`
` UNITED STATES PATENT AND TRADEMARK OFFICE
` BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
` __________________________
` :
` FACEBOOK, INC., LINKEDIN :
` CORP., and TWITTER, INC., :
` :
` Petitioners : Case
` : Nos. IPR2013-00478
` v. : IPR2013-00479
` : IPR2013-00480
` SOFTWARE RIGHTS ARCHIVE, : IPR2013-00481
` LLC, :
` Patent Owner :
` __________________________:
`
` Volume 2 of 2
`
` VIDEOTAPE DEPOSITION of PAUL S. JACOBS,
` Ph.D., taken before Ryan K. Black, RPR, CLR
` Notary Public, in and for the District of
` Columbia, at the offices of Rothwell, Figg,
` Ernst & Manbeck, P.C., 607 14th Street, N.W.,
` Washington, D.C., on Wednesday, August 6, 2014,
` commencing at 9:06 a.m.
`
` Veritext Legal Solutions
` Mid-Atlantic Region
` 1250 Eye Street, NW - Suite 1201
` Washington, D.C. 20005
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`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
`
`Facebook, Inc. EXHIBIT 1030
`
`001
`
`
`
`Page 255
`
` A P P E A R A N C E S:
`
` ROTHWELL, FIGG, ERNST & MANBECK, P.C.
` By: MARTIN ZOLTICK, ESQ.
` MICHAEL JONES, ESQ.
` 607 14th Street, N.W., Suite 800
` Washington, D.C. 20005
` 202.783.6040
` Representing - Software Rights Archive, LLC
`
` DINOVO PRICE ELLWANGER & HARDY LLP
` By: VICTOR G. HARDY, ESQ.
` MITCH YANG, ESQ.
` 7000 North MoPac Expressway, Suite 350
` Austin, Texas 78731
` 512.539.2626
` Representing - Software Rights Archive, LLC
`
` KEKER & VAN NEST
` By: DAVID J. SILBERT, ESQ.
` 633 Battery Street
` San Francisco, California 94111
` 415.676.2261
` Representing - LinkedIn Corp. & Twitter, Inc.
`
` COOLEY LLP
` By: LOWELL MEAD, ESQ.
` 3175 Hanover Street
` Palo Alto, California 94304
` 650.849.5734
` Representing - Facebook, Inc.
`
` ALSO PRESENT:
`
` Edward A. Fox, Ph.D
` Amy Langville, Ph.D.
` Jason Levin, Legal Videographer
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`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
`
`Facebook, Inc. EXHIBIT 1030
`
`002
`
`
`
`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 256
`
` I N D E X
`
` TESTIMONY OF: PAUL S. JACOBS, Ph.D
`
` BY MR. SILBERT.............................257
`
` BY MR. MEAD................................293
`
` BY MR. HARDY...............................389
`
` - - -
`
` E X H I B I T S
`
` - - -
`
` NUMBER DESCRIPTION MARKED
`
` Petitioner
`
` No. 1036 a document titled Text-based
`
` Intelligence Systems..............257
`
` Ptnt Owner
`
` No. 2116 an uncertified rough draft
`
` Transcript of Dr. Jacobs'
`
` deposition testimony given on
`
` August 5, 2014, morning session...388
`
` Ptnt Owner
`
` No. 2117 an uncertified rough draft transcript
`
` of Dr. Jacobs' deposition testimony
`
` given on August 5, 2014...........403
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`003
`
`
`
`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 257
`
` THE VIDEOGRAPHER: This begins Volume
`
` 2 in the deposition of Paul Jacobs. The witness
`
` has been previously sworn in. We're going on
`
` record at 9:06 a.m.
`
` CROSS EXAMINATION (Cont'd)
`
` BY MR. SILBERT:
`
` Q. Good morning, Dr. Jacobs.
`
` MR. SILBERT: I'm going to ask the
`
` court reporter to mark this document as Exhibit
`
` 1036.
`
` (Petitioner's Deposition Exhibit No.
`
` 1036, a document titled Text-Based Intelligence
`
` Systems, was marked.)
