I/P Engine, Inc. v. AOL, Inc. et al

Filing 892

Reply to Motion re 820 MOTION for Judgment as a Matter of Law Defendants Renewed Motion For Judgment As A Matter Of Law On Invalidity Or New Trial filed by AOL Inc., Gannett Company, Inc., Google Inc., IAC Search & Media, Inc., Target Corporation. (Noona, Stephen)

UNITED STATES DISTRICT COURT
EASTERN DISTRICT OF VIRGINIA
NORFOLK DIVISION
I/P ENGINE, INC.

Plaintiff,
Civil Action No. 2:11-cv-512

v.
AOL, INC., et al.,

Defendants.

REPLY BRIEF IN SUPPORT OF DEFENDANTS’ RENEWED MOTION FOR
JUDGMENT AS A MATTER OF LAW ON INVALIDITY OR NEW TRIAL
As Plaintiff’s Opposition confirms, there are only three basic ways in which Plaintiff
seeks to distinguish the asserted claims from the Culliss reference (on anticipation) or the other
prior art references (on obviousness). First, Plaintiff cites its expert’s testimony that Culliss does
not engage in content analysis because the content portion of Culliss’ scores will supposedly get
“swamped” or diluted to insignificance by the feedback portion. This allegation is simply
irrelevant. Even if the content portion of Culliss’ scores was diluted to insignificance, it still
exists, and that is all that is required for anticipation as a matter of law.
Second, Plaintiff argues that Culliss does not “filter” for relevance to the query because
Culliss’ filtering is supposedly not enabled. But Plaintiff bears the burden of proving that
Culliss’ filtering is not enabled, and Plaintiff has failed to carry that burden. Plaintiff merely
cites its expert’s testimony about two supposed deficiencies with Culliss’ filtering – one of which
has no support in Culliss itself and the other of which shows only that Culliss’ filtering could be
imperfect. But perfect efficacy is not required to show enablement as a matter of law.
Third, Plaintiff’s only argument for distinguishing the asserted claims from Defendants’
obviousness references is to say that Defendants’ references did not use the search query when
filtering. In the words of Plaintiff’s expert, these references “tossed” search results “over the
wall” to a content/collaborative filter, but did not toss the search query over the wall. Thus, even

00197.52009/5168680.1

1

under Plaintiff’s view of the art, the only obviousness question posed for the Court is whether it
would be obvious to toss the search query over the wall along with the search results.1
Obviousness is a question of law, and (as explained below) it would be utterly obvious to toss the
search query over the wall – i.e., use the search query when filtering. Indeed, the Asserted
Patents themselves acknowledge that conventional prior art search engines used the query when
filtering. (DX-001 at 2:15-18.) Thus, using the query when filtering is not innovative at all.
I.

THERE WAS NO LEGALLY SUFFICIENT BASIS TO FIND THAT CULLISS
FAILS TO ANTICIPATE THE ASSERTED CLAIMS
A.

Plaintiff’s Arguments for Distinguishing Culliss Rest on Dr. Carbonell’s
Unsupported Opinions, Rather Than the Text of Culliss Itself

With respect to Culliss, Defendants’ Rule 50(b) Motion and Plaintiff’s Opposition show
a consistent pattern – Defendants rely on the text of Culliss, while Plaintiff relies almost
exclusively on the unsupported opinions of its validity expert, Dr. Carbonell. For example,
Plaintiff argues that Culliss does not disclose content analysis because Dr. Carbonell says that it
does not. (Opp., 7 (“Dr. Carbonell thus refuted Defendants’ assertion that Culliss contained
content analysis sufficient to meet the claim element.”)) Similarly, Plaintiff argues that Culliss’
filtering disclosure will not work because Dr. Carbonell says that it will not. (Id., 8 (quoting Dr.
Carbonell’s testimony).) As discussed below, however, Dr. Carbonell’s testimony on these
points runs contrary to the plain text of Culliss itself. Even at the JMOL stage, Plaintiff cannot
distinguish Culliss based on bare expert testimony that defies the plain language of Culliss

1

The technique of using the search query when filtering is reflected in the independent
claims of the ‘420 Patent (which require filtering “for relevance to the query”) and the
independent claims of the ‘664 Patent (which require receiving information found to be relevant
“to the query” by other users). Plaintiff does not dispute that Defendants’ obviousness
references disclose all the dependent claims limitations, nor does Plaintiff dispute that all the
dependent claim limitations are also found in Culliss.

00197.52009/5168680.1

2

itself.2 See PharmaStem Therapeutics, Inc. v. ViaCell, Inc., 491 F.3d 1342, 1361 (Fed. Cir.
2007) (directing JMOL of invalidity, despite expert testimony to the contrary, where the expert
testimony “cannot be reconciled . . . with the prior art references themselves”)
B.

