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
`
`Page 1
`
`Michael Mitzenmacher, PhD
` IN THE UNITED STATES PATENT AND TRADEMARK
`OFFICE BEFORE THE PATENT TRIAL AND APPEAL BOARD
`APPLE INC.,
`PETITIONER
`v.
`CALIFORNIA INSTITUTE OF TECHNOLOGY
`PATENT OWNER
`
`IPR2017-00210
`IPR2017-00700
`IPR2017-00219
`IPR2017-00728
`IPR2017-00701
`IPR2017-00297
`
`DEPOSITION OF
`MICHAEL MITZENMACHER, PhD, Volume I
`Portland, Oregon
`Wednesday, December 20, 2017
`
`Reported by:
`LISA TRONCOSO, RPR, CSR, CLR
`JOB NO. 135227
`
`TSG Reporting - Worldwide 877-702-9580
`
`CALTECH - EXHIBIT 2038
`Apple Inc. v. California Institute of Technology
`IPR2017-00701
`
`
`2
`3
`4
`5
`6
`7
`8
`9
`
`10
`
`11
`
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`
`22
`23
`24
`25
`
`Page 1
`
`Michael Mitzenmacher, PhD
` IN THE UNITED STATES PATENT AND TRADEMARK
`OFFICE BEFORE THE PATENT TRIAL AND APPEAL BOARD
`APPLE INC.,
`PETITIONER
`v.
`CALIFORNIA INSTITUTE OF TECHNOLOGY
`PATENT OWNER
`
`IPR2017-00210
`IPR2017-00700
`IPR2017-00219
`IPR2017-00728
`IPR2017-00701
`IPR2017-00297
`
`DEPOSITION OF
`MICHAEL MITZENMACHER, PhD, Volume I
`Portland, Oregon
`Wednesday, December 20, 2017
`
`Reported by:
`LISA TRONCOSO, RPR, CSR, CLR
`JOB NO. 135227
`
`TSG Reporting - Worldwide 877-702-9580
`
`CALTECH - EXHIBIT 2038
`Apple Inc. v. California Institute of Technology
`IPR2017-00701
`
`
`
`Michael Mitzenmacher, PhD
`December 20, 2017
`9:00 a.m.
`
`Page 2
`
`Deposition of MICHAEL MITZENMACHER,
`PhD, Volume I, held at the offices of Wilson,
`Sonsini, Goodrich & Rosati, 701 Fifth Avenue,
`Suite 5100, Seattle, Washington, before
`Lisa Troncoso, a Registered Professional
`Reporter, Certified Livenote Reporter, and
`Oregon Certified Shorthand Reporter
`Number 14-0429.
`
`TSG Reporting - Worldwide 877-702-9580
`
`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
`
`
`
`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
`
`Page 3
`
` Michael Mitzenmacher, PhD
`A P P E A R A N C E S
`WILMERHALE
`Attorneys for Petitioner
` 60 State Street
` Boston, MA 02109
`BY: RICHARD GOLDENBERG, ESQ.
` - AND -
` MICHAEL SMITH, ESQ.
`
`Wilson Sonsini Goodrich & Rosati
`Attorneys for Patent Owner
` 701 Fifth Avenue
` Seattle, Washington 98104
`BY: MICHAEL ROSATO, ESQ.
` - AND -
` QUINCY LU, ESQ.
` - AND -
` MATTHEW ARGENTI, ESQ. (via telephone)
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 4
`
`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
`
` Michael Mitzenmacher, PhD
` -O0O-
` MICHAEL MITZENMACHER, PhD,
`called as a witness, having first been sworn by
`the Certified Shorthand Reporter, was examined
`and testified as follows:
` MR. ROSATO: Before you start, just
` so you -- I'm just organizing stuff and
` noticed this. I think these are just --
` looks like just clean copies of his
` declaration, if you want to take a look.
` MR. GOLDENBERG: Let me just take a
` peak. That's fine, that will spare us some
` exhibits.
` (Michael Mitzenmacher's folder of
` Declarations, marked.)
` MR. GOLDENBERG: So, sorry, we did
` that on the record, right?
` MR. ROSATO: It looks like we did
` actually.
