Ex Parte El-Charif et al

Patent Trial and Appeal Board

Jul 31, 2017

12694191 (P.T.A.B. Jul. 31, 2017)

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.
12/694,191 01/26/2010 Rami El-Charif P2395.10270US01 5526
132897 7590 08/02/2017
Maschoff Brennan/ PayPal
1389 Center Drive, Ste 300
Park City, UT 84098
EXAMINER
ROLAND, GRISELLE CORBO
ART UNIT PAPER NUMBER
2158
NOTIFICATION DATE DELIVERY MODE
08/02/2017 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):
docket @ mabr. com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte RAMI EL-CHARIF, SANJAY PUNDLKRAO GHAT ARE,
STEVEN CHEN, OLIVIER G. DUMON,
MUHAMMAD FAISAL REHMAN,
and
GUANGLIE SONG
Appeal 2017-003575
Application 12/694,191
Technology Center 2100
Before CARLA M. KRIVAK, HUNG H. BUI, and JON M. JURGOVAN,
Administrative Patent Judges.
KRIVAK, Administrative Patent Judge.
DECISION ON APPEAL
Appellants appeal under 35 U.S.C. § 134(a) from a Final Rejection of
claims 1—15, which are all the claims pending in the application. We have
jurisdiction under 35 U.S.C. § 6(b).
We affirm-in-part.
Appeal 2017-003575
Application 12/694,191
STATEMENT OF THE CASE
Appellants’ invention is directed to “methods and systems for
simulating a search to generate an optimized scoring function” for “ordering
the search results when presenting the search results to an end-user of a
computer-based trading or e-commerce application” (Title (capitalization
altered); Spec. 11).
Independent claim 1, and dependent claims 2 and 3, reproduced
below, are exemplary of the subject matter on appeal.
1. A computer-implemented method comprising:
with a processor-based data importing module, importing
data representing a plurality of search result sets resulting from
a plurality of executed search queries, each search result set
including a plurality of item listings satisfying a search query
from the plurality of executed search queries and a listing slot
identifier identifying a position within a search results page of a
particular item listing that has resulted in a transaction being
concluded, the listing slot identifier determined utilizing a
production scoring function specified with one or more
parameters that are weighted with a first set of one or more
weighting factors;
with a processor-based parameter optimization module,
processing the plurality of search result sets to derive a new
scoring function, the new scoring function having a second set
of one or more weighting factors for one or more parameters,
the second set of one or more weighting factors selected to
satisfy a set of constraints, the second set of one or more
weighting factors being different than the first set of one or
more weighting factors; and
with a processor-based scoring function evaluation
module, evaluating the performance of the new scoring function
relative to a production scoring function.
2. The computer-implemented method of claim 1,
wherein the second set of one or more weighting factors are
selected to maximize an average rank shift metric for the
2
Appeal 2017-003575
Application 12/694,191
plurality of search result sets, the average rank shift metric
determined by comparing the listing slot identifier as derived by
the production scoring function with a listing slot identifier as
derived by a new scoring function for each item listing that has
resulted in a transaction being concluded for each search result
set of the plurality of search result sets.
3. The computer-implemented method of claim 1,
further comprising:
filtering the plurality of search result sets resulting from
the plurality of executed search queries to select search result
sets having a predetermined number of item listings assigned to
a particular category of item listings, wherein processing the
plurality of search result sets to derive a new scoring function
includes processing the search result sets having the
predetermined number of item listings assigned to the particular
category.
REFERENCES and REJECTIONS
(1) The Examiner rejected claims 1, 5—8, and 12—14 under
35 U.S.C. § 103(a) based upon the teachings of Watson (US 7,444,327 B2;
issued Oct. 28, 2008) and Kim (US 8,938,463 Bl; issued Jan. 20, 2015).
(2) The Examiner rejected claims 2—4, 9-11, and 15 under
35 U.S.C. § 103 based upon the teachings of Watson, Kim, and Dumon (US
2010/0262602 Al; published Oct. 14, 2010).
ANALYSIS
Claims 1, 5—8, and 12—15
With respect to independent claim 1, Appellants contend the
Examiner erred in finding Kim discloses a “new scoring function having a
second set of one or more weighting factors for one or more parameters, the
second set of one or more weighting factors selected to satisfy a set of
3
Appeal 2017-003575
Application 12/694,191
constraints, the second set of one or more weighting factors being different
than the first set of one or more weighting factors” (Br. 16 (citing Kim col.