The Only Claim Elements Allegedly Missing From Culliss Were “Content
Analysis” and “Filtering” for Relevance to the Query

Plaintiff argues that Defendants’ Rule 50(b) Motion does not address the actual claim
limitations that the jury found lacking from Culliss. Plaintiff argues that Defendants instead
focus on whether Culliss discloses “content analysis” and “filtering for relevance to the query.”
(See Opp., 5.) As an initial matter, Defendants’ Rule 50(b) Motion expressly incorporated their
Rule 50(a) Motion, which provided an element-by-element analysis of how Culliss discloses all
claim elements. (D.N. 776, 3-11; D.N. 821, 4.) While the Rule 50(b) Motion itself did focus on
“content analysis” and “filtering,” that is because Dr. Carbonell stated at trial that these are the
two things allegedly absent from Culliss:
Q: Now, on your direct examination I think you said that you thought Culliss
didn’t disclose filtering and didn’t disclose content-based analysis; is that right?
A: That’s right.
(Trial Tr. 1924:10-13.) Right after this exchange, Plaintiff’s counsel told the Court that Dr.
Carbonell “talked about two things that were missing and he didn’t talk about anything else in
the reference.” (Id., 1924:17-18.) To be sure, Dr. Carbonell opined that Culliss’ lack of “content
analysis” and “filtering” rendered several claim limitations unmet – but his only stated basis to
2

Plaintiff cites Kinetic Concepts, Inc. v. Smith & Nephew, Inc., 688 F.3d 1342 (Fed. Cir.
2012) for the proposition that Dr. Carbonell’s testimony must be accepted as credible simply
because the jury adopted his conclusions. (See Opp., 4.) But Kinetic actually held that plaintiff’s
expert testimony supported a verdict of non-obviousness given that the expert’s testimony was
“consistent with the [prior art] reference” and corroborated by statements by the prior art
inventor. See Kinetic, 688 F.3d at 1363, 1364-65. Kinetic nowhere suggested that expert
testimony which defied the plain text of a prior art reference would be sufficient to validate a
patent over that reference.

00197.52009/5168680.1

3

find these limitations unmet was Culliss’ alleged lack of content analysis and filtering. Thus, if
Culliss does disclose content analysis and filtering, then Dr. Carbonell has provided no other
rationale for distinguishing Culliss from the asserted claims.
C.

Culliss Discloses Content Analysis

As explained in Defendants’ Opening Brief, Culliss discloses content analysis because an
article’s key term score in Culliss can be initially set by counting how often terms from the
user’s query appear in the article. (DX-058, 14:34-36.) These initial content-based scores are
then altered based on user feedback (id., 4:37-49), so that the final scores are a combination of
the initial content-based analysis and the subsequent feedback-based adjustments.
Neither Plaintiff nor Dr. Carbonell disputes this basic functionality of Culliss. Dr.
Carbonell merely opined that Culliss’ initial content score plays “virtually” no role when it
comes time to serve articles based on their scores (Trial Tr. 1859:2-4) because the feedbackbased adjustments allegedly dilute the initial content-based portion of the scores to
insignificance. Dr. Carbonell explained that the initial content-based score gets “swamped” by
the feedback-based adjustments (id., 1932:14-15), so that “for all practical purposes, [Culliss] is
not content-based filtering.” (Id., 1932:25-1933:2 (emphasis added).) Dr. Carbonell gave the
specific hypothetical example of an article whose score is initialized at 1 based on content and
then clicked on 1 million times. (Id. 1932:12-21.) According to Dr. Carbonell, the initial
content-based score simply “doesn’t matter” in this situation. (Id., 1932:15.)
However, Dr. Carbonell’s position that the content portion of Culliss’ scores gets diluted
to insignificance and thus “doesn’t matter” is not a legally valid basis to find that the content
portion is absent from Culliss. As the Federal Circuit held in SmithKline Beecham Corp. v.
Apotex Corp., 403 F.3d 1331, 1345 (Fed. Cir. 2005), a prior art reference will be deemed to
contain a claimed compound for purposes of anticipation even if the reference contains only
00197.52009/5168680.1

4

“trace amounts” of the compound. Here, even if the content-based portion of Culliss’ scores
could be diluted and “swamped” by the feedback-based adjustments, the final scores will still
contain at least “trace amounts” of the initial content analysis. Under SmithKline, this compels a
finding that the content analysis is still present for purposes of anticipation.
Plaintiff does not directly dispute this point. Plaintiff does not even address SmithKline,
despite Defendants’ reliance on SmithKline in their Opening Brief. Instead, Plaintiff cites Dr.
Carbonell’s testimony that Culliss’ content analysis occurs during a preliminary “initialization”
stage of Culliss’ system. As Plaintiff states: “Dr. Carbonell distinguished initialization from the
actual operation of the claimed invention, and explained that, while initialization may consider
content, ‘[t]he initial operation of the system is purely collaborative, pure profile, pure
popularity-based system, and that’s what governs.’” (Opp., 6-7 (Plaintiff’s emphasis).)
But the fact that Culliss’ content scores are generated at an “initialization” stage is
irrelevant, because these initial content-based scores are the same scores that are then altered by
feedback during the operation of Culliss’ system. This is precisely how Culliss’ scores come to
embody the claimed combination of content and feedback. (See DX-058, 14:34-36 (scores
initialized based on content); id., 4:37-49 (scores altered based on feedback).) Therefore,
according to Dr. Carbonell’s own testimony, the score for a given article in Culliss consists of
the article’s initial content score plus any feedback-based adjustments to this score generated
from user selections of the article. (Trial Tr. 1858:18-1859:17.)
Thus, Dr. Carbonell’s testimony provided no sufficient basis to find that Culliss lacks
content analysis. Even if the jury credited Dr. Carbonell’s position that Culliss’ content analysis

00197.52009/5168680.1

5

occurs at an initialization stage and then gets diluted to insignificance, that is insufficient as a
matter of law to conclude that the requisite content analysis is absent from Culliss.3
D.

