2100
`Los Angeles, California 90071
`
`TODD M. BRIGGS, ESQUIRE
`Quinn Emanuel Urquhart & Sullivan, LLP
`555 Twin Dolphin Drive, 5th Floor
`Redwood Shores
`California 94065
`
`The above-entitled matter came on for hearing on Tuesday, May 8,
`
`2018, commencing at 9:01 a.m., at the U.S. Patent and Trademark Office,
`600 Dulany Street, Alexandria, Virginia.
`2
`
`
`
`
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`P R O C E E D I N G S
`JUDGE BARRETT: You may be seated. Good morning, everyone.
`We are on the record in IPR 2017-00700, IPR 2017-00701 and IPR 2017-
`00728, Apple Incorporated v. California Institute of Technology. Is that
`correct?
`MR. GOLDENBERG: Yes, Your Honor.
`JUDGE BARRETT: Thank you. If we can have the attorneys'
`appearances, please, Petitioner?
`MR. GOLDENBERG: Good morning, Your Honor. My name is
`Richard Goldenberg, I am Lead Counsel for Apple Incorporated, and with
`me at the counsel table is Michael Smith, my Co-Counsel.
`MR. SMITH: Hi.
`MR. GOLDENBERG: And with me right across the table is Jim
`Dowd, also my Co-Counsel.
`JUDGE BARRETT: Good morning. Welcome. Patent Owner?
`MR. ROSATO: Good morning, Your Honor. Mike Rosato for Cal
`Tech; I have Quincy Lu who is at the counsel table with me, Matthew
`Argenti behind me, and Todd Briggs. Thank you.
`JUDGE BARRETT: Welcome. If you are prepared to discuss, a
`preliminary matter, the Supreme Court's decision in SAS. I know you had a
`chance to meet and confer, and I want to commend you for that. What are
`your thoughts on how should we proceed; Petitioner?
`MR. GOLDENBERG: Yes, Your Honor. So, we did reach an
`agreement outside before the hearing this morning, and so we have
`agreement --
`
`
`
`3
`
`
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`JUDGE JEFFERSON: I'm sorry to interrupt you. Why don't you step
`to that podium, just in case the recording device doesn’t pick you up there?
`MR. GOLDENBERG: Thank you, Your Honor. Yes. Mr. Rosato
`and I spoke outside the hearing this morning, before the hearing, and we do
`have an agreement on a joint proposal to request that the Board dismiss or
`withdraw all the claims that were originally not instituted, in these IPRs as
`well as the IPRs that were argued a few weeks ago, 210, 219 and --
`MR. ROSATO: 297.
`MR. GOLDENBERG: 297. So, as long as the Board would agree to
`withdraw those claims such that there's no final written decision on them,
`then we would jointly request that the Board dismiss that.
`JUDGE BARRETT: All right. I tell you what I will do, after the
`hearing today or tomorrow, I will issue an order authorizing a joint motion,
`and we'll go from there.
`MR. GOLDENBERG: Thank you.
`JUDGE BARRETT: Thank you, both of you. All right then, on to
`the hearings. We set forth the procedure in the hearings order, but we do
`have some flexibility here. We set aside 30 minutes per side, per case. I
`suspect it won't take that long. We will have one continuous transcript for
`all three proceedings. There is no need to repeat arguments. I believe most
`of the arguments in these cases the issues are the same across all three. So,
`no need to repeat.
`The working assumption is, we will handle these sequentially. Is that
`still how you want -- the parties want to handle it?
`
`
`
`4
`
`
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`MR. GOLDENBERG: For Petitioner, yes, Your Honor. We agree
`that second and third hearings may be substantially shorter than the first, but
`we had planned to address them sequentially.
`MR. ROSATO: That's fine with us, Your Honor.
`JUDGE BARRETT: All right. Wonderful! As always, for clarity of
`the transcript when you refer to an exhibit on the screen, please state for the
`record the exhibit and the page number, or the demonstrative page, so we'll
`be able to understand that later.
`So, we are going to handle these sequentially. Petitioner, you'll go
`first. You can reserve time for rebuttal, and Patent Owner then will present
`its response, and then you can use your rebuttal time, Petitioner. And I'll
`give each -- I'm going to use the clock, and I'll give each Counsel a warning
`when you're reaching the end. Any questions?
`MR. GOLDENBERG: I'd like to reserve about 10 minutes for
`rebuttal, if I can, if Mr. Smith can give me a warning when I'm getting near
`20 minutes.
