FAIANS, SHAHAR et al.

Patent Trials and Appeals Board

Feb 3, 2020

15436020 - (D) (P.T.A.B. Feb. 3, 2020)

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.
15/436,020 02/17/2017 SHAHAR FAIANS P-76715-US1 7810
49443 7590 02/03/2020
Pearl Cohen Zedek Latzer Baratz LLP
1500 Broadway
12th Floor
New York, NY 10036
EXAMINER
COLUCCI, MICHAEL C
ART UNIT PAPER NUMBER
2658
NOTIFICATION DATE DELIVERY MODE
02/03/2020 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):
Arch-USPTO@PearlCohen.com
USPTO@PearlCohen.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte SHAHAR FAIANS, AVRAHAM LOUSKY, ELAD HOFFMAN,
ALON MOSHE SABBAN, JADE TARNI KAHN, and ROIE MANDLER
Appeal 2018-003595
Application 15/436,020
Technology Center 2600

Before ST. JOHN COURTENAY III, JENNIFER S. BISK, and
IRVIN E. BRANCH, Administrative Patent Judges.

BRANCH, Administrative Patent Judge.

Appeal 2018-003595
Application 15/436,020
2
DECISION ON APPEAL
STATEMENT OF THE CASE1
Pursuant to 35 U.S.C. § 134(a), Appellant2 appeals from the
Examiner’s decision to reject claims 1, 3–17, and 19–23.3 See Final Act. 1.
We have jurisdiction under 35 U.S.C. § 6(b). An oral hearing was held on
October 24, 2019.
We AFFIRM.
CLAIMED SUBJECT MATTER
The claims are directed to a computer-implemented method for
enrolling voice prints generated from audio streams. Spec. Abstract. Claim
1, reproduced below with disputed limitations emphasized in italics, is
representative of the claimed subject matter:
1. A computer-implemented method comprising:
receiving audio streams from an audio server that archives
audio data of historical communication sessions between
customers and representatives;
for each of at least some of the audio streams generating
and enrolling a respective enrolled voice print by:
a management server using a processor to create a
preliminary association between an audio stream related
to a particular one of the historical communication
sessions and an identity of a customer that has engaged in

1 We refer to the Specification, filed March 29, 2017 (“Spec.”); Final Office
Action, mailed August 29, 2017 (“Final Act.”); Appeal Brief, filed
November 5, 2017 (“Appeal Br.”); Examiner’s Answer, mailed January 19,
2018 (“Ans.”); and Reply Brief, filed February 15, 2018 (“Reply Br.”).
2 We use the word Appellant to refer to “applicant” as defined in 37 C.F.R.
§ 1.42. Appellant identifies the real party in interest as NICE Ltd. Appeal
Br. 1.
3 Claims 2 and 18 have been indicated to have allowable subject matter.
Final Act. 22.
Appeal 2018-003595
Application 15/436,020
3
the particular historical communication session based on
stored identification information;
determining using the processor a confidence level
of the preliminary association based on stored
authentication information related to the customer;
if the confidence level is higher than a threshold,
sending a request to generate from the audio stream a
current voice print associated with the customer;
prior to enrolling the current voice print, comparing
using the processor the current voice print associated with
the customer to a voice print previously enrolled for the
customer and comparing using the processor the current
voice print associated with the customer to voice prints
associated with other customers; and
enrolling using the processor the current voice print,
if the current voice print does not match any other voice
prints associated with any of the other customers and if the
current voice print is consistent with the voice print
previously enrolled for the customer, in a customer voice
print database as an enrolled voice print; and
automatically authenticating the customer during a
succeeding communication session based on the enrolled voice
print.
Appeal 2018-003595
Application 15/436,020
4
REFERENCES
The prior art relied upon by the Examiner is:
Name Reference Date
Wasserblat
Davis
US 8,145,562 B2
US 9,197,736 B2
Mar. 27, 2012
Nov. 24, 2015
GROSS US 2009/0319270 A1 Dec. 24, 2009
Schmidt US 2011/0010543 A1 Jan. 13, 2011
REJECTIONS4
Claims 1, 4–8, 10–12, 14, 17, and 20–23 stand rejected under
35 U.S.C. § 103(a) as being unpatentable over Wasserblat and GROSS.
Final Act. 11–19.
Claims 3, 13, and 19 stand rejected under 35 U.S.C. § 103(a) as being
unpatentable over Wasserblat, GROSS, and Schmidt. Final Act. 19–20.
Claims 9, 15, and 16 stand rejected under 35 U.S.C. § 103(a) as being
unpatentable over Wasserblat and Davis. Final Act. 21–22.5