Culliss Discloses “Filtering” for Relevance to the Query

As explained in Defendants’ Opening Brief, Culliss discloses filtering for relevance to
the query in the embodiment where articles’ key terms include “rating” key terms like X-rated,
G-rated, etc. (DX-058, 11:8-12:41.) Like the other key term scores, the rating key term scores
can be initially set by content analysis (id., 14:23-25) and then altered by user feedback. (Id.,
11:47-51.) And these rating key term scores are used to filter the articles for relevance to the
user’s query: For example, articles with an X-rated key term score above a certain threshold will
be filtered out and not displayed to searchers who entered G-rated queries. (Id., 12:1-5.)4
In arguing that this embodiment does not disclose the requisite “content-based filtering,”
Plaintiff cites Dr. Carbonell’s testimony for the proposition that “[i]n the ‘X-rated filtering’ addon part, [] the system does not perform an analysis of the content of an item at all – the scores in
this part are entirely based on user feedback.” (Opp., 8.) But this defies the plain language of
Culliss, which states that the rating key term scores are initialized by content analysis just like
every other key term score: “Initially, the key terms, category key terms and rating key terms
may be associated with words or other information in the article . . . . For example, the scores can
be initially set to correspond with the frequency of the term occurrence in the article.” (DX-058,
3

Plaintiff’s invalidity position is also inconsistent with its infringement position. For
invalidity, Plaintiff argues that Culliss’ content scores are irrelevant because they might be only a
very small portion of the overall scores given to Culliss’ articles. But for infringement, Plaintiff
argues that Google meets the “content” limitations by merely using content-based templates to
look up corresponding feedback scores – even though the scores themselves are generated
entirely by feedback and have no content portion at all. (See Trial Tr. 533:1-16; 538:24-539:10.)
4

Culliss knows whether a given user’s query is G-rated or X-rated because the system
“would permit or require the user to enter a rating key term in the search query. The invention
would operate in a similar manner for the rating key terms as described above for the key terms
alone . . . .” (DX-058, 11:40-43.)

00197.52009/5168680.1

6

14:23-25, 14:34-36 (emphasis added).) Again, Plaintiff cannot distinguish Culliss based on
expert testimony that defies the plain language of Culliss itself. PharmaStem, 491 F.3d at 1361.5
Plaintiff next argues that Culliss’ X-rated filtering is not enabled because it is “not even a
workable filtering.” (Opp., 8.) As recounted in Defendants’ Opening Brief, and as Plaintiff does
not dispute in its Opposition, Dr. Carbonell gave two reasons why this X-rated filtering is
supposedly not workable: (1) he said that some G-rated searchers might have to view X-rated
articles before the articles could be labeled as X-rated and filtered out of subsequent G-rated
search results (Trial Tr. 1864:8-11); and (2) he said that Culliss may improperly label articles as
X-rated just because adults like them. (Id., 1863:12-1864:7.)
But Dr. Carbonell’s Point #1 is legally irrelevant. The fact that some G-rated searchers
might have to view X-rated articles before those articles can be filtered out of subsequent Grated search results simply means that Culliss’ X-rated filtering is not perfect. That does not
mean that the filtering is non-enabled. See Decca Ltd. v. United States, 544 F.2d 1070, 1077 (Ct.
Cl. 1976) (“The mere fact that the system has some drawbacks, or that under certain postulated
conditions it may not work . . . does not detract from the operability of the disclosed equipment
to perform its described function.”).6
Meanwhile, Dr. Carbonell’s Point #2 is utterly unsupported by Culliss itself. As
discussed above, articles get X-rated scores in Culliss’ system through a combination of two
5

Plaintiff states that Defendants ignored Dr. Carbonell’s position that Culliss’ X-rated
filtering does not use content and only addressed Dr. Carbonell’s position that Culliss’ X-rated
filtering is unworkable. (Opp., 8.) Defendants’ Opening Brief focused on the “unworkability”
point instead of the “no content” point because Dr. Carbonell’s position that the X-rated
embodiment does not use content is derivative of his position that Culliss’ overall system does
not use content. Nonetheless, as discussed above, Dr. Carbonell’s position that the X-rated
embodiment does not use content defies the plain language of Culliss.
6

By analogy, Plaintiff argued at summary judgment that the patented system might not
work “perfectly,” stating: “the system may not always be perfect at identifying a user’s exact
information need.” (D.N. 427 at 20 (emphasis in original).)

00197.52009/5168680.1

7

factors: having X-rated terms in their content (DX-058, 14:23-25) and being clicked on by users
who entered X-rated terms in their search queries. (Id., 11:39-12:26.) Nowhere does Culliss
state that an article will be deemed X-rated just because adults like the article. As noted above,
Plaintiff cannot defeat invalidity JMOL by proffering expert testimony that has no support in the
applicable prior art reference itself. PharmaStem, 491 F.3d at 1361.
Also, because Culliss is a patent, there is a presumption that Culliss is enabled. Plaintiff
bears the burden of persuading the Court otherwise. Amgen Inc. v. Hoechst Marion Roussel,
Inc., 314 F.3d 1313, 1355 (Fed. Cir. 2003) (“[A]n accused infringer should be similarly entitled
to have the district court presume the enablement of unclaimed (and claimed) material in a prior
art patent,” and the plaintiff bears the burden of rebutting this presumption by “present[ing]
evidence of nonenablement that a trial court finds persuasive.”) As stated above, Dr. Carbonell’s
arguments about the supposed unworkability of Culliss are far from “persuasive” and cannot
carry Plaintiff’s burden of proving that Culliss’ filtering is so deficient as to be non-enabled.7
E.