`MR. SMITH: Okay.
`JUDGE BARRETT: With that then, Petitioner, you may begin.
`MR. GOLDENBERG: Good morning, Your Honor. So, if we can
`turn to slide 2 for the 700 slides, I plan to follow a similar roadmap that we
`did in the prior hearings, but as always, if the Court has questions or would
`like to hear things in a different order, just let me know, I'm happy to
`address the Board's questions at any time.
`But turning to the first point on slide 4, and this is an overview of the
`invalidity of the challenged claims, here, on slide 4 we are showing the two
`
`
`
`5
`
`
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`
`grounds that have been instituted. All of the challenge claims here, there's
`only one independent claim, claim 11 is the independent claim, and all the
`claims are invalid over combinations of the Ping, MacKay and Divsalar
`references, with the Luby97 reference contributing to the invalidity of one
`claim, claim 13.
`Briefly, the Ping reference, teaches most of the limitations required by
`these claims, Ping could be thought of as missing two features. One would
`be irregularity, Ping's irregular code, and the claims require irregularity, and
`then the other would be an explicit teaching of repetition of bits; and
`MacKay, as to this combination, the teaching of irregularity, and Divsalar as
`the explicit teaching of repetition.
`And then finally, for the one remaining claim, the Luby97 reference
`provides a teaching of crossing the bits in a stream. It's a well-known detail,
`it's typified by Luby97 here, Luby97, is providing that same teaching as it
`did in the prior hearings that we had a few weeks ago.
`Now, turning to slide 5, this is claim 11, the one independent claim,
`and as you can see on the left claim 11 has some text, and then on the right,
`claim 11 also includes this graph, the Tanner graph. And I will try and
`show, at least at a high level, how the combination of Ping, MacKay and
`Divsalar, teach all the limitations required by this claim.
`And if we turn to slide 6, this is the construction for the Tanner graph
`term, and this is as it was presented in our petition, it was the Board's
`construction in the prior IPRs on the same patent involving the Hughes
`party. The Board here adopted the same construction, so this is the
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`6
`
`
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`
`construction of the Tanner graph, just broken up into three pieces, so we can
`deal with it one at a time.
`And I'll try and show how each of these pieces are meant by the art,
`and then I'm going to go in reverse order starting with the (iii) term. And so
`if we turn to slide 7, this is the (iii) portion of the construction, which
`requires parity bits, you know, parity bit is determined -- is the function of
`both information bits and other parity bits as shown by the configuration of
`nodes and edges in this Tanner graph.
`And at the bottom of the slide here, the bottom left, this is Ping's
`equation 4. And I'll try and show how Ping's equation teaches this
`limitation, or this portion of the construction.
`But just backing up for a moment to give an overview of Ping, in
`Ping's equation, the Ps stand for parity bits, the Ds stand for information bits,
`and then the Hs are elements of Ping's HD matrix.
`And every one of the parity bits involves a summation of Hs times Ds.
`That's a binary code, the elements of H will always be either one or zero,
`where H is zero it just zeros out the corresponding information bit. Where H
`is a 1, the summation turns into just a sum of information bits, because at
`one time the information bit has the same value as the information bit. So,
`the sum becomes -- the summation becomes a sum of some collection of
`information bits.
`And if we start with the first parity bit, P1, P1 is determined entirely
`as a sum of some number of information bits, or a subset of the information
`bits, wherever the H1 equals the 1.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`7
`
`
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`And then to get to P2 or all the remaining bits, you use the prior bit.
`So, if we take the example of the second parity bit, we can substitute number
`two in for the value (i) in the orange box, in Ping's equation, and that would
`tell us that P2, the second parity bit equals P1, the parity bit that we just
`calculated a moment ago, plus another sum of information bits, likely
`different information bits, but the basic formula holds, any parity bit after
`the first one will be the sum of the prior parity bit, and then plus a collection
`of information bits.
`And that's what's shown on the right. On the left of slide 7 is Ping's
`actual equation 4, on the right, we've just broken it out, to show what the
`explicit equation would be for a few of the parity bits.
`And now if we compare this back to the claim construction, the
`current construction is that the parity bit is determined as a function of both
`information bits and other parity bits. Well, that's just what Ping does.
`Every parity bit is a sum of the prior parity bit plus other information bits.
`And in fact, many of the parity bits depend on more than one parity bit.