`
` BY MR. SILBERT:
`
` Q. And, Dr. Jacobs, please review Exhibit
`
` 1036 as much as you'd like, but my question,
`
` when you're through reviewing it is, do you
`
` recognize this exhibit as a copy of the book
`
` that you edited, titled Text-Based Intelligence
`
` Systems? And then the subtitle is, Current
`
` Research and Practice in Information Extraction
`
` and Retrieval.
`
` A. Yes. I haven't reviewed the whole
`
` thing, but -- but, yes, I recognize it.
`
` Q. Okay. Could I ask you to please
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`004
`
`
`
`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 258
`
` -- well, strike that.
`
` This book was published in 1992,
`
` correct?
`
` A. That's the -- that's the copyright
`
` date on the copyright page. I'm not sure if
`
` that actually means, necessarily, that it was
`
` published in that year, but -- but that's the
`
` copyright date.
`
` Q. Do you know when it was published?
`
` A. No. I'm not certain. And the reason
`
` I'm not certain whether it's 1991 or 1992 is,
`
` it's possible that it actually went to press in
`
` 1991, with a 1992 copyright date.
`
` Q. If it was -- if it went to press in
`
` 1992, do you know the month in 1992 that it went
`
` to press?
`
` A. I -- I don't recall. That was 22
`
` years ago, and I really don't recall exactly
`
` which month it would have been.
`
` Q. The book collects papers that were
`
` presented at a symposium in 1990; is that
`
` correct?
`
` A. I recall that this book was based
`
` on a AAAI symposium that I chaired, that I
`
` believe was in the AAAI spring symposium series.
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`005
`
`
`
`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 259
`
` And to the best of my recollection, that
`
` symposium was held in the spring of 1990.
`
` Q. Yeah. I'm just looking at the -- if
`
` you look at the preface, which has the Page
`
` Roman Numeral vii, small vii, it's after the
`
` table of contents, it says, this volume started
`
` with a symposium in 1990, sponsored by AAAI,
`
` entitled, Text-Based Intelligence Systems.
`
` Do you see that?
`
` A. Yes.
`
` Q. Okay. Would you please turn to Page
`
` 127?
`
` Are you there?
`
` A. Mm-hmm.
`
` Q. Do you see the title of this paper is,
`
` Text Retrieval and Inference?
`
` A. Text Retrieval and Inference, yes.
`
` Q. And it's got two authors. One is W.
`
` Bruce Croft, and the other is Howard R. Turtle;
`
` is that correct?
`
` A. Yes.
`
` Q. And it says that Mr. Turtle worked for
`
` West Publishing Company, correct?
`
` A. That's what it says, yes.
`
` Q. And West Publishing Company is the
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`006
`
`
`
`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 260
`
` company that publishes the Westlaw legal
`
` database, correct?
`
` A. Yes. But you said is, so I assume
`
` you meant at the time of this publication that
`
` that's what it would have been called.
`
` Q. Yes. Okay. Thank you. I had -- that
`
` is right.
`
` A. Right.
`
` Q. Thanks for that clarification.
`
` Yes. That at the -- at -- at the time
`
` of this publication, West Publishing Company
`
` published the Westlaw legal database; is that
`
` correct?
`
` A. That's my understanding, yes.
`
` Q. Okay. Would you please look at Page
`
` 130?
`
` Are you there?
`
` A. Yes.
`
` Q. I'm going to start reading from the
`
` first full paragraph on that page. It says,
`
` within the information retrieval community, a
`
` number of techniques have been developed that
`
` can represent the content of documents
`
` and -- and information needs. These
`
` representations have a much different flavor
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`007
`
`
`
`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 261
`
` than NLP representations. They are generally
`
` based on simple, very general features of
`
` documents; e.g., words, citations, and represent
`
` simple relationships between features; e.g.,
`
` phrases, and between documents; e.g., two
`
` documents cite the same document.
`
` And I'll stop reading there.
`
` Now, two documents citing the same
`
` document, that's the relationship known as
`
` bibliographic coupling, correct?
`
` A. Yes. In an abstract sense, if two
`
` documents cite the same document, they are
`
` bibliographically coupled.
`
` Q. Okay. And in the language of the
`
` patents, that's an indirect relationship between
`
` documents, correct?