Plaintiff Cannot Defeat Invalidity JMOL Merely Because Culliss was Before
the PTO During the Prosecution of the Patent-in-Suit

At various points in its Opposition, Plaintiff references the fact that Culliss was before
the PTO during prosecution of the Asserted Patents. (See Opp., 3, 10-11.) Of course, there are
innumerable court decisions invalidating patents (at both JMOL and summary judgment) based

7

Plaintiff’s non-enablement argument also rests on an improper legal standard.
Specifically, Plaintiff argues that Culliss is non-enabled “because Culliss does not adequately
teach how to make and use the filter required by the asserted claims.” (Opp., 9.) But a prior art
reference can be enabled for anticipation purposes even if it is not adequately described for one
to use it. Rasmusson v. SmithKline Beecham Corp., 413 F.3d 1318, 1325-26 (Fed. Cir. 2005).
“The reason is that section 112 provides that the specification must enable one skilled in the art
to ‘use’ the invention whereas [section] 102 makes no such requirement as to an anticipatory
disclosure.” Id. at 1325 (emphasis added). Thus, even if there was any factual support (which
there is not) for Dr. Carbonell’s position that Culliss’ X-rated filtering is so flawed as to be
unusable, this filtering is still adequate for anticipation.

00197.52009/5168680.1

8

on prior art that was before the PTO during prosecution.8 Indeed, “there is no heightened burden
of proof when a reference was previously considered by the PTO.” Sciele Pharma Inc. v. Lupin
Ltd., 684 F.3d 1253, 1260 (Fed. Cir. 2012). Thus, Plaintiff cannot defeat JMOL merely by
arguing that Culliss was before the PTO during prosecution.
This is particularly true given that Culliss appears only in a list of references considered
by the PTO; there is no explanation in the prosecution history as to why the Asserted Patents are
patentable over Culliss. Furthermore, given the ex parte nature of prosecution, the PTO did not
have the benefit of an adversarial party to argue for why Culliss invalidates the Asserted Patents.
Nor did the PTO have the benefit of Plaintiff’s infringement theory, which broadly interpreted
the Asserted Patents to merely require filtering items based on scores that include a combination
of content and feedback. Amazon.com, Inc. v. Barnesandnoble.com, Inc., 239 F.3d 1343, 1351
(Fed. Cir. 2001) (“A patent may not, like a ‘nose of wax,’ be twisted one way to avoid
anticipation and another to find infringement.”). For all these reasons, the fact that Culliss was
before the PTO during prosecution is no grounds to deny anticipation JMOL based on Culliss.
II.

THERE WAS NO LEGALLY SUFFICIENT BASIS TO FIND THE ASSERTED
CLAIMS NON-OBVIOUS
A.

The Obviousness Art Discloses Virtually Every Claimed Element

Before addressing the difference between the asserted claims and Defendants’
obviousness references – and whether this difference would be obvious to overcome – it is
important to clarify what Plaintiff cannot dispute. Plaintiff cannot dispute that Defendants’

8

See, e.g., PharmaStem, 491 F.3d at 1366-67 (ordering JMOL of obviousness based on
prior art considered by the PTO during prosecution); Liebel-Flarsheim Co. v. Medrad, Inc., 481
F.3d 1371, 1381 (Fed. Cir. 2007) (affirming summary judgment of anticipation based on prior art
considered by the PTO during prosecution); Celeritas Techs., Ltd. v. Rockwell Int’l Corp., 150
F.3d 1354, 1360-61 (Fed. Cir. 1998) (ordering JMOL of anticipation based on prior art
considered by the PTO during prosecution).

00197.52009/5168680.1

9

obviousness references combine content-based filtering with collaborative filtering. For
example, the Fab reference states: “By combining both collaborative and content-based filtering
systems, Fab may eliminate many of the weaknesses found in each approach.” (DX-050, 66.)
The WebHound reference states: “This thesis claims that content-based filtering and automated
collaborative filtering are complementary techniques, and the combination of ACF with some
easily extractable features of documents is a powerful information filtering technique.” (DX049, Abstract.) And the Rose reference states: “The prediction of relevance is carried out by
combining data pertaining to the content of each item of information with other data regarding
correlations of interests between users . . . . The user correlation data is obtained from feedback
information provided by users when they retrieve items of information.” (DX-034, Abstract.)
Nor can Plaintiff dispute that these references perform content/collaborative filtering on
search results. For example, WebHound states that “a WEBHOUND like front-end to a popular
search engine such as Lycos, could enable users with WEBHOUND accounts to filter the results
of their searches on the extensive databases complied by these search engines in a personalized
fashion.” (DX-049, 78.) Rose states that “[t]he relevance predicting technique of the present
invention . . . can be employed to filter messages provided to a user in an electronic mail system
and search results obtained through an on-line text retrieval service.” (DX-034, 2:51-55.)9
B.