`We could see that, for example, with P3 -- or P4, P4 equals P3 plus a
`bunch of information bits, P3 itself depends on P2 which depends on P1,
`therefore, once you get past the first couple, parity bits are determined as a
`function of both information bits and other parity bits, plural.
`So that's the arithmetic part of this construction, the remaining part is
`that it matches the configuration and nodes and edges of the Tanner graph.
`And Ping's equation does that, and I can show that by flipping back and
`forth between slides 7 and slide 5, that has the Tanner graph. And just to
`orient on this paragraph, here on the left, the U nodes, those represent
`
`
`
`8
`
`
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`
`information bits. The V nodes that are kind of in the middle, just to the right
`of the random permutation, they represent the check bits, so the parity
`checks, and then the bubbles on the right, the Xs, those represent the parity
`bits.
`
`The V nodes, every one of them enforces the constraint saying that the
`sum of all the edges connected to it must be zero. So, if we start with the
`first parity check, V1, V1 has, if we ignore ellipses, V1 has four edges
`connected to it. Three edges reach out to the left, to three of the information
`bits, three of the U-s, and then one of the edges reaches out to the right, to
`the first parity bit, X1.
`So the equation enforced by V1 is going to be: X1 plus the three
`information bits must equal zero. And in binary, Mod 2 arithmetic, anything
`plus itself equals zero, so that equation is the same as X1 equals the sum of
`three information bits.
`And that is exactly what Ping does on slide 7, the first parity bit, P1,
`which is X1 in terms of the Tanner graph, equals a sum of information bits.
`And now if we look at the next parity check back to slide 5, let's take
`V2, the second parity check, the check equation enforced by V2 is all the
`edges connected to it must be zero, so X1 plus X2 plus three more
`information bits, must all equal zero, if we add X2 to both sides that
`equation becomes X2 equals X1 plus 3 information bits. And that is, again,
`exactly what Ping does as we see on slide 5. Every parity bit after the first
`one equals the prior parity bit plus the sum of information bits.
`So, Ping doesn’t describe his code in terms of Tanner graph, he
`describes it using parity check matrices and equations, but as we explained
`
`
`
`9
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`
`to the petition, there is a duality between parity check matrices and Tanner
`graphs. One of skill in the art will understand if they see a parity check
`matrix they know what the corresponding Tanner graph would be, and that's
`not a point that's in dispute. So, that is how Ping teaches this portion of the
`construction.
`If we move to slide 8, and the (ii) part of the construction, the part to
`focus on here, is the word "randomly", so here the check nodes are randomly
`connected to the repeated message bits to enforce constraints determined by
`the parity bits. And Ping explicitly teaches that random aspect.
`In the code here, on slide 8, it explains that in the construction of HD
`we randomly create the distribution of 1s -- the number of 1s per column,
`and the number of 1s per row. Ping's random placement of the 1s in HD
`exactly corresponds to the random requirement in this part of the
`construction.
`And then if we move to slide 9; and now are at the single (i) part of
`the construction, and the part to focus on here is the repeat. I'll come back to
`irregular in a moment, because irregular is taught by MacKay, but if we just
`focused on the repeat part. So, it's a graph representing a repeat to
`accumulate code, and so on.
`Ping doesn’t explicitly teach repeating the information bits, but doing
`that would have been obvious particularly in view of Divsalar. The point
`here is that in Ping's HD matrix, in Ping's original disclosure, every column
`of HD has weight four, and that means that every information bit is going to
`contribute to exactly four parity checks.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`10
`
`
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`An obvious way for one ordinary skill to implement that would be to
`repeat the parity bits four times, so that there's one copy of the parity bit
`available for use in every one of the checks.
`Now, if you would have to implement it that way, it would have been
`obvious to implement Ping with or without such repetition, but the point is
`that implementing Ping with the repetition would have been obvious, and
`Divsalar shows an obvious way to do that repetition. Repeaters are well
`known, that's also not in dispute. So, Ping, in view of Divsalar, teaches this
`repetition aspect.
`That, I think covers everything required by the claim, claim 11, except
`for the notion of irregularity. And if we turn to slide 12, it's MacKay that is
`providing the teaching of irregularity, and --
`JUDGE BARRETT: Before you move on to that, the check nodes?
`MR. GOLDENBERG: Yes.
`JUDGE BARRETT: My understanding is, Petitioner's position is the
`check nodes provide constraints.
`MR. GOLDENBERG: Yes.
`JUDGE BARRETT: And then it is met by Ping's constraints,
`including the one about -- I believe I'm reading from Ping, "We randomly
`create exactly one element one per column," as being one of the constraints.