`
` A. No. That's not really correct.
`
` Q. Why not?
`
` A. Because in the language of the patent,
`
` indirect relationships is used in claims that
`
` have other specific language related to objects
`
` in databases.
`
` Q. Okay. And do you recall in the
`
` patent -- the patents, and we can get them out,
`
` but they -- they provide a figure where they
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`008
`
`
`
`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 262
`
` show a number of different relationships between
`
` documents?
`
` I believe it's --
`
` A. Yeah. Why don't you show me
`
` specifically what we're discussing?
`
` Q. Okay. Do you have any of the -- any
`
` of the patents in front of you?
`
` A. I believe the answer is, no, because
`
` my --
`
` Q. Oh, we had the water spillage.
`
` A. Yeah. We had the water damage
`
` yesterday.
`
` Let's see. No. No. No. I do have
`
` another. I think Dr. Fox gave me another copy.
`
` So I -- I have a copy -- oh, this is a different
`
` one. Yeah, I have a copy of the '494 patent.
`
` Q. Okay. Would you please look at Figure
`
` 6 of the '494 patent, which is titled, Schematic
`
` Representations of the 18 Primary Patterns?
`
` Do you have that in front of you?
`
` A. Yes.
`
` Q. Now, the second -- no, it
`
` would be -- do you see one of these that
`
` corresponds to bibliographic coupling?
`
` A. No.
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`009
`
`
`
`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 263
`
` Q. Okay. Is it -- is it your opinion
`
` that bibliographic coupling between documents is
`
` not an indirect relationship between documents,
`
` as the patent uses the term indirect
`
` relationship?
`
` MR. HARDY: Objection; form.
`
` THE WITNESS: No. That's not my
`
` opinion.
`
` BY MR. SILBERT:
`
` Q. Okay. So then do you think that
`
` bibliographic coupling between documents is an
`
` indirect relationship between documents, as the
`
` patent uses the term indirect relationship?
`
` A. Okay. So your question says, as the
`
` patent uses the term indirect relationship, --
`
` Q. Well, we --
`
` A. -- so should we look at a particular
`
` claim where that's used so I can be more
`
` specific about it?
`
` Q. I don't mean to interrupt. If
`
` you'd like, you can look at the -- the PTAB's
`
` claim construction, which is relation -- of
`
` indirect relationships, which is relationships,
`
` where at least one intermediate object exists
`
` between two objects, and where the intermediate
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`0010
`
`
`
`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 264
`
` objects connect the two objects through a chain
`
` of citations.
`
` A. Okay. So which -- which institution
`
` decision are we talking about?
`
` Q. Well, this one --
`
` A. You're referring to the claim
`
` construction in the institution decision?
`
` Q. In the '352, --
`
` A. The '352 decision?
`
` Q. -- yeah, but I think it's the same.
`
` Actually, in the '494, it appears to
`
` be slightly different.
`
` A. Mm-hmm.
`
` Q. It's relationships that are
`
` characterized by at least one intermediate node
`
` between two nodes, so under either of those
`
` constructions.
`
` A. Either?
`
` And, again, are we talking about
`
` -- you said, as used in the patent, so can we
`
` refer to a specific claim?
`
` Q. I'm referring to the claim
`
` construction. I'm referring to the term as
`
` it's been construed by the P -- PTAB.
`
` A. Independent of -- you -- you did say
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`0011
`
`
`
`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 265
`
` the way it's used in the patent.
`
` MR. HARDY: Objection.
`
` BY MR. SILBERT:
`
` Q. My question is -- my question is,
`
` referring to the claim constructions of indirect
`
` relationship that I've just recited for you, is
`
` bibliographic coupling an indirect relationship
`
` between documents?
`
` MR. HARDY: Objection; form.
`
` THE WITNESS: You haven't given me
`
` enough information to say.
`
` BY MR. SILBERT:
`
` Q. Why not?
`
` A. Because, first of all, my opinions
`
` are based on the way the terms are used in the
`
` patent. And, second, the Board's construction
`
` refers to objects, in the case of '352, and
`
` nodes, in the case of the other example you gave
`
` me, which I believe was the '494.