The Only Element Allegedly Missing from the Obviousness Art is Using the
Search Query When Content/Collaborative Filtering Search Results

Plaintiff argues that these obviousness references are nonetheless lacking because they do
not use the search query when performing their content/collaborative filtering on search results.10

9

This disclosure of how Rose can “filter . . . search results” occurs in Rose’s Summary
of the Invention section, not “Rose’s background section” as Plaintiff alleges. (See Opp., 14.)
10

At various points in its Opposition and at trial, Plaintiff also referred to using the
search query when filtering as “tightly integrating” search with filtering. (See, e.g., Opp.,14
00197.52009/5168680.1

10

For example, as for WebHound, Plaintiff cites Dr. Carbonell’s testimony that search results are
“thrown over the wall” to WebHound’s content/collaborative filter, but that this filter does not
consider the search query. (Opp., 13-14.) Similarly, as for Rose, Plaintiff argues that “any
subsequent filtering allegedly disclosed in Rose fails to filter items for ‘relevance to the query’
because that filtering does not consider the query at all.” (Id., 15.) Thus, the obviousness
question in this case is as follows: would it be obvious to use the search query when performing
content/collaborative filtering on search results generated through that query? This obviousness
question is a question of law that is the Court’s province to answer. Soverain Software LLC v.
Newegg Inc., -- F.3d --, 2013 WL 216406 (Fed. Cir. Jan. 22, 2013) (reversing district court’s
judgment of non-obviousness and finding patent obvious as a matter of law).
C.

It Would Be Obvious to Use the Search Query When Performing
Content/Collaborative Filtering on Search Results

The Court should answer this obviousness question in the affirmative: it would be
obvious to use the search query when performing content/collaborative filtering on search
results. For one thing, the PTO found that all the search-related elements of the ‘420 Patent –
performing filtering on search results and using the search query when filtering – were obvious
improvements over a system that just performed content/collaborative filtering of information
with no search elements at all. The PTO made this decision when rejecting the ‘420 Patent as
obvious over the ‘799 Patent – a predecessor patent that taught content/collaborative filtering but
did not teach any search limitations. (Opening Br., 25-29; DX-004, IPE2092-93.) While the
Court excluded testimony about this double-patenting rejection from reaching the jury, the

(arguing that Rose “does not disclose a tightly integrated search system that could filter items for
relevance to a query”); Trial Tr. 1888:10-12 (“I myself tried and did not succeed in arriving at
this kind of tight integration in performing all these operations with respect to the query.”).)

00197.52009/5168680.1

11

double-patenting rejection is still part of the trial record. Thus, the Court can and should
consider it in deciding the legal question of whether the claims are obvious.
The PTO’s obviousness opinion is supported by the testimony of Defendants’ expert (Dr.
Ungar), and by sheer common sense. As Dr. Ungar pointed out at trial: “[I]f you are filtering
search results, it’s obvious to keep around the query and use that also for filtering. . . . [J]ust
think about it. If you ask a query of a search engine, you get a result, you just have the query
sitting there with the result, why not use that also for filtering.” (Trial Tr. 1317:25-1318:7.)
Critically, Plaintiff’s position that it would be non-obvious to use the search query when
filtering is belied by the Asserted Patents themselves. The Asserted Patents’ shared specification
states that “conventional search engines initiate a search in response to an individual user’s query
and use content-based filtering to compare the query to accessed network informons.” (DX-001,
2:15-18 (emphasis added).) Thus, using the query when filtering was in the prior art, according
to the Asserted Patents. This admission is dispositive – Plaintiff cannot allege that using the
query when filtering is a non-obvious technique when the Asserted Patents admit that this
technique was in the prior art. Constant v. AMD, Inc., 848 F.2d 1560, 1570 (Fed. Cir. 1988) (“A
statement in a patent that something is in the prior art is binding on the applicant and patentee for
determinations of anticipation and obviousness”); PharmaStem, 491 F.3d at 1362 (same).
Culliss also shows how using the query for filtering was known in the prior art, because
Culliss uses the search query when scoring and filtering items. (Opening Br., 17-18.) Plaintiff
argues that Defendants cannot cite Culliss for obviousness because Dr. Ungar asserted Culliss as
an anticipation reference instead of an obviousness reference. (Opp., 12.) But Dr. Ungar opined
that any elements missing from his asserted obviousness references would be obvious to supply
from the prior art as a whole. (Trial Tr. 1311:6-14.) This would include Culliss, which was

00197.52009/5168680.1

12

explicitly referenced during Dr. Ungar’s obviousness testimony (id., 1309:6-1310:2) and was
later discussed extensively during his anticipation testimony. See Johns Hopkins Univ. v.
CellPro, Inc., 152 F.3d 1342, 1357 fn. 21 (Fed. Cir. 1998) (“[A]nticipation is the epitome of
obviousness.”) Moreover, Culliss is indisputably part of the trial record, and thus it should be
considered by the Court in examining the legal issue of whether the asserted claims are obvious.
D.

The Level of Ordinary Skill in the Art Further Shows Obviousness

In the face of all this obviousness evidence, Plaintiff relies heavily on Dr. Carbonell’s
testimony that search and filtering were different fields at the time of the Asserted Patents, so
that one of ordinary skill in the art would supposedly be unable to apply the techniques of one
field to the other. (See Opp., 14, 21.) But the agreed definition for a person of ordinary skill in
the art (POSITA) in this case was merely “an individual with a bachelor’s degree in computer
science with at least 2 years of experience in the field.” (D.N. 789, 8, 9.) This definition does
not suggest that the POSITA would be in the filtering field alone (and thus supposedly ignorant
of search techniques) or in the search field alone (and thus supposedly ignorant of filtering
techniques). Thus, there is no basis for Dr. Carbonell’s position that the POSITA could not
understand both search and filtering techniques. See also KSR Intern. Co. v. Teleflex Inc., 550
U.S. 398, 417 (2007) (“When a work is available in one field of endeavor, design incentives and
other market forces can prompt variations of it, either in the same field or a different one.”)
(emphasis added).
As noted above, the binding admissions in the Asserted Patents also refute Dr.
Carbonell’s position that the prior art did not know how to apply search techniques to filtering.
And Dr. Carbonell’s own trial metaphor also refutes his position. According to Dr. Carbonell,
prior art references such as WebHound and Rose ran a search and “tossed” the search results
“over the wall” to a profile-based filtering system, but did not toss the search query over the wall
00197.52009/5168680.1