6,11. 25—65; col. 10,1. 45—col. 11,1. 14)). Rather, Appellants contend Kim
“simply teaches re-running a model,” and “does not disclose or suggest two
scoring functions (a production scoring function and a new scoring
function)” (Br. 16). Appellants additionally contend the Examiner’s
combination of Watson and Kim does not teach the claimed first and second
sets of weighting factors because Kim and Watson, individually or in
combination, “only teach a first set of weighting factors” (Br. 16). Lastly,
Appellants argue the Examiner’s combination of Kim and Watson does not
teach evaluating the performance of a new scoring function relative to a
production scoring function, as claimed; rather, combining Kim’s “teaching
of using ‘implicit user feedback’ with Watson would, at best, result in a
ranking system that uses implicit user feedback as one of the ranking
parameters” (Br. 16—17). We do not agree.
We agree with and adopt the Examiner’s findings as our own.
Particularly, we agree with the Examiner that Kim discloses multiple
functions and models “used [by the ranking engine] for scoring and ranking”
search results, thereby teaching two different scoring functions as required
by claim 1 (Ans. 3 (citing Kim col. 6,11. 5—19 and 11. 27—65); see also Kim
col. 14,11. 29-37 (discussing two models with “[a]ny difference between
these two models . . . likely to be due only to presentation bias”)). Further,
we agree with the Examiner that Kim’s scoring functions are specified by
features/parameters weighted with adjustable weights, and therefore “Kim
teaches ... a first set of weighting factors (weighted features prior to
adjustments [of a scoring function]) and a second set of weighting[sic]
4
Appeal 2017-003575
Application 12/694,191
factors (adjusted weighted features)” for a second scoring function that
reduces display bias in search results (Ans. 2—3 (citing Kim col. 6,11. 27—65;
col. 8,11. 65—67; col. 20,11. 1—17; Abstract)). Appellants’ arguments have
not addressed these specific findings by the Examiner in a Reply Brief.
We also are not persuaded by Appellants’ argument that the
combination of Kim and Watson does not teach the claimed evaluating the
performance of a new scoring function relative to a production scoring
function (Br. 16—17). Rather, we agree with the Examiner that “Watson
creates a new scoring function (. . . optimizer applies adjustments), and
evaluates the performance of the new scoring function (. . . adjustments are
tested)” (Ans. 3 (citing Watson col. 5,11. 19-42; col. 6,11. 25—60; Figs. 4A,
4B, and 5B)).
As to Appellants’ argument that combining Kim’s “teaching of using
‘implicit user feedback’ with Watson would, at best, result in a ranking
system that uses implicit user feedback as one of the ranking parameters”
(Br. 17), we note the broadly recited limitations in claim 1 do not exclude
Appellants’ proffered scenario. Particularly, claim 1 does not preclude a
new scoring function specified by (1) the production scoring function’s
parameters and weighting factors, together with (2) a new ranking parameter
(e.g., Kim’s implicit user feedback) weighted by its attendant (new)
weighting factor.
Thus, for the above reasons, we sustain the Examiner’s rejection of
independent claim 1, and independent claims 8 and 15 argued for
substantially the same reasons (Br. 17). We also sustain the Examiner’s
rejection of dependent claims 5, 6, 12, and 13, argued for their dependency
on claims 1 and 8 (Br. 23).
5
Appeal 2017-003575
Application 12/694,191
Claims 2 and 9
The Examiner finds Dumon’s “shifting of weights [in a listing quality
score] which takes into account mean values” teaches the average rank shift
metric recited in claims 2 and 9 because Dumon’s listing quality score ranks
an item listing over time by shifting a weight from a predicted score—based
on similar items’ observed mean price—to an observed score associated
with the item listing (Ans. 4 (citing Dumon H 35, 36, 39)).
Appellants argue Dumon’s listing quality score merely measures “the
likelihood that an item listing, if presented in a search results page, will
result in a transaction being concluded,” and does not teach the claimed
“average rank shift metric determined by comparing” listing slot identifiers
from multiple scoring functions “for each item listing that has resulted in a
transaction being concluded” (Br. 19 (citing Dumon H 35, 39)). Appellants
also argue Dumon does not teach the claimed “one or more weighting
factors . . . selected to maximize an averase rank shift metric for the
plurality of search result sets” (Br. 19).