4 Because Appellant does not Appeal the nonstatutory obviousness-type
double patenting rejection of all pending claims, we summarily affirm the
rejection. Final Act. 6–7; Appeal Br. 5.
5 To the extent the Examiner's oversight resulted in omitting an additional
citation to the GROSS reference to properly support the rejection of claims
9, 15, and 16 (which each depend directly from claim 1), we find the Office
Action nevertheless provides Appellant with sufficient notice of the basis for
the rejection, and therefore complies with the minimum notice requirements
of 35 U.S.C. § 132(a). Under § 132(a), an Office action must at least contain
“such information and references as may be useful in judging of the
propriety of continuing the prosecution of his application.” “Section 132 is
violated when a rejection is so uninformative that it prevents the applicant
from recognizing and seeking to counter the grounds for rejection.” Chester
v. Miller, 906 F.2d 1574, 1578 (Fed. Cir. 1990).
Appeal 2018-003595
Application 15/436,020
5
ANALYSIS
We have reviewed the Examiner’s rejections in light of Appellant’s
arguments. We have considered in this Decision only those arguments
Appellant actually raised in the Briefs. Any other arguments Appellant
could have made but chose not to make in the Briefs are deemed to be
waived. See 37 C.F.R. § 41.37(c)(1)(iv) (2017).
To the extent consistent with our analysis herein, we adopt as our own
the findings and reasons set forth by the Examiner in (1) the action from
which this appeal is taken (Final Act. 11–22) and (2) the Examiner’s Answer
in response to Appellant’s Appeal Brief (Ans. 3–12) and concur with the
conclusions reached by the Examiner. We highlight the following for
emphasis.
35 U.S.C. § 1036
The Examiner finds most elements of claim 1 are taught by
Wasserblat. Final Act. 11–13. The Examiner reasons that one of ordinary
skill in the art would have found it obvious at the time of the invention “to
modify the system of Wasserblat to incorporate fraud detection as taught by
embodiments of Wasserblat to allow for both generation in addition to
retrieval of profiles for combined scoring and profiling component.” Id. at
13. The Examiner further finds that
[w]hile Wasserblat teaches comparison with known fraudster
lists prior to adding a voice print to a user profile, it fails to teach
prior to enrolling the current voice print, comparing using the
processor the current voice print associated with the customer to
a voice print previously enrolled for the customer and comparing