13

so it could be used by the filtering system. (See Trial Tr. 1834:9-1835:2; 1874:2-17; 1880:1015.) Thus, the supposed “wall” between search and filtering had already been surmounted in the
prior art, because prior art references like WebHound and Rose were able to send search results
“over the wall” with ease. Bridging the gap between the obviousness art on the one hand and the
Asserted Patents on the other would merely require tossing the search query over the wall (along
with the search results) so that the filtering system could use the query when filtering. Noninnovative techniques such as “throwing the query over the wall” are precisely the type of
patents that the Supreme Court held should be found obvious in KSR.
Contrary to Plaintiff’s position, the jury interrogatory responses are not probative of nonobviousness on this point. The jury found that Defendants’ obviousness references “do not
disclose a tightly integrated search system and could not filter information relevant to a query.”
(D.N. 789, 8, 9.) But “tightly integrated” (or any variant thereof) appears nowhere in the
asserted claims. And as for filtering information relevant to the query, none of the interrogatory
responses purported to answer the pivotal legal question of whether this technique – i.e.,
throwing the query over the wall – would be obvious to perform. Nor could they, as this
question is the Court’s to answer. Soverain, 2013 WL 216406, at *2. As discussed above,
performing this technique in the context of the claims would simply require using the search
query for filtering – a technique that was in the prior art according to the Asserted Patents.
E.

No Secondary Considerations Can Rebut the Obviousness Showing

Given the prima facie case of obviousness, Plaintiff bears the burden of rebutting the
obviousness case by proving secondary considerations of non-obviousness. Muniauction, Inc. v.
Thomson Corp., 532 F.3d 1318, 1327 (Fed. Cir. 2008). Where a prima facie case of obviousness
is strong, evidence of secondary considerations must likewise be strong to overcome the prima
facie case – thus, even “substantial evidence” of secondary considerations will often be
00197.52009/5168680.1

14

insufficient to overcome a strong prima facie case of obviousness. Leapfrog Enters., Inc. v.
Fisher-Price, Inc., 485 F.3d 1157, 1162 (Fed. Cir. 2007). Here, Plaintiff failed to produce
sufficient evidence to overcome the very strong prima facie obviousness case.
(a)

No Evidence of Copying

Plaintiff argues that the jury’s finding of copying was supported by Plaintiff’s argument
and evidence that “Google did not present any documents to explain how the system was
developed, there was no Google development story, and the ‘420 patent was cited on one of
Google’s patents prior to development of the infringing system.” (Opp., 29-30.) But this
provides no basis to find copying. For example, the mere citation of the ‘420 Patent in a Google
patent is no evidence that Google copied the ‘420 Patent. If anything, the fact that the PTO
allowed the Google patent over the ‘420 Patent suggests that the PTO thought the Google patent
was an improvement over the ‘420 Patent, not a copy of it. Nor does Plaintiff’s innuendo that
“there was no Google development story” show copying. It is Plaintiff’s burden to prove
copying, and Plaintiff cannot meet this burden by merely claiming that Google did not produce
evidence of how it developed its system. Finally, Plaintiff’s allegation that “Google’s system . . .
incorporated all features of the asserted claims” (Opp., 25) does not show copying – if it did,
then any plaintiff could prove copying by merely proving infringement. This is not the law.
Tokai Corp. v. Easton Enter., Inc., 632 F.3d 1358, 1370 (Fed. Cir. 2011) (even “[a] stipulation of
infringement, taken alone, is not probative of copying”).
(b)

Inconsistent Findings on Independent Invention and
Unsuccessful Attempts by Others

Plaintiff argues that the “scanning system” in the ‘420 Patent, a limitation not present in
the ‘664 Patent, explains the jury’s inconsistent findings as to the independent invention of the
two patents. (Opp., 26-27.) But Plaintiff made identical non-obviousness arguments for both

00197.52009/5168680.1

15

patents, referring to them collectively as “the Lang and Kosak invention.” (Trial Tr. 1875:1023.) Plaintiff presents no evidence or explanation for how the “scanning system” limitation
justified the jury’s finding that the ‘664 invention was independently invented by others while
the ‘420 invention was not. For example, there is no evidence that any third-party invention
contained all the ‘664 limitations yet lacked the ‘420 “scanning system” limitation.
Plaintiff again cites the scanning system limitation to explain why the jury could have
found “unsuccessful attempts by others” for the ‘420 Patent but not for the ‘664 Patent. (Opp.,
24-25.) Yet Plaintiff does not even try to explain why the presence of the scanning system in the
‘420 Patent but not the ‘664 Patent justifies a finding that others had unsuccessfully tried to
create the ‘420 invention but not the ‘664 invention.
(c)

No Evidence of Acceptance by Others as Shown by Praise or
Licensing of the Claimed Invention

Plaintiff’s only cited evidence for “acceptance by others” is the fact that other patents cite
the ‘420 Patent. (Opp., 26.) But Plaintiff cannot argue that a bald citation on the face of a patent
constitutes “praise from others in the field.” Accordingly, Plaintiff presented no evidence of
“[a]cceptance by others . . . as shown by praise from others in the field or from licensing of the
claimed invention.” (D.N. 789, 9-10.)
(d)

No Evidence of Commercial Success

In support of its assertion that the patented invention “has proven a commercial success,”
Plaintiff cites Dr. Carbonell’s testimony about “modern search engines, Google included, that
use the teachings of these patent claims.” (Opp., 23.) But Plaintiff presented no evidence that
Google’s search engine or any other search engine practices the patents, much less that their
success is attributable to the patents. (Trial Tr. 1886:25-1887:5.) In fact, Plaintiff stated at trial
that search was not accused and not relevant to the case. (Id., 116:7-12.) Furthermore, Dr.