`EXAMINATION BY:
`MR. GOLDENBERG:
` Q. Dr. Mitzenmacher, good morning.
` A. Good morning.
` Q. My name is Richard Goldenberg and
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 5
`
`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
`
` Michael Mitzenmacher, PhD
`I'll be taking your deposition today.
` A. Okay.
` Q. Have you been deposed before?
` A. Yes.
` Q. How many times?
` A. Many. Like, I don't recall, but
`probably more than twenty.
` Q. All for patent cases?
` A. Some of the cases involved like
`trade secrets or copyright, most of them have
`been patent cases.
` Q. And where are you presently
`employed?
` A. Harvard University.
` Q. And are you doing a post in Seattle?
`I'm just curious why the deposition is in
`Seattle and not Boston.
` A. The lawyer sort of said, you know,
`where would you like to do it and I really
`didn't have a preference. You know, my parents
`lived in L.A. on the west coast, so I said I'd
`be more than happy to come out on the west
`coast given the cold weather in Boston, and
`then give me a chance -- I stopped and saw my
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 6
`
`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
`
` Michael Mitzenmacher, PhD
`parents before coming up here which, you know,
`was a nice opportunity for me.
` Q. Sounds good. I was just curious.
` So I was -- just before I saw the
`binder that your counsel showed me, I was going
`to give you one of the -- one of your
`declarations, but you already have a copy.
` A. However you want to do it. Whatever
`is easier for you.
` Q. Can you turn to the declaration for
`the '781 patent in your binder there?
` MR. ROSATO: Counsel, if you have
` copies, can I also have a copy?
` MR. GOLDENBERG: Sure.
` MR. ROSATO: Thank you.
` THE WITNESS: The '781, so for the
` IPR00297?
`BY MR. GOLDENBERG:
` Q. Correct.
` A. I got it.
` Q. And if we just turn to the last
`page, page 88, you see it's signed Michael
`Mitzenmacher, PhD?
` A. Uh-huh.
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 7
`
`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
`
` Michael Mitzenmacher, PhD
` Q. Is that your signature?
` A. Yes.
` Q. And you authored this declaration?
` A. Yes.
` Q. And you authored the other
`declarations that are at issue in this case?
` A. Yes.
` MR. ROSATO: Objection, form.
`BY MR. GOLDENBERG:
` Q. Maybe we can just go through your
`binder. Did you author the declaration of
`Michael Mitzenmacher for IPR2017-00210 that's
`labeled as Exhibit 2004 in that IPR?
` A. Sorry, which one again?
` Q. You probably want to look at the
`first page.
` A. 2017-210. Yes, that's my signature,
`and this is my report expressing my opinions
`that I wrote and worked on with counsel.
` MR. ROSATO: Counsel, I don't want
` to cause unnecessary time, just noting that
` there are other declarations besides
` Mitzenmacher declarations.
` MR. GOLDENBERG: Fair enough.
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 8
`
`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
`
` Michael Mitzenmacher, PhD
` MR. ROSATO: That's all.
` MR. GOLDENBERG: Yup.
`BY MR. GOLDENBERG:
` Q. Just to capture this in one quick
`question for the record, Dr. Mitzenmacher, you
`were the author of declaration of Michael
`Mitzenmacher for IPR2017-00210 that's marked as
`Exhibit 2004, correct?
` A. Yes. Reporting my opinions, I
`authored it and obviously I worked on it with
`counsel.
` Q. And just flip to the declaration for
`the next IPR2017-219.
` A. Yes.
` Q. So you were the author of the
`declaration of Michael Mitzenmacher for
`IPR2017-00219 that's marked as Exhibit 2004,
`correct?
` A. Yes.
` Q. And if you turn to the next
`declaration that's the IPR2017-00297.
` A. Yes.
` Q. So, Dr. Mitzenmacher, you're also
`the author of the declaration of Michael
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 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
`
` Michael Mitzenmacher, PhD
`Mitzenmacher for IPR2017-00219 that's marked as
`Exhibit 2004, correct?
` A. Yes.
` Q. And if you turn to -- let me start
`again. Are you ready?
` A. Yeah.
` Q. Do you have your declaration for
`IPR2017-700, 701 and 728?
` A. Yes.