`MR. GOLDENBERG: Yes.
`JUDGE BARRETT: So if --
`MR. GOLDENBERG: Or at least four -- I'm sorry -- it's four per
`column we take before -- I'm sorry for the interruption, Your Honor, I
`
`
`
`11
`
`
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`
`believe you're right. HD is broken up into segments, and then each of the
`segments there's one, a single one per column?
`JUDGE BARRETT: Correct. So, my understanding is, Petitioner is
`using that to map to the check node requirement of the Tanner graph. The
`constraints imposed by that one element, one per column.
`MR. GOLDENBERG: I think it's -- I'd say it a little bit differently.
`In the columns of HD -- the number of 1s in the columns of HD determine
`the number of checks that a given information bit would contribute to; the
`number of 1s in a row of HD determine the number of information bits that
`would be summed for any given check. So, for a particular parity check,
`you would actually have to look at the 1s in the row, in the corresponding
`row of HD. But it is correct that it's -- HD is providing the constraints that
`match up with the Tanner graph.
`Does that -- I hope I haven't made things more confusing.
`JUDGE BARRETT: No. I think this is helpful to make sure we are
`all understanding, and the confusion may be on my end. Okay. So, you are
`saying the HD, generally is the constraint that corresponds to the check
`nodes?
`MR. GOLDENBERG: That's right.
`JUDGE BARRETT: Okay. As opposed to the specific requirement
`that Ping sets out as one element, one per column.
`MR. GOLDENBERG: That's right. It's the nature of HD itself that
`matches the check nodes in the Tanner graph, and there's no dispute about
`this in Ping's example, there are exactly four 1s per column, which would
`make Ping be a regular code. The irregularity comes from the teaching of
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`
`
`12
`
`
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`
`MacKay which would suggest that one of ordinary skill to make the weights
`of HD, the column weights of HD uneven, and once that is done, then you
`have an irregular code that matches claim 11.
`JUDGE BARRETT: Why I'm asking this line of questions, and I
`don't know if this is exactly the Patent Owner's position, but I want to make
`sure I understand, because as my initial questions suggested, my
`understanding is you were establishing that Ping disclosed constraints, and
`then by modifying to include irregularity, wouldn't you be destroying the
`constraint you just needed for the check node limitation?
`MR. GOLDENBERG: So, no, Your Honor. So, whether Ping has
`uneven weights per column or even weights per column, it still matches the
`check node requirement, making the column weights uneven then matches
`the claims requirement that the information bits be used irregularly.
`So, I think that the point that -- the Patent Owners argue that there
`would not have been a motivation to use MacKay's irregularity in Ping,
`because Ping sets forth these so-called rules that there are going to be so
`many 1s per column, and so many 1s per row.
`I think that's the wrong way to look at it, and it's not how one of
`ordinary skill would have looked at it. Ping's overall parity check matrix
`includes two portions, one called HP, and then one called HD. The HP
`portion can have only one -- and this was Ping's contribution really, is to
`take an LDPC matrix and turn it into something that could be encoded
`quickly.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`
`
`
`13
`
`
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`One of the criticisms of LDPCs is -- prior to this, was if you had a low
`density parity check matrix, the generator would be high density, and high
`density generator matrix is going to make the encoding slide.
`What Ping did, was he disclosed this way of splitting up the parity
`check matrix where you have something very regular and determined in HP,
`and it's not in dispute that there's only one form for HP, one of skill in the art
`would not have messed with HP.
`HD though, has a random structure. It's both low density and random.
`That's what makes the encoding fast, because HD is actually used as the
`generator matrix in Ping. And you can see that from this equation on slide 5,
`we compute the parity bits from the information bits in HD, mathematically,
`in terms of the (inaudible), the same thing would be represented by
`multiplying the HD matrix times the column of the information bits to get
`the parity bits, that makes HD be a generator.
`So, backing up, Ping has got irregular structure in HP, a random
`structure in HD, as long as HD is both random and low density, it meets the
`spirit of Ping. And what MacKay adds, is to take Ping's explicit example,
`HD had the same number of 1s per row, and the same number of 1s per
`column, but MacKay is teaching that irregularity improves things, and
`would have motivated one of ordinary skill to make the HD -- make the
`weights uneven. Is that --
`JUDGE BARRETT: I think that's quite helpful. By my count you
`have about five more minutes.