We agree with Appellants that Dumon’s weight shifting in the listing
quality score does not teach an average rank shift metric “determined by
comparing the listing slot identifiers]” derived by two scoring functions, as
required by claims 2 and 9. Rather, Dumon’s weight shifting is determined
by “historical data [that] becomes available to assess the actual performance
of the item listing” (see Dumon 139; Br. 19). The cited portions of Dumon
also fail to teach selecting “one or more weighting factors ... to maximize
[the] average rank shift metric for the plurality of search result sets,” as
recited in claims 2 and 9 (Br. 19). We, therefore, do not sustain the
6
Appeal 2017-003575
Application 12/694,191
Examiner’s rejection of claims 2 and 9, as the Examiner has not identified
sufficient evidence to support this rejection.
Claims 3, 4, 10, and 11
The Examiner finds Dumon’s intermingling of item listings assigned
to different groups to “ensure that the search results page includes item
listings from groups in quantities established by some predefined ratio,”
teaches search result sets having a predetermined number of item listings
assigned to a particular category of item listings, as recited in claims 3 and
10 (Ans. 5 (citing Dumon || 40, 69)). The Examiner finds Dumon’s “item
listing [that] has been presented in a search results page a predetermined
number of times” also discloses this limitation (Final Act. 14 (citing Dumon
1 69)).
Appellants argue Dumon’s grouping and intermingling item listings
are not equivalent to selecting search result sets having a predetermined
number of item listings, as claimed (Br. 21). We agree with Appellants.
Dumon’s predefined ratio is not a predetermined number of item listings, as
recited in claims 3 and 10 (see Dumon 140; Br. 21). Additionally, Dumon’s
predetermined number of times an item listing has appeared in search results
is not a predetermined number of item listings of a search result set, as
claimed (see Dumon | 69; Br. 21).
We, therefore, do not sustain the Examiner’s rejection of claims 3 and
10, as the Examiner has not identified sufficient evidence to support this
rejection. Because we reverse the Examiner’s rejection of claims 3 and 10,
we also reverse the rejection of claims 4 and 11, dependent therefrom.
7
Appeal 2017-003575
Application 12/694,191
Claims 7 and 14
Appellants contend Watson does not teach a new scoring function
specified with parameters that differ from the production scoring function, as
recited in claims 7 and 14 (Br. 23). Particularly, Appellants argue the
claimed “‘parameter’ is defined as ‘a variable that must be given a specific
value during the execution of a program or of a procedure within a
program’,’’ and Watson’s limiting an application to consideration of less
than all sources of performance data “does not infer that [such] parameters
differ'” (Br. 23 (citing Watson col. 12,11. 6—32) (emphases added)). We do
not agree. Rather, we agree with the Examiner that Watson’s “various
sources of collected relevance performance data” are commensurate with
the description of “parameters” in Appellants’ Specification (Ans. 5 (citing
Watson col. 10,1. 45—col. 11,1. 14) (emphasis added); see Spec. 125; Fig.
2).1 In contrast, Appellants’ argument that the term “‘parameter’ is defined
as ‘a variable that must be given a specific value during the execution of a
program or of a procedure within a program’” (Br. 23) is not commensurate
with the scope of the claim or with the “parameters” described in
Appellants’ Specification.
Additionally, we agree with the Examiner that Watson’s search result
relevance optimizer “limited in application to consideration of less than all
1 Appellants’ Specification describes an “example format for a scoring
function is a formula or equation having any number of weighted
parameters'1'’ including “a weighted relevance score or parameter 46,” “a
weighted listing quality score or parameter 48” including “demand metrics,”
“item attributes,” and “similar items,” and “a weighted business rules score
or parameter 50'” including “promotions” and “demotions” data (Spec. 125
(emphases added); Fig. 2 (capitalization altered)).
8
Appeal 2017-003575
Application 12/694,191
sources of performance data” teaches the claimed new scoring function
“specified with parameters that differ from the production scoring function”
(see Watson col. 12,11. 25—28; Ans. 6). Particularly, Watson’s optimized
function has differing parameters from a previous scoring function because
the optimized function’s parameters include “less than all [previously used]
sources of performance data” (Ans. 6 (citing Watson col. 10,1. 45—col. 11,1.
14; col. 12,11. 6-32)).
Accordingly, we sustain the Examiner’s rejection of claims 7 and 14,
as Appellants’ arguments have not persuaded us of error in the Examiner’s
rejection.
DECISION
The Examiner’s decision rejecting claims 1, 5—8, and 12—15 is
affirmed.
The Examiner’s decision rejecting claims 2-4 and 9-11 is reversed.
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
9