6 Because Appellant argues all pending claims based on claim 1 (Appeal Br.
20) we base our analysis on claim 1 and, except for our ultimate conclusion,
do not mention the remaining claims further.
Appeal 2018-003595
Application 15/436,020
6
using the processor the current voice print associated with the
customer to voice prints associated with other customers.
Id. The Examiner finds that GROSS teaches what is missing from
Wasserblat and that claim 1 would have been obvious based on the
combined teachings and suggestions of the references. Id. at 13–14.
Specifically, the Examiner finds, as follows:
GROSS expressly teaches a machine voice print database
in addition to a database of known fraudster information, as well
as a human speaker database separate from machine and known
fraudster, wherein a when trying to create a new voice print the
enrolling customer if determined not to be fraudulent is either
permitted to create a new voice print or denied due to the same
voice print already existing with another account, such a second
verification step would prevent a non-fraudster nonmachine user
from creating a new account or print with an existing voice print
with another user and thereby safely allowing a new user to
create a voice print, thereby reducing duplicate accounts for
spam purposes (GROSS 0136, 0127 with fig. 2), wherein this
avoids the creation of a fraudulent profile that would otherwise
go undetected.
Therefore, it would have been obvious to one of ordinary
skill in the art at the time of the invention to modify the system
of Wasserblat to incorporate the above limitation as taught by
GROSS to allow for 1) preventing a non-fraudster non-machine
user (recorded synthetic input) from creating a new account or
print with an existing voice print with another user and 2) thereby
safely allowing a new user to create a voice print, thereby
reducing duplicate accounts for spam purposes (GROSS 0136,
0127 with fig. 2), such that the new user will not be a
recorded/synthetic voice or known-fraudster i.e. imposter, the
system of Wasserblat is thereby enhanced to provide another
layer of security for duplicate users as expressly shown in
GROSS even if the first fraud test is passed.
Id.
Appeal 2018-003595
Application 15/436,020
7
Appellant contends that Wasserblat does not teach or suggest “a
preliminary association between an audio stream related to a particular one
of the historical communication sessions and an identity of a customer that
has engaged in the particular historical communication session based on
stored identification information” and “determining using the processor a
confidence level of the preliminary association based on stored
authentication information related to the customer.” Appeal Br. 6–11.
Appellant contends “Wasserblat teaches two risk scores: a) a risk score as to
how [likely] an interaction is fraudulent . . . ; and b) a risk score for the
transaction.” Id. at 7 (citing Wasserblat ¶¶ 32–34).
The Examiner finds the claimed “preliminary association” means “the
comparison of features of the input user to existing features of previous
users.” Ans. 4. The Examiner further finds that Wasserblat discloses:
extracting a voice print or a voice sample from one or more of
the interactions; comparing the voice print or voice sample to one
or more speaker models to obtain a similarity score; determining
a fraud risk score from the similarity score and from one or more
characteristics associated with a transaction related to the
interaction; and taking a safety measure when said fraud risk
score exceeds a predetermined threshold.
Id. at 4 (quoting Wasserblat, ¶ 14). The Examiner also cites Wasserblat’s
Abstract, which discloses:
one or more interactions is correlated with one or more
transactions, the interactions . . . and transactions features are
combined, and features are extracted from the combined
structure. The features are compared against one or more
profiles, and a combined risk score is determined for the
interactions or transactions. If the risk score exceeds a
predetermined threshold, a preventive/corrective action can be
taken.
Appeal 2018-003595
Application 15/436,020
8
Id.
We understand the Examiner to map the claimed “preliminary
association” to Wasserblat’s “similarity score,” which is a comparison of a
voice print from a customer interaction to a speaker model. We understand
the Examiner to map the claimed confidence level to Wasserblat’s fraud risk
score, which is based on the similarity score. We note that, consistent with
the Examiner’s findings, Wasserblat discloses numerous comparisons of
interactions, including voice samples and transactions therefrom, with
historical interactions, including stored voice prints. See Wasserblat
Abstract, ¶ 14. We note that, while Wasserblat discloses these comparisons
for the purpose of fraud detection, consistent with the Examiner’s findings,
Wasserblat discloses taking some action based on a comparison exceeding a
threshold. Id. See Final Act. 11–14; Ans. 4–11. We agree with the
Examiner’s conclusion as follows:
it would have been obvious to one of ordinary skill in the art at