00197.52009/5168680.1

16

Carbonell admitted that he did not know any details about which aspects of Google’s systems
were accused of infringement. (Id., 1889:20-1890:10.) Thus, he could not and did not opine on
what portion of Google’s commercial success is attributable to the alleged infringement.
Further, contrary to Plaintiff's assertion, Plaintiff's damages expert did not testify that
“inclusion of the infringing technology increased revenues by 20-40% for Google.” (Opp. at
23). Rather, he testified that Google’s implementation of the entire Smart Ads system
supposedly led to a 20% increase in revenue. (Trial Tr. 902:11-17.) He then agreed that
Google’s advertising revenue is generated from innumerable technologies that are not
encompassed by the asserted patents (see, e.g., id., 918:8-23; 919:4-13) and admitted that he did
not evaluate whether the patented technology is the basis for customer demand for the entirety of
the accused systems. (Id., 919:14-22.) Thus, his testimony is of no import.
(e)

No Evidence of Other Secondary Considerations

Plaintiff also argues that “the long-felt need for better search results existed and was
‘recognized even in the cited prior art.’” (Opp., 23-24 (citing Trial Tr. 1887:13-17).) But as
Plaintiff itself admitted, “this case is not about search engine results.” (Trial Tr. 107:21.) Both
Google and other pre-existing search engines were already satisfying the need for high-quality
search results, and, again, Plaintiff does not even accuse Google’s search engine of infringement.
(Id., 116:7-12.) Thus, the supposed long-felt need for “better search results” had already been
met, and could not have been met by the Asserted Patents in any event.
As to “unsuccessful attempts by others,” Plaintiff cites Dr. Carbonell’s conclusory
statements about the supposed failure of himself and others. (Opp., 24.) But Dr. Carbonell
provides no details about how he (or others) tried and failed to come up with the patented
invention. He did not describe when he did the work, what work he completed, or why he failed.
Thus, his conclusory statements on this point carry little to no weight. Brooke Group Ltd. v.
00197.52009/5168680.1

17

Brown & Williamson Tobacco Corp., 509 U.S. 209, 242 (1993) (“When an expert opinion is not
supported by sufficient facts to validate it in the eyes of the law . . . it cannot support a jury’s
verdict.”); Ashland Oil, Inc. v. Delta Resins & Refractories, Inc., 776 F.2d 281, 294 (Fed. Cir.
1985) (“Lack of factual support for expert opinion going to factual determinations [] may render
the testimony of little probative value in a validity determination.”). Plaintiff’s position appears
to be that it can prove a secondary consideration of non-obviousness by having a single witness
state, with no detail, that he failed to create the patented invention. This position is implausible
and insufficient as a matter of law.
Plaintiff next argues that the Asserted Patents achieved “unexpected” results because they
achieved “better” results than the prior art. (Opp., 25.) Initially, Plaintiff did not provide
evidence of “better” results and Plaintiff points to nothing beyond unsubstantiated conclusions.
Further, Plaintiff must show that its invention is an unexpected improvement over the prior art,
not simply any improvement. “An applicant may make this showing with evidence that the
claimed invention exhibits some superior property or advantage that a person of ordinary skill in
the relevant art would find surprising or unexpected.” In re Mayne, 104 F.3d 1339, 1343 (Fed.
Cir. 1997). Plaintiff presented no evidence for why the one alleged improvement of the Asserted
Patents – using the search query when filtering search results – was surprising or unexpected.
III.

IN THE ALTERNATIVE, DEFENDANTS ARE ENTITLED TO A NEW TRIAL
ON ANTICIPATION AND OBVIOUSNESS
A.

The Double-Patenting Exclusion Warrants a New Trial on Obviousness

Plaintiff argues that the Court properly excluded DDX 3.35—a demonstrative showing
the PTO’s obviousness-type double patenting rejection of the ‘420 Patent—and Dr. Ungar’s
testimony about DDX 3.35 under Fed. R. Evid. 402 and 403. However, the double-patenting
rejection is both highly relevant and non-prejudicial, and its exclusion warrants a new trial.