`
` And bibliographic coupling, as I tried
`
` to explain in my earlier answer is -- is an ab
`
` -- describes an abstract relationship that could
`
` be just referring to the bibliographies
`
` of paper documents. So there's not enough
`
` information.
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`0012
`
`
`
`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 266
`
` Q. Okay. So your testimony is that,
`
` if two documents are bibliographically coupled,
`
` you're unable to tell whether that means they
`
` have an indirect relationship with each other
`
` under the PTAB's constructions of indirect
`
` relationship?
`
` MR. HARDY: Objection; form.
`
` Objection; scope.
`
` THE WITNESS: I said I would be able
`
` to if you gave me enough information.
`
` BY MR. SILBERT:
`
` Q. But they -- but -- but with the
`
` information that the two documents are
`
` bibliographically coupled with each other,
`
` you're unable to determine whether they're
`
` indirectly related under the PTAB's
`
` constructions of indirect relationship?
`
` THE WITNESS: Can you read back the
`
` question, please?
`
` THE REPORTER: Sure.
`
` (Referred-to testimony read back.)
`
` THE WITNESS: What do you mean by
`
` unable?
`
` BY MR. SILBERT:
`
` Q. Well, I thought you told me you
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`0013
`
`
`
`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 267
`
` couldn't answer the question.
`
` A. I said that I could not answer
`
` the question because there was not enough
`
` information.
`
` Q. Okay. Would you -- do you still have
`
` the '494 patent in front of you?
`
` A. Yes.
`
` Q. Would you turn to Column 14, please?
`
` Do you have that in front of you?
`
` A. Yes.
`
` Q. Would you look at Line 27, Pattern
`
` Number 2, it says, A cites C and B cites C.
`
` Do you see that?
`
` A. Yes.
`
` Q. Does that correspond to bibliographic
`
` coupling?
`
` A. Yes.
`
` Q. Okay. I'll state for the record,
`
` I -- it appears to me there might be a typo
`
` in Figure 6 of this patent, which I had never
`
` noticed before.
`
` But, in any event, would you please
`
` now turn back to Exhibit 1036, the copy of your
`
` book, and back to Page 130, which is the page we
`
` were looking at before?
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`0014
`
`
`
`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 268
`
` So just to get our frame of reference,
`
` again, I'll -- I'll, again, read the text that
`
` I'm focusing on. It's the first full paragraph
`
` on Page 130. It says, within the information
`
` retrieval community, a number of techniques have
`
` been developed that can represent the content of
`
` documents and information needs. These
`
` representations have a much different flavor
`
` than NLP representations. They are generally
`
` based on simple, very general features of
`
` documents; e.g., words, citations, and represent
`
` simple relationships between features; e.g.,
`
` phrases, and between documents; e.g., two
`
` documents cite the same document.
`
` Do you -- do you agree, based
`
` on reading this passage, that, as of at
`
` least 1992 when your book was published, using
`
` bibliographic coupling for search was nothing
`
` new?
`
` A. No.
`
` Q. Why not?
`
` A. So that's a very broad que --
`
` question. Would you like me to focus on this
`
` paragraph of my book, in particular?
`
` Q. Sure.
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`0015
`
`
`
`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 269
`
` A. Okay. So what -- this paragraph
`
` discusses techniques used in the information
`
` retrieval community. It's from the book, as
`
` we've discussed, that presents leading-edge
`
` research, at least as of 1990, in information
`
` retrieval. And -- and that's very broad-based.
`
` As you can see, it discusses Natural
`
` Language Processing and a variety of other
`
` techniques. My own research area was Natural
`
` Language Processing.
`
` And this discusses document
`
` representations. There is no question that
`
` the use of citations, and even co-citations, for
`
` example, as -- in -- in the general area of
`
` document representations, had been explored
`
` between at least -- well, in the case of
`
` bibliographic coupling, 1963; in the case of
`
` co-citations, at least 1972.
`
` So those concepts as concepts in
`
` the area of document representation, which is
`
` certainly a fundamental component of information
`
` retrieval, had been in existence for quite some
`
` time, at the time of publication of this, 27
`
` years.
