Ex Parte Steck

Patent Trial and Appeal Board

Dec 29, 2016

13916132 (P.T.A.B. Dec. 29, 2016)

United States Patent and Trademark Office
UNITED STATES DEPARTMENT OF COMMERCE
United States Patent and Trademark Office
Address: COMMISSIONER FOR PATENTS
P.O.Box 1450
Alexandria, Virginia 22313-1450
www.uspto.gov
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR ATTORNEY DOCKET NO. CONFIRMATION NO.
13/916,132 06/12/2013 Harald STECK NETF/0080US (060340) 1066
108911 7590 01/03/2017
Arte.ois T aw firm in T T P / Netflix
EXAMINER
7710 Cherry Park Drive Suite T #104
Houston, TX 77095
TESFAYE, AKLIL M
ART UNIT PAPER NUMBER
2423
NOTIFICATION DATE DELIVERY MODE
01/03/2017 ELECTRONIC
Please find below and/or attached an Office communication concerning this application or proceeding.
The time period for reply, if any, is set in the attached communication.
Notice of the Office communication was sent electronically on above-indicated "Notification Date" to the
following e-mail address(es):
algdocketing @ artegislaw. com
kcruz @ artegislaw.com
mmccauley @ artegislaw.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte HARALD STECK
Appeal 2016-004014
Application 13/916,132
Technology Center 2400
Before JOSEPH L. DIXON, JAMES R. HUGHES, and ERIC S. FRAHM,
Administrative Patent Judges.
HUGHES, Administrative Patent Judge.
DECISION ON APPEAL
STATEMENT OF THE CASE
Appellant appeals under 35 U.S.C. § 134(a) from a rejection of claims
1—20. We have jurisdiction under 35 U.S.C. § 6(b).
We affirm-in-part.
The invention relates to “techniques for promoting new media titles to
targeted audiences” (Spec. 11). Claim 1, reproduced below, is illustrative of
the claimed subject matter:
1. A computer-implemented method comprising:
Appeal 2016-004014
Application 13/916,132
determining scores representing similarities between a
given first data type item and each of a plurality of other first data
type items as represented in a first latent space;
generating a representation of the given first data type item
in a second latent space by at least one of (1) averaging
representations in the second latent space of a given number of
the plurality of other first data type items associated with the
highest similarity scores, and (2) solving an objective function to
obtain a representation of a dummy user who has viewed the
given number of the plurality of other first data type items
associated with the highest similarity scores;
determining scores representing similarities between the
given first data type item and a plurality of users as represented
in the second latent space; and
directing content that promotes the given first data type
item to one or more of the users based on the scores representing
similarities between the given first data type item and the users.
REFERENCES
The prior art relied upon by the Examiner in rejecting the claims on
appeal is:
Weare
Li
Fleischman
Nice
Fleischman
US 2010/0083318 A1
US 2012/0239645 A1
US 2013/0018896 A1
US 2013/0218907 A1
US 2013/0305282 A1
Apr. 1,2010
Sept. 20, 2012
Jan. 17, 2013
Aug. 22, 2013
Nov. 14, 2013
REJECTIONS
The Examiner made the following rejections:
Claims 1,2, 7—9, and 14—16 stand rejected under 35 U.S.C. § 103(a)
as being unpatentable over Fleischman ‘282 and Nice.
2
Appeal 2016-004014
Application 13/916,132
Claims 3, 5, 10, 12, 17, and 19 stand rejected under 35 U.S.C.
§ 103(a) as being unpatentable over Fleischman ‘282, Nice, and Li.
Claims 4, 11, and 18 stand rejected under 35 U.S.C. § 103(a) as being
unpatentable over Fleischman ‘282, Nice, Li, and Weare.
Claims 6, 13, and 20 stand rejected under 35 U.S.C. § 103(a) as being
unpatentable over Fleischman ‘282, Nice, and Fleischman ‘896.
ANALYSIS
Claims 1, 6—8, 13—15, and 20
Appellant contends Fleischman ‘282 fails to teach “determining
scores representing similarities between a given first data type item and each
of a plurality of other first data type items,” as recited in claim 1 (App. Br.
13—15). We are not persuaded by Appellant’s arguments.
Fleischman ‘282 describes a way of providing a targeted
advertisement to a user based on the user’s likely interests on social media
(see Fleischman ‘282, || 5—6). This is accomplished by correlating a Web
ID of a user with one or more social media (SM) IDs of the user, correlating
social media content authored by the user with time-based events, for
example, a TV show, and finally, correlating an event with an advertisement
(Fleischman ‘282, 21—24, 28, 35—37). Fleischman ‘282 performs the
correlation between the Web ID and SM IDs “by matching uniform resource
locators (URLs) that appear in tracking cookies and also in SM content
items, and by also matching the times that those URLs appear” (Fleischman
‘282, | 111). “The more matches between a web ID and a SM ID, the
greater the confidence value. The confidence may be determined as any
numerical value (e.g., ranks, probabilities, percentages, real number
values).” (Fleischman, ‘282,1117). The Examiner finds Fleischman ‘282’s
3
Appeal 2016-004014
Application 13/916,132
correlation of Web ID to SM IDs teaches the claim 1 limitation of
“determining scores representing similarities between a given first data type
item and each of a plurality of other first data type items” (Final Act. 5; Ans.
3—4). We agree with the Examiner.
We are not persuaded by Appellant’s argument that Fleischman
‘282’s correlations are “between an identifier of one type (a web ID) and
identifiers of a different type (an SM ID)” (App. Br. 14; see also Reply Br.