` Q. So you are the author of the
`declaration of Dr. Michael Mitzenmacher for
`IPR2017-700, 2017-701 and 2017-728 that's
`marked as Exhibit 2004, correct?
` A. Yes. So same answer, I am the
`author, these express my opinions and, again,
`obviously I worked on them with counsel in
`preparing the reports.
` Q. Understood. Thank you.
`Dr. Mitzenmacher, every parity-check matrix has
`a corresponding tanner graph, correct?
` MR. ROSATO: Objection, form.
` THE WITNESS: I'd say that you can
` take a parity-check matrix and one of the
` ways you can express it is through a tanner
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 10
`
`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
`
` Michael Mitzenmacher, PhD
` graph. As a scientist or engineer I always
` hesitate to say every, but scientists don't
` always like to work in absolutes, but I
` can't think of an example of a, you know,
` certainly standard parity-check matrix that
` you cannot express as a tanner graph.
`BY MR. GOLDENBERG:
` Q. Dr. Mitzenmacher, just to make sure
`I understood your testimony there, if one has a
`parity-check matrix one can generate a tanner
`graph that expresses the same code, correct?
` A. If one has a parity-check matrix,
`again, in the manner described, for example, in
`some of my work, and if you're using the same
`definition of a parity-check matrix then, yes,
`one of the ways you can express the code that
`that matrix represents is also through a tanner
`graph.
` Q. And that works in the reverse, too,
`correct. If you have a tanner graph you could
`generate a corresponding parity-check matrix,
`correct?
` A. I would say only for limited classes
`of tanner graphs that would depend on how you
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 11
`
`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
`
` Michael Mitzenmacher, PhD
`were using the term. Typically, in the modern
`parlance, tanner graphs would be understood
`that you could have more complicated operations
`and simple parity checks, in which case you
`might not be able to represent that more
`complex type of operation within a parity-check
`matrix.
` Q. Now, not that previous question, but
`the one before you'd mention if we're using the
`same definition of parity-check matrix.
` A. Yes.
` Q. What is your definition of
`parity-check matrix?
` A. So I would look back, for instance,
`in some of my papers, or I can look for the
`report, but typically a parity-check matrix you
`have a matrix where the entries in the matrix
`represent that the values represent that
`certain -- what are sometimes referred to as,
`in this setting, information bits, or bits that
`are going to, or depending on the setting
`they're also referred to as message bits or
`message nodes.
` Typically I use like message nodes.
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 12
`
`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
`
` Michael Mitzenmacher, PhD
`In this context, you know, they do have message
`nodes, and you have certain constraints on the
`message nodes and parity-check matrix
`represents these constraints, the constraints
`being of the form that the sum of the values
`represented by the one entries in the row of a
`matrix, zero, generally using arithmetic, but
`that depends on the setting as well.
` Q. With that understanding of the
`parity-check matrix, is it correct to say that
`a parity-check matrix will have a corresponding
`tanner graph?
` A. Yes, in the manner that we've
`described.
` Q. Is it also correct that if a
`parity-check matrix is multiplied by a valid
`code word that the result will be zero?
` A. Yes, that's what the constraints of
`the parity-check matrix represent.
` Q. Is it also correct that if a
`parity-check matrix is multiplied by a
`generator matrix for the same code that the
`result will be zero?
` MR. ROSATO: Objection, foundation.
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 13
`
`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
`
` Michael Mitzenmacher, PhD
` THE WITNESS: Do you mean a zero
` matrix, like it won't actually be zero or
` zero valued, so you mean like a zero
` matrix?
`BY MR. GOLDENBERG:
` Q. Yes, a matrix of all zeros.
` A. Let's see, I would have to double
`check, but I think that's right.
` Q. So is it correct to say that a
`parity-check matrix will have a corresponding
`set of generator matrices?
` MR. ROSATO: Objection, form,
` foundation.
` THE WITNESS: I don't think I would
` put it that way. I would say that there
` are relationships between parity matrices
` and generator matrices that, you know, a
` parity-check matrix in that manner
` corresponds to a code, and the generator
` matrix corresponds to a code, and there are
` various ways and means that you might
` transform one into the other that are known
` in the art.
`BY MR. GOLDENBERG:
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 14
`
`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
`
` Michael Mitzenmacher, PhD
` Q. And what is a generator matrix?