`
` So this -- this paragraph does explain
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`0016
`
`
`
`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 270
`
` that the -- the use of bibliographic coupling,
`
` for example, in information retrieval, the part
`
` of information retrieval concerned with document
`
` representations was -- was not new. But that
`
` does not say that those representations were
`
` used in search, per se. It doesn't say anything
`
` about that.
`
` Q. Okay. Are -- are you familiar
`
` with this article that the -- that this
`
` quotation is from? I mean, you obviously read
`
` it some time ago. Have you read it recently?
`
` A. Okay. There were multiple questions
`
` in there. You want to take it apart so that I
`
` can answer?
`
` Q. Sure. How familiar are you, as you
`
` sit here today, with this Croft and Turtle
`
` article that we're looking at?
`
` A. I am not very familiar with it
`
` today, because I do not recall having reviewed
`
` this particular article since I read it in
`
` preparation for the publication of my book,
`
` which, as you saw, was around 1991.
`
` Q. Do you -- is it your testimony that a
`
` person of ordinary skill in the art, reading the
`
` passage that we've read, would not understand
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`0017
`
`
`
`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 271
`
` that a document representation, such as is -- is
`
` described here, could be used for search?
`
` A. Okay. I -- in order to answer
`
` that question, I really have to respond to two
`
` different aspects of the question. One is, can
`
` document representations be used for search?
`
` And the second aspect is, if there is research
`
` in a particular document representation, does
`
` that mean that that representation could be used
`
` for search.
`
` So with respect to the first part,
`
` one of ordinary skill would understand, in
`
` general -- I'm not sure how much this paragraph
`
` contributes, but one of ordinary skill would
`
` understand that document representations, in
`
` general, could be used for search.
`
` Now, that said, there was a great
`
` deal of research on document representations,
`
` and this mentions a number of different
`
` representations, and certainly implies that
`
` Natural Language Processing, for example, can
`
` be used as a way of contributing to a document
`
` representation -- now -- as well as some of
`
` these other methods, such as citations and
`
` phrases, which are characterized here as being
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`0018
`
`
`
`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 272
`
` separate Natural Language Processing.
`
` Now, the -- the fact that something
`
` can be used as a document representation,
`
` in theory, means that that might be used for
`
` search. But one of ordinary skill would
`
` understand that the results of using such a
`
` representation would depend on the quality of
`
` the representation and on empirical results in
`
` terms of whether it worked or not.
`
` So in the case, for example, of some
`
` of these advanced methods that we were exploring
`
` -- exploring, like, Natural Language Processing,
`
` a lot of the data indicated that, even if
`
` high-quality, high-accuracy representations were
`
` available using NLP, it -- it would likely fail
`
` to contribute to the empirical results and
`
` search accuracy. And the same happened to be
`
` true, as Dr. Fox's work showed, in the area of
`
` representations such as co-citation.
`
` So I don't think this particular
`
` paragraph sheds any light on whether a -- a
`
` particular representation should be used for
`
` search, and, at -- at most, in the abstract,
`
` says that -- that -- that these areas of
`
` different representations of documents have
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`0019
`
`
`
`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 273
`
` been explored.
`
` Q. Okay. So you -- in your opinion, this
`
` passage doesn't say anything about which is the
`
` best type of representation to use for search.
`
` Is that fair?
`
` A. Or even which ones would be effective
`
` or useful at all.
`
` Q. Okay. But -- but you agree
`
` that this passage does teach that a document
`
` representation that includes bibliographic
`
` coupling relationships has been used for search?
`
` A. No.
`
` Q. Why not?
`
` A. This paragraph says that within the
`
` information retrieval community, techniques for
`
` document representations, such as representing
`
` simple relationships between documents, such as
`
` two documents cite the same document, have been
`
` developed.
`
` So it -- it does say that the
`
` technique of representing relationships,
`
` such as bibliographic coupling in the
`
` information retrieval community had been
`
` developed. That does not say what those
`
` representations had been used for, and, in
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`0020
`
`
`
`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 274
`
` particular, does not say that they had been used
`
` for search. It doesn't say that co-citation has
`
` been used for search in this particular
`
` paragraph. So one of ordinary skill would not
`
` understand from this paragraph that it says
`
` that.