00197.52009/5168680.1

18

As to relevance, it is undisputed that a patent’s prosecution history should be considered
when determining obviousness. (Opening Br., 27.) Plaintiff does not rebut this point, but rather
admits that it successfully used portions of the prosecution history “to show the conclusions
reached by the patent office” on the issue of invalidity. (Opp., 29.) Yet Plaintiff argues that
Defendants’ attempt to use the prosecution history for invalidity was improper.
In doing so, Plaintiff cites cases addressing a different question—whether ongoing
reexaminations are probative of invalidity. (Opp., 27 & n.13.) But the reason why these cases
give little weight to ongoing reexaminations is because they are incomplete and thus the final
result is uncertain. Transam. Life Ins. Co. v. Lincoln Nat'l Life Ins. Co., 597 F. Supp. 2d 897,
906 (N.D. Iowa 2009) (“Federal district courts have . . . conclude[d] that evidence of incomplete
patent reexamination proceedings are inadmissible at trial on patent infringement claims and
invalidity defenses.”) (emphasis added). This is not the case with the double-patenting rejection
that Defendants sought to introduce. The PTO’s determination that the ‘420 Patent11 was
obvious over a content/collaborative filtering system was never withdrawn or backtracked from.
The inventors got around the rejection by filing the terminal disclaimer, but the obviousness
finding by the PTO remained. And the prosecution history at issue is closed and final.
Furthermore, admission of the double-patenting rejection would not be misleading or
unfairly prejudicial as Plaintiff suggests. (Opp., 28.) Plaintiff's argument that a terminal
disclaimer does not constitute acquiescence by the patentee to the PTO’s obviousness finding is a
red herring. The excluded slide specifically and accurately showed the PTO’s conclusion that
combining a content/collaborative filter with a search engine was obvious. This is not
misleading because it accurately recites what occurred during prosecution and it places the ‘420
11

Defendants have explained how this finding is equally applicable to the ‘664 Patent, a
point which Plaintiff does not dispute. (Opening Br., 28-29.)

00197.52009/5168680.1

19

Patent’s allowance into its proper and complete context. Plaintiff’s position that the jury should
not be informed about the circumstances of the Patent’s allowance is a more certain way to
mislead the jury, as the jury could only be left with the false impression that the PTO did not find
the Patent to be an obvious improvement over a content/collaborative filtering system. Indeed,
this was precisely the false impression that was created when Plaintiff was allowed to introduce
portions of the prosecution history supporting its invalidity position (Trial Tr. 1487:11-1490:10)
while Defendants were not allowed to do the same through the double-patenting rejection.
Moreover, contrary to another of Plaintiff’s arguments, the jury is fully capable of
understanding the non-binding effect of PTO actions, as juries are tasked with understanding this
principle every time invalidity is an issue in a patent case. Further, the jury had already been
informed about the back-and-forth nature of patent prosecution through the video the Court
showed at the beginning of trial, and Plaintiff would have been free to bring out the rationale
behind filing a terminal disclaimer on cross-examination or through rebuttal testimony.
Finally, because of the importance of the double-patenting rejection, the Court’s error
was highly prejudicial. Plaintiff asserts that because the Court allowed a discussion of the
terminal disclaimer, somehow this rendered any error harmless. (Opp., 29.) But Plaintiff does
not explain how Defendants could have placed the terminal disclaimer in context without being
able to discuss the PTO’s obviousness decision that caused the terminal disclaimer to be filed.
Because the exclusion of DDX 3.35 and Dr. Ungar’s related testimony constituted
prejudicial error, Defendants are entitled to a new trial on obviousness.
DATED: February 15, 2013

00197.52009/5168680.1

/s/ Stephen E. Noona
Stephen E. Noona
Virginia State Bar No. 25367
KAUFMAN & CANOLES, P.C.
150 West Main Street, Suite 2100
Norfolk, VA 23510

20

Telephone: [phone]
Facsimile: [phone]
[e-mail]
David Bilsker
David A. Perlson
QUINN EMANUEL URQUHART &
SULLIVAN, LLP
50 California Street, 22nd Floor
San Francisco, California 94111
Telephone: [phone]
Facsimile: [phone]
[e-mail]
[e-mail]
Counsel for Google Inc., Target Corporation,
IAC Search & Media, Inc., and Gannett Co., Inc.

/s/ Stephen E. Noona
Stephen E. Noona
Virginia State Bar No. 25367
KAUFMAN & CANOLES, P.C.
150 W. Main Street, Suite 2100
Norfolk, VA 23510
Telephone: [phone]
Facsimile: [phone]
[e-mail]
Robert L. Burns
FINNEGAN, HENDERSON, FARABOW, GARRETT &
DUNNER, LLP
Two Freedom Square
11955 Freedom Drive
Reston, VA 20190
Telephone: [phone]
Facsimile: [phone]
Cortney S. Alexander
FINNEGAN, HENDERSON, FARABOW, GARRETT &
DUNNER, LLP
3500 SunTrust Plaza
303 Peachtree Street, NE
Atlanta, GA 94111
Telephone: [phone]
Facsimile: [phone]
Counsel for Defendant AOL Inc.
00197.52009/5168680.1

21

CERTIFICATE OF SERVICE
I hereby certify that on February 15, 2013, I will electronically file the foregoing with the
Clerk of Court using the CM/ECF system, which will send a notification of such filing (NEF) to
the following:
Jeffrey K. Sherwood
Kenneth W. Brothers
DICKSTEIN SHAPIRO LLP
1825 Eye Street NW
Washington, DC 20006
Telephone: [phone]
Facsimile: [phone]
[e-mail]
[e-mail]
Donald C. Schultz
W. Ryan Snow
Steven Stancliff
CRENSHAW, WARE & MARTIN, P.L.C.
150 West Main Street, Suite 1500
Norfolk, VA 23510
Telephone: [phone]
Facsimile: [phone]
[e-mail]
[e-mail]
[e-mail]
Counsel for Plaintiff, I/P Engine, Inc.

/s/ Stephen E. Noona
Stephen E. Noona
Virginia State Bar No. 25367
KAUFMAN & CANOLES, P.C.
150 West Main Street, Suite 2100
Norfolk, VA 23510
Telephone: [phone]
Facsimile: [phone]
[e-mail]

12214084v1
00197.52009/5168680.1

22