` A. In this context, a generator matrix
`is often used to describe a way of mapping a
`message or an input into a code word or an
`output.
` Q. Is it correct to say that a
`generator matrix multiplied by a vector of
`information bits will equal a code word for the
`code?
` MR. ROSATO: Objection, form,
` foundation.
` THE WITNESS: I would be careful of
` using the term information bits. You know,
` the information bits, the way I understand
` them, Dr. Davis describes information bits
` as a message that's meant to be sent, and
` there are various transformations that you
` can put on these information bits before
` undergoing this sort of encoding.
` So I might prefer the terminology
` of, you know, message bits, or depending on
` the context, again, assuming that we're
` talking about a binary code and working on
` bits, they're codes that work on,
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 15
`
`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
`
` Michael Mitzenmacher, PhD
` obviously, data units that are larger than
` bits.
`BY MR. GOLDENBERG:
` Q. And if this helps avoid confusion, I
`think we can restrict all questions and answers
`today to binary codes.
` A. All right.
` Q. Are you okay with that?
` A. That would be fine.
` Q. If for some reason we need to refer
`to a nonbinary code, then we'll make explicit
`reference to it, but for now we can assume that
`all questions and answers pertain only to
`binary codes, okay?
` A. Okay.
` Q. Now, I think you said -- withdrawn.
` A message bit could be either an
`information bit or a parity bit, correct?
` A. That would depend on the context or
`setting, or how people -- yes, but typically
`often it's used in that regard.
` Q. I mean, is that how you use the
`term?
` A. I believe in at least some papers
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 16
`
`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
`
` Michael Mitzenmacher, PhD
`that's how I've used the term.
` Q. And you would not multiply a
`generator matrix times parity bits in order to
`encode a code word, correct?
` A. No, you would -- the parity bits
`would be the -- produced by the coding
`operations.
` Q. So you expressed some reticence
`about using the term information bits. What's
`a good term to describe the bits of a code word
`that in a systematic code that are not parity
`bits?
` MR. ROSATO: Objection, form.
` THE WITNESS: Sorry, if you parse a
` bunch of things, bits of the code in a
` systematic code?
`BY MR. GOLDENBERG:
` Q. Right.
` A. I guess more talking about the
`generator matrices in particular, what I'm
`talking about is independent of whether they
`were systematic or nonsystematic. You know, I
`think generally these might be referred to as
`the systematic bits.
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 17
`
`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
`
` Michael Mitzenmacher, PhD
` Q. Okay. So is it accurate to say then
`that a generator matrix times a vector of
`systematic bits will equal a code word?
` MR. ROSATO: Objection, form,
` foundation.
` THE WITNESS: In the context of
` systematic codes, right, where you were
` also transmitting it, that would be one way
` you can refer to it.
`BY MR. GOLDENBERG:
` Q. Well, even in a nonsystematic code
`there's some set of information that's desired
`to be transmitted, correct?
` A. Yes.
` Q. And what's a good term for that
`information?
` A. So, again, I don't mind information
`bits as representing in some sense the
`information that you end up desiring to be
`transmitted. It's just that that might not
`actually be the information that you input or
`multiply within your parlance to create the
`code word.
` There may be other transformations
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 18
`
`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
`
` Michael Mitzenmacher, PhD
`you put on those bits beforehand before
`actually doing the matrix multiplication for
`the code, so that's something that I think was
`unclear in Dr. Davis' description.
` Q. Is it correct that a generator
`matrix times a vector of bits will equal a code
`word?
` MR. ROSATO: Objection, form,
` foundation.
` THE WITNESS: Yes.
`BY MR. GOLDENBERG:
` Q. And what's a good term to describe
`that vector of bits?
` MR. ROSATO: Same objection.
` THE WITNESS: Again, I would call it
` perhaps just the input to the code.
`BY MR. GOLDENBERG:
` Q. So a generator matrix times a vector
`of code input equals a code word, correct?
` MR. ROSATO: Same objection.
` THE WITNESS: I think that's good
` terminology.
`BY MR. GOLDENBERG:
` Q. And a parity-check matrix times a
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 19
`
`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
`
` Michael Mitzenmacher, PhD
`code word will equal zero, correct?