`MR. GOLDENBERG: Thank you, Your Honor. One point, as long
`as we are on this, that I wanted to mention is, there has been a lot of
`
`
`
`14
`
`
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`
`discussion in this case about unpredictability, and the Patent Owners argue
`that the unpredictability of this field would have made it -- would dissuade
`one of ordinary skill from modifying Ping to use MacKay's irregularity.
`We, of course, disagree with that. Let me see if I can find the slide
`here, so on slide 11, we've got the motivation coming from MacKay for one
`of ordinary skill to use MacKay's irregularity in Ping. MacKay taught that
`irregularity clearly improved the code. One or ordinary skill would have
`been motivated to use that.
`Now, according to the Patent Owner, one of ordinary skill would have
`looked at this and said, irregularity, I'd better stay away from that, too
`unpredictable, don't use it, don't touch it.
`That's not the mindset of one of ordinary skill, it makes -- that would
`turn one of ordinary skill into an automaton. One of ordinary skill reading
`Ping and MacKay would have been motivated to try and obtain the superior
`part -- performance offered by MacKay's irregularity in Ping.
`Now there are many ways one of ordinary skill could have done that.
`There are lots of -- as the Patent Owner points out, there are lots of
`modifications that could be done to HD. Every one of those would have
`invalidated the claims. There's no dispute that making HD irregular results
`in irregular code that meets the limitations. And every single modification
`to HD that makes it irregular, meets the claims.
`JUDGE HUDALLA: Can I ask you about?
`MR. GOLDENBERG: Yes.
`JUDGE HUDALLA: In the last hearing, you were asked about
`obvious-to-try, and that's what you're, again, sounding like, there's a million
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`
`
`15
`
`
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`
`ways we could do this, right? But you are not making an obvious-to-try
`argument in this case. I mean, that's your position, right?
`MR. GOLDENBERG: That's true. And we actually try to avoid that,
`Your Honor.
`JUDGE HUDALLA: Okay.
`MR. GOLDENBERG: What I'm saying is, one of ordinary skill
`would have been motivated to use the irregularity and the matter -- there
`would have been, with certainty, the modification would have resulted in a
`code that worked. The only thing that's in dispute is with that -- with the
`performance of that code have been optimal. Well, that would have -- that
`probably would have required some experimentation that one of the skill in
`the art was used to doing in this field, to arrive at an optimal code, or one
`that had desired performance.
`But there's no performance requirement in any of these claims, or the
`claims that we discussed a few weeks ago. So, the point I'm trying to make
`is that once one of ordinary skill uses the irregularity of MacKay in Ping, as
`they would have been motivated to do, these claims are invalid, because we
`now have a combination that meets every limitation of the claims, and that's
`true no matter which way one of ordinary skill might have decided to modify
`Ping's HD.
`Now, the specific example that we have posited was MacKay talks
`about column weights of 9 and 3, it would have been obvious to use 9 and 3
`column weights in HD, any which way one of ordinary skill in the art did
`that, they get the results in the claim -- the combination that invalidates the
`claims. Does that --
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`
`
`16
`
`
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`JUDGE JEFFERSON: I think it does. I just -- you know, I guess
`that's how you would respond to Patent Owner's argument that there's -- I
`think they said 10 to the 15th, or something, number of permutations that
`could come out of this combination.
`MR. GOLDENBERG: And every one of those permutations would
`invalidate the claim.
`JUDGE HUDALLA: Okay.
`MR. GOLDENBERG: Every one results in a code that functions as a
`code, there's no doubt they would produce parity bits, there's no doubt they
`could be decoded. The only thing that's in doubt that might have required
`some experiment would be the performance. Is this an optimally-performing
`code, but that's actually irrelevant for all the claims under consideration.
`JUDGE HUDALLA: I mean, now you're turning it into -- that kind of
`moves us into reasonable expectation of success. I mean, obviously a person
`of ordinary skill in the art wouldn't have wanted a worse code to come out of
`this combination you're proposing, right?
`MR. GOLDENBERG: That's true. And if MacKay had -- if the
`teachings of MacKay had been that irregularity is something that makes
`things worse and should be avoided at all cost, then maybe there would not
`have been a reasonable expectation of success. But the teaching of MacKay
`is the opposite. MacKay is teaching, this is exciting, the irregularity
`improved the performance. That's what would have motivated one of
`ordinary skill to make the combination, and they would have had an
`expectation that they modified code would work as a code.
`JUDGE HUDALLA: I think I understand your position.
`
`
`
`17
`
`
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`JUDGE BARRETT: Your initial 20 minutes is up.