`
` Q. Okay. But it does say, you agree,
`
` that one representation form of a document is
`
` to include bibliographic coupling information in
`
` the representation, correct?
`
` THE WITNESS: Could you read back the
`
` question, please?
`
` THE REPORTER: Sure.
`
` (Referred-to testimony read back.)
`
` THE WITNESS: It's a little bit
`
` convoluted, but -- but, basically, it does
`
` say that you can include bibliographic coupling
`
` information in a document representation.
`
` BY MR. SILBERT:
`
` Q. Okay. Would you look down a few
`
` paragraphs to the one that starts, given?
`
` That sentence says, given the availability of a
`
` number of representation techniques that capture
`
` some of the meaning of a document or information
`
` need, our -- our basic premise is that decisions
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`0021
`
`
`
`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
`
` about which documents match an information need
`
` should make use of as many of the representation
`
`Page 275
`
` forms as practical.
`
` A. Mm-hmm.
`
` Q. Do you see that?
`
` A. Yes.
`
` Q. Doesn't that teach that what these
`
` authors are proposing is that people should
`
` use as many of the representation forms,
`
` including -- as possible, including the ones
`
` that they have just itemized above?
`
` A. No.
`
` Q. Why not?
`
` A. Because --
`
` THE WITNESS: Can you read back the
`
` question, please?
`
` THE REPORTER: Sure.
`
` (Referred-to testimony read back.)
`
` THE WITNESS: Okay. There
`
` are at least two problems with the way you
`
` characterized this paragraph. First is that,
`
` you're assuming the representation forms must
`
` include all of the representation forms that are
`
` mentioned elsewhere in this document. I don't
`
` know that that's correct.
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`0022
`
`
`
`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 276
`
` And the second problem is that, you
`
` used the word possible. And this specifically
`
` says, our premise is that decisions about
`
` what -- which documents match an information
`
` need should make use of as many of the
`
` representation forms as practical.
`
` Now, as I explained at the time of
`
` publication of this article, there's a lot of
`
` research that's reported here that -- that, in
`
` some ways, may have looked promising, but in the
`
` case of an information retrieval -- a particular
`
` information retrieval system, would have viewed
`
` as being impractical.
`
` So, for example, my own research
`
` area was semantic interpretation. There were
`
` elements of semantic interpretation that might
`
` have been useful and practical in information
`
` retrieval systems, but there were many others
`
` that would have been viewed as -- as not helpful
`
` and impractical. And, as I've said, many of
`
` these citation techniques could also be
`
` characterized as not helpful and impractical.
`
` So in the context of this
`
` statement, one of ordinary skill would, under
`
` no circumstances, have -- have assumed that it
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`0023
`
`
`
`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 277
`
` would have been practical to use, for example,
`
` bibliographic coupling. And so it really
`
` doesn't -- it really doesn't teach what you --
`
` what you said, in terms of using as many as
`
` possible at all. It teaches something contrary
`
` to that. The word very practical is different
`
` from possible.
`
` BY MR. SILBERT:
`
` Q. Can you go back to the paragraph
`
` above, the one we were reading previously, that
`
` refers to the representations that represent
`
` simple relationships between features and
`
` between documents; e.g., two documents cite to
`
` the same document?
`
` Do you have that language in front of
`
` you?
`
` A. I -- I -- I have the paragraph in
`
` front of me. I don't believe you read it
`
` correctly.
`
` Q. Yeah. Okay.
`
` The -- the next sentence says,
`
` the focus here is on simple but general
`
` representations that can be applied to most
`
` texts, rather than on specialized techniques,
`
` which capture more information, but are
`
`VERITEXT NATIONAL COURT REPORTING COMPANY
`215-241-1000 ~ 610-434-8588 ~ 302-571-0510 ~ 202-803-8830
`
`Facebook, Inc. EXHIBIT 1030
`
`0024
`
`
`
`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 278
`
` applicable only in narrow contexts.
`
` Do you see that?
`
` A. Yes.
`
` Q. Do you understand that to say that
`
` the representations that are described in the
`
` preceding sentence or sentences are viewed by
`
` these authors as simple, general representations
`
` that can be applied to most texts?
`
` THE WITNESS: Could you read back