` A. Yes, in standard codes.
` Q. And I'm looking for a way to
`describe corresponding parity check and
`generator matrices. Would it be correct to say
`that if a generator matrix times a code input
`equals a zero, and a parity-check matrix times
`all the code words that can be produced by that
`generator also equal zero, then the generator
`matrix and parity-check matrix are for the same
`code?
` MR. ROSATO: Objection, form.
` THE WITNESS: Sorry, I missed at
` least part of it. So the generator matrix
` times --
`BY MR. GOLDENBERG:
` Q. Let me withdraw the question and
`start over. Rather than me stating the answer,
`I'll ask you to state the answer.
` What's the right criteria for a
`parity-check matrix and generator matrix to be
`said to be used for the same code?
` MR. ROSATO: Objection, form,
` foundation.
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 20
`
`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
`
` Michael Mitzenmacher, PhD
` THE WITNESS: I would say typically,
` you know, if a generator matrix based code
` is using as an input some K bits, and
` outputting N bits, then the collection of K
` bit inputs will produce a collection of N
` bit outputs, the code words, and if for all
` of those code words the parity-check matrix
` times that code word leads to a zero, or
` that is that those code words and only
` those code words satisfy the constraints,
` then they represent the same code.
`BY MR. GOLDENBERG:
` Q. And every parity-check matrix will
`have a set of generator matrices that represent
`the same code as the parity-check matrix,
`correct?
` MR. ROSATO: Objection, form,
` foundation.
` THE WITNESS: Right. If you're
` getting that they're -- conceivably can be
` multiple generator matrices that can lead
` to the same code there would be multiple
` parity check matrices that lead to the same
` R=1 code.
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 21
`
`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
`
` Michael Mitzenmacher, PhD
`BY MR. GOLDENBERG:
` Q. So multiple parity check matrices
`can define the same code, correct?
` MR. ROSATO: Same objection.
` THE WITNESS: Yes. Again, when we
` talk about the same code, maybe we should
` be a bit clearer, you know, exactly what it
` means subject to, you know.
` You can either mean they must come
` out with the exact same code words or you
` might mean the same code words up to some
` sort of permutations of the order of the
` bits, so on.
` But I'd say under those criteria
` it's possible that multiple parity check
` matrices can define the same code certainly
` up to at least permutation of the bits.
`BY MR. GOLDENBERG:
` Q. Any LDPC code could be described by
`multiple parity check matrices, correct?
` MR. ROSATO: Same objection.
` THE WITNESS: With that
` understanding that I stated, yes.
`BY MR. GOLDENBERG:
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 22
`
`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
`
` Michael Mitzenmacher, PhD
` Q. And there are also multiple
`generator matrices that could be used to
`produce code words for the same code, correct?
` MR. ROSATO: Same objection.
` THE WITNESS: Again, with the
` understanding of, you know, we're talking
` about up to permutation of ordering things
` like that, then certainly.
` MR. GOLDENBERG: Mr. Rosato, what's
` the basis for your objection?
` MR. ROSATO: Form and foundation.
` MR. GOLDENBERG: What specifically
` about the form did you feel was incorrect?
` MR. ROSATO: And foundation. I
` mean, what code words are we talking about?
` There's a lot of generalities here. I'm
` preserving objections because there is open
` questions in terms of scope, how this
` relates to his direct testimony. If this
` ends up going somewhere then we'll deal
` with the objections accordingly.
` MR. GOLDENBERG: What specifically
` about the form of the question did you feel
` was improper?
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 23
`
`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
`
` Michael Mitzenmacher, PhD
` MR. ROSATO: I'll have to look at
`the specific question you're asking, as you
`seem to ask a general question about all of
`the objections.
` The question is, like I said,
`there's -- the context is fairly unclear.
`I don't want to coach the witness, so we're
`very limited in the practice and limited to
`very specific types of objections.
` MR. GOLDENBERG: I just, to the
`extent I can, I want to ask questions that
`you find that's non-objectionable and
`that's why I'm asking you.
` MR. ROSATO: I understand.
` MR. GOLDENBERG: But what
`specifically about the form did you find
`improper? If there's something I can
`correct, I'll correct it.