4). The broadest reasonable interpretation of a “first data type item”
encompasses both a Web ID and an SM ID because they are both data
representing an identification of a user. Appellant has not shown claim 1,
read in light of the Specification, requires a more specific definition of a
“first data type item” that would preclude a finding that both Fleischman
‘282’s Web ID and SM ID are “first data type items.”
We are also not persuaded by Appellant’s argument that Fleischman
‘282 calculates “correlation scores by determining the probability of an
event. . . . However, the language of claim 1 explicitly requires scores that
represent similarities between a given first data type item and each of a
plurality of other first data type items. Importantly, probability and
similarity are very different concepts.” (App. Br. 14). Fleischman ‘282
matches a Web ID to a SM ID based on whether the Web ID user visited a
URL around the same time that the SM ID user authored social media
content containing the URL (see Fleischman ‘282, 13—15). A match is
thus based on the Web ID and SM ID having a URL in common, where “the
greater the number of matches that are detected between a web ID and a SM
ID, the more likely it is that the two are correlated” (Fleischman ‘282, |
116). We find Fleischman ‘282’s matches based on a common URL
accessed by a Web ID and posted by a SM ID meets the limitation of
4
Appeal 2016-004014
Application 13/916,132
“similarities” in claim 1. Fleischman, ‘282 further discloses that for each
Web ID and SM ID pair, a confidence value is determined based on the
number of matches (Fleischman, ‘282,1117). We find Fleischman ‘282’s
determining a confidence value for each Web ID and SM ID pair, based on
the number of matches between each pair, meets the claim 1 limitation of
“determining scores representing similarities between a given first data type
item and each of a plurality of other first data type items.” Nothing in claim
1 precludes the claimed “scores” from being probabilities, as Appellant
argues (App. Br. 14). And in any case, Fleischman ‘282 teaches the
confidence value is not limited to a probability, but can be any numerical
value, including a real number value (Fleischman, ‘282,1117).
We are, therefore, not persuaded the Examiner erred in rejecting claim
1, and claims 6—8, 13—15, and 20 not specifically argued separately.
Claims 2—5, 9—12, and 16 19
Appellant contends Nice fails to disclose “generating a first vector
which is an average of, or is obtained as a dummy user vector of, a given
number of vectors in the fifth matrix that represent the plurality of first data
type items associated with the highest similarity scores,” as recited in claim
2 (App. Br. 15—16).
In the Final Action, the Examiner does not provide a specific mapping
of the claimed “first vector” to any feature in Nice. The Examiner cites to
paragraphs 35 to 38 of Nice for disclosing the disputed limitation (Final Act.
7). However, the Examiner has not identified, and we have not found,
specific teachings in the cited portions of Nice that support the Examiner’s
findings.
5
Appeal 2016-004014
Application 13/916,132
In the Answer, the Examiner finds, citing paragraphs 36 to 40, that
“Nice reads on the claimed first vector which is the trait vector generated
and the dummy user vector of a given number of vectors which is the
selected set of P users from among the N users for determining trait vectors”
(Ans. 9). We do not agree with this finding, however, because claim 2
requires the “dummy user vector” be based on “a given number of vectors in
the fifth matrix that represent the plurality of first data type items associated
with the highest similarity scores” (emphasis added). That is, the recited
“dummy user vector” is based on item vectors, not user vectors as in the
Examiner’s findings regarding Nice. Accordingly, the Examiner’s findings
in the Answer are not commensurate with the limitations of claim 2.
We are, therefore, constrained by the record to find the Examiner
erred in rejecting claim 2, claims 9 and 16, which recite similar limitations to
claim 2, and claims 3—5, 10-12, and 17—19, which depend from claims 2, 9,
and 16.1 2
1 Independent claims 1, 8, and 15 each recite a similar limitation to the claim
2 limitation discussed above, namely, “generating a representation of the
given first data type item in a second latent space by at least one of (1)
averaging representations in the second latent space of a given number of the
plurality of other first data type items associated with the highest similarity
scores, and (2) solving an objective function to obtain a representation of a
dummy user . . . .” However, Appellant has not presented specific
arguments showing the combination of Fleischman ‘282 and Nice fails to
disclose the “generating” limitation of claims 1, 8, and 15. Therefore, we
make no finding as to whether the Examiner erred with respect to the
“generating” limitation in our decision to affirm the rejections of claims 1, 8,
and 15.
6
Appeal 2016-004014
Application 13/916,132
CONCLUSIONS
Under 35 U.S.C. § 103(a), the Examiner erred in rejecting claims 2—5,
9—12, and 16—19, but did not err in rejecting claims 1, 6—8, 13—15, and 20.
DECISION
For the above reasons, the Examiner’s rejections of claims 2—5, 9—12,
and 16—19 are reversed, and the Examiner’s rejections of claims 1, 6—8, IS­
IS, and 20 are affirmed.
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a)(l)(iv).
AFFIRMED-IN-PART
7