`MR. GOLDENBERG: Okay. Thank you, Your Honor. I'm just
`going to mention a couple things quickly, and I'll reserve the rest of my time.
`I think that covers the overview of why the claims are invalid.
`On slide 16, as we saw the last time the Patent Owner had chosen not
`to depose or reply declarant, not Dr. Frey or Dr. Davis in a second
`declaration. As the record stands, their testimony is unchallenged.
`I think, I'm just going to address one point about the row weights. So
`the Patent Owners argued that MacKay does not teach uninform row
`weights, so I think it's undisputed that MacKay teaches non-uniform column
`weights.
`This pertains to just one dependent claim, I believe it's dependent
`claim 12. MacKay, and here we have on slide 21, MacKay teaches, looking
`at the bottom part of the code we select a profile that describes the desired
`number of columns of each weight, and the desired number of rows of each
`weight.
`The desired number of rows of each weight tells one of ordinary skill
`in the art that there can be at least two rows with different weights, and that's
`all it takes to invalidate claim 12.
`This wasn’t the focus of MacKay. In MacKay's examples all the row
`rates were even, but he's got a clear teaching that the row weights can be
`uneven, and are part of the selection process.
`I think our point is that just making the row weight uneven is a trivial
`detail, the claims are obvious whether the row weights -- and MacKay
`
`
`
`18
`
`
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`
`teaches both regular, all even, and all not even row weights, but that detail of
`claim 12 is clearly met by this teaching that we have on slide 21.
`If there are no questions on that point, I'll go ahead and reserve the
`rest of my time for rebuttal. Thank you.
`JUDGE BARRETT: I have eight minutes left.
`MR. GOLDENBERG: Thank you.
`MR. ROSATO: Good morning. And may it please the Board. Now,
`
`consistent with what I understand to be the preference, we've reserved any
`objections during Counsel's speaking time, to not interrupt. But I do want to
`note an objection to, of course, any new argument that was advanced there,
`and there was quite a bit, particular on this issue of the Tanner graph recited
`in the claim.
`And I'm going to turn, if I can find my slide here, but one of the issues
`identified in the briefing was that while claims 11, and in claim 11 and
`particularly in the claims of the patents that are being challenged, recite
`generating parity bits specifically in accordance with an illustrated, and
`provided Tanner graph. One of the points of criticism with Petitioner's
`petition materials, was there is no mapping at all to that particular Tanner
`graph, that they instead relied on a verbal description of aspects of the
`Tanner graph, that verbal description than referring back to the Tanner graph
`itself.
`And we had pointed out that as part of a petition case and, you know,
`consistent with the statutory requirements for specificity and clarity in the
`stated case, but there is a lack of mapping to the Tanner graph, particularly
`with regard to this -- piece 3 of the verbal description.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`
`
`19
`
`
`
`Case IPR2017-00700 (Patent 7,421,032 B2)
`Case IPR2017-00701 (Patent 7,421,032 B2)
`Case IPR2017-00728 (Patent 7,421,032 B2)
`
`
`1
`2
Accessing this document will incur an additional charge of $.
After purchase, you can access this document again without charge.
Accept $ Charge
Still Working On It
This document is taking longer than usual to download. This can happen if we need to contact the court directly to obtain the document and their servers are running slowly.
Give it another minute or two to complete, and then try the refresh button.
A few More Minutes ... Still Working
It can take up to 5 minutes for us to download a document if the court servers are running slowly.
Thank you for your continued patience.
This document could not be displayed.
We could not find this document within its docket. Please go back to the docket page and check the link. If that does not work, go back to the docket and refresh it to pull the newest information.
Your account does not support viewing this document.
You need a Paid Account to view this document. Click here to change your account type.
Your account does not support viewing this document.
Set your membership
status to view this document.
With a Docket Alarm membership, you'll
get a whole lot more, including:
- Up-to-date information for this case.
- Email alerts whenever there is an update.
- Full text search for other cases.
- Get email alerts whenever a new case matches your search.
One Moment Please
The filing “” is large (MB) and is being downloaded.
Please refresh this page in a few minutes to see if the filing has been downloaded. The filing will also be emailed to you when the download completes.
Your document is on its way!
If you do not receive the document in five minutes, contact support at support@docketalarm.com.
Sealed Document
We are unable to display this document, it may be under a court ordered seal.
If you have proper credentials to access the file, you may proceed directly to the court's system using your government issued username and password.
Access Government Site