` MR. ROSATO: There are aspects of
`the questions that are vague.
` MR. GOLDENBERG: I'm asking you
`about the specific last question. You made
`a form objection, what was your basis?
` MR. ROSATO: There are aspects that
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 24
`
`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
`
` Michael Mitzenmacher, PhD
` are vague, where the relevance to the
` direct testimony is at a minimum, unclear,
` and this line of questioning has gone on
` for a fair amount of time without any
` specific reference to any specific document
` that's been entered into the record, relied
` upon, or discussed by Mr. Mitzenmacher in
` his direct testimony.
` MR. GOLDENBERG: Thank you, counsel.
`BY MR. GOLDENBERG:
` Q. Is it common in error correcting
`code literature to describe a code by a
`parity-check matrix?
` A. Common? I mean, I would say that's
`one of the ways that's used. I have seen it in
`some standard textbooks and so on. I mean, I
`can't speak to the wide commonality, but I
`believe that most people in coding would
`understand the parity-check matrix in that
`context.
` Q. But you're an expert in error
`correcting codes, right?
` A. Yes.
` Q. Within your expertise, have you seen
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 25
`
`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
`
` Michael Mitzenmacher, PhD
`many references that describe a code in terms
`of this parity-check matrix?
` MR. ROSATO: Objection, form, scope.
` THE WITNESS: Like I've said, I've
` seen in textbooks and numerous papers and
` so on, you know, so in that sense
` certainly, yes. To the extent that's
` commonality, I don't think I'd be prepared
` to judge or say, but I answered.
`BY MR. GOLDENBERG:
` Q. People of ordinary skill in error
`correcting can determine a generator matrix for
`a parity-check matrix, correct?
` MR. ROSATO: Objection, form.
` THE WITNESS: Again, I think one
` skilled in the art would be able to find
` methods for doing that. For example,
` various textbooks and so on, they did not
` know it of themselves.
`BY MR. GOLDENBERG:
` Q. And that is one of ordinary skill
`could produce a generator matrix that is for
`the same code as a parity-check matrix,
`correct?
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 26
`
`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
`
` Michael Mitzenmacher, PhD
` MR. ROSATO: Objection, form, scope.
` THE WITNESS: I'd say generally I
` think that's possible. Although, again,
` we'd have to look at specific instances to
` see if there were any complications. I'd
` say generally I would think that would be
` something one of skill in the art could do.
`BY MR. GOLDENBERG:
` Q. And so this may seem repetitive, but
`I just want to ask the question a slightly
`different way. Are you ready?
` A. Okay.
` Q. Given a parity-check matrix, one of
`ordinary skill can find a generator matrix for
`the same code, correct?
` MR. ROSATO: Same objection.
` THE WITNESS: Again, so given a
` specific parity-check matrix one would, in
` the skill of the art, would know methods
` that would allow one to transform it into a
` generator.
` Sorry. To use that to determine a
` corresponding generator matrix by various
` means. You know, again, whether they
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 27
`
`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
`
` Michael Mitzenmacher, PhD
` specifically are be able to do it might
` matter on things such as the actual size
` and so on, but for methodology for
` connecting parity-check matrix and
` generator matrices, no.
`BY MR. GOLDENBERG:
` Q. And in particular, given a
`parity-check matrix for an LDPC code, one of
`ordinary skill could determine a generator
`matrix for that code, correct?
` MR. ROSATO: Same objection.
` THE WITNESS: With the understanding
` and caveats I've described, yes, if one is
` given a specific matrix there are known
` methodologies.
`BY MR. GOLDENBERG:
` Q. In LDPC code -- withdrawn.
` For any given LDPC code there will
`always exist a generator corresponding to a
`systematic version of the code, correct?
` MR. ROSATO: Objection, form.
` THE WITNESS: I'm not clear on what
` you mean exactly by that question.
`BY MR. GOLDENBERG:
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 28
`
`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
`
` Michael Mitzenmacher, PhD
` Q. Well, just for clarity, what's a
`systematic code?
` A. A systematic code is one in which
`we're referring to before, I guess, as to the
`code bits are transmitted as part of the code
`word or form part of the code word.
` Q. So in a systematic code, the input
`of the code forms part of the code word,
`correct?
` A. Yes.
` Q. And a parity-check matrix doesn't
`necessarily determine whether a code will be
`systematic or not systematic, correct?
` MR. ROSATO: Objection, form,
` foundation.
` THE WITNESS: Correct.
`BY MR. GOLDENBERG:
` Q. But for any given LDPC parity-check
`matrix there exists a generator for that code
`that will produce a systematic code, correct?
` MR. ROSATO: Same objection.
` THE WITNESS: So if I understand
` that question, you're saying is there a way
` of setting up a generator that would make
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 29
`
`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
`
` Michael Mitzenmacher, PhD
` a -- such a code systematic, right, that is
` potentially possible. Although, again,
` that doesn't mean that any specific low
` density parity check code or code
` implementation is itself systematic.
`BY MR. GOLDENBERG:
` Q. Well, what do you mean by
`implementation? The reason I ask is that part
`of the implementation for any parity-check
`matrix is picking a particular generator, so if
`that's what you mean by implementation -- is
`that what you meant by implementation?
` MR. ROSATO: Objection, form.
` THE WITNESS: I mean, I guess I take
` some issue with what you're saying, that an
` implementation requires picking a
` generator. You know, I think we've
` established that for low density
` parity-check matrix there exists
` generators, but I'm not clear that any
` given implementation has to utilize or have
` a generator matrix in order to construct or
` create code words.
`BY MR. GOLDENBERG:
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 30
`
`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
`
` Michael Mitzenmacher, PhD
` Q. I may have muddied the waters by
`guessing at what you meant by implementation.
` A. Okay.
` Q. And it seems like I guessed wrong.
`Let me just ask the question again.
` A. Okay.
` Q. For any given LDPC matrix there
`exists a generator matrix for the same code
`that is a systematic generator, correct?
` A. I believe that's the case for, you
`know, the general class of low density parity
`check codes that are -- that have valid sense
`of constraints, yeah.
` Q. Do you have the LUBY98 -- you don't
`have the prior references in your binder there,
`do you?
` A. No.
` MR. GOLDENBERG: Can I have your
` copy of the exhibit back. I just want to
` make a note on it.
` MR. ROSATO: Counsel, excuse me, I
` don't mean to interrupt.
` MR. GOLDENBERG: Let's go off the
` record.
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 31
`
`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
`
` Michael Mitzenmacher, PhD
` MR. ROSATO: Okay.
` (Discussion off the record.)
` (Exhibit Apple 1004 for
` IPR2017-00210, marked.)
`BY MR. GOLDENBERG:
` Q. Dr. Mitzenmacher, you've been handed
`an exhibit. It's marked Apple 1004, and it
`also says IPR2017-00210. Do you see that?
` A. Yes.
` Q. And do you recognize -- let me pause
`for a second.
` This exhibit that you have in front
`of you has been filed in multiple IPRs. It may
`have different exhibit numbers of the different
`six IPRs, but at least for IPR2017-00210, this
`document was Exhibit 1004. Do you understand?
` A. Okay.
` Q. And do you recognize the document
`you have in front of you?
` A. Yeah, it appears to be the --
`looking at the table of content for, "The
`Thirtieth Annual ACM Symposium on Theory of
`Computing," and then a paper on that conference
`which I am a coauthor.
`
`TSG Reporting - Worldwide 877-702-9580
`
`
`
`Page 32
`
`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
`
` Michael Mitzenmacher, PhD
` Q. Can we refer today to this exhibit
`that was Exhibit 1004 and IPR2017-00210, can we
`refer to that document as LUBY98?
` A. All right.
` Q. You'll know what I mean when I say
`LUBY98?
` A. I'll try to keep that in mind.
` Q. So Luby98 describes some LDPC codes,
`correct?
` A. Yes.
` Q. Every one of the codes described in
`LUBY98 has a corresponding systematic generator
`matrix, correct?
` A. We don't describe general matrices.
`There are various issues with your question.
`First, we tend to define families of code,
`although we did do experiments on some specific
`codes, so I think that would apply only to the
`extent that you're talking about the specific
`codes that we tested, but also we don't -- to
`the extent it applies, it would apply only to
`those and certainly not to the families that
`we're describing. A
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