the time of the invention to modify the system of Wasserblat to
incorporate the above limitation as taught by GROSS to allow
for 1) preventing a non-fraudster non-machine user (recorded
synthetic input) from creating a new account or print with an
existing voice print with another user and 2) thereby safely
allowing a new user to create a voice print, thereby reducing
duplicate accounts for spam purposes (GROSS 0136, 0127 with
fig. 2), such that the new user will not be a recorded/synthetic
voice or known-fraudster i.e. imposter, the system of Wasserblat
is thereby enhanced to provide another layer of security for
duplicate users as expressly shown in GROSS even if the first
fraud test is passed.
Final Act. 14.
Appellant does not persuasively rebut the Examiner’s construction of
“preliminary association.” See generally Appeal Br.; Reply Br. Nor does
Appeal 2018-003595
Application 15/436,020
9
Appellant address the Examiner’s mapping. Id. Specifically, Appellant
does not rebut the Examiner’s mapping of the claimed preliminary
association to Wasserblat’s “similarity score,” upon which the fraud risk
score is based. Reply Br. 1–2. See Wasserblat, ¶ 14. Appellant attempts to
traverse the Examiner’s findings by focusing on Wasserblat’s “risk scores”
(see, e.g., Appeal Br. 7), but nowhere considers the possibility of
Wasserblat’s “risk score” being the claimed “preliminary association.”
Accordingly, we are unpersuaded of error in the Examiner’s finding
that Wasserblat teaches “a preliminary association between an audio stream
related to a particular one of the historical communication sessions and an
identity of a customer that has engaged in the particular historical
communication session based on stored identification information” and
“determining using the processor a confidence level of the preliminary
association based on stored authentication information related to the
customer.”
Appellant makes several additional arguments, which do not persuade
us of error for the reasons which follow. Specifically, Appellant argues
“Wasserblat’s risk scores are never applied in Wasserblat as the confidence
level is in the current claims, and to modify Wasserblat to do so would
impermissibly change the principle of operation of Wasserblat.” Appeal Br.
11. This argument does not persuade us that merely changing from fraud
detection based on compared voice samples to gating a voice sample update
amounts to an impermissible change in Wasserman’s principle of operation.
Appellant also argues “Wasserblat does not teach or suggest actually
enrolling voice prints.” Appeal Br. 14; see id. at 12–15. This line of
argument is unpersuasive because it is based on the teachings of Wasserblat
alone and does not consider the teachings of GROSS or, more importantly,
Appeal 2018-003595
Application 15/436,020
10
what the combined teachings of the references would have taught or
suggested to one of ordinary skill in the art. “Non-obviousness cannot be
established by attacking references individually where the rejection is based
upon the teachings of a combination of references.” In re Merck & Co., 800
F.2d 1091, 1097 (Fed. Cir. 1986).
Appellant also argues as follows:
to modify Wasserblat so that its combined fraud probability was
only computed if a risk score passed a threshold, or so that its
comparison to its black list of fraudsters was only done if a score
was a certain amount, would prevent Wasserblat from working
as intended, which is to produce a combined fraud probability
based on the factors presented in para. [0068] regardless of the
individual factors.
Appeal Br. 18.
Appellant also argues “Gross only teaches using voice prints to
authenticate a person, not to decide whether or not to enroll a voice print.
Neither Gross nor Wasserblat teaches enrolling voice prints” and “if Gross
were combined with a system such as the Office Action’s interpretation of
Wasserblat, the resulting system would not work.” Appeal Br. 19–20. This
argument is unpersuasive because it argues against a bodily incorporation of
the teachings of Wasserblat and Gross. “The test for obviousness is not
whether the features of a secondary reference may be bodily incorporated
into the structure of the primary reference. . . . Rather, the test is what the
combined teachings of [those] references would have suggested to those of
ordinary skill in the art.” In re Keller, 642 F.2d 413, 425 (CCPA 1981). See
also In re Sneed, 710 F.2d 1544, 1550 (Fed. Cir. 1983) (“[I]t is not necessary
that the inventions of the references be physically combinable to render
obvious the invention under review”); and In re Nievelt, 482 F.2d 965, 968
Appeal 2018-003595
Application 15/436,020
11
(CCPA 1973) (“Combining the teachings of references does not involve an
ability to combine their specific structures”). Rather, “if a technique has
been used to improve one device, and a person of ordinary skill in the art
would recognize that it would improve similar devices in the same way,
using the technique is obvious unless its actual application is beyond his or
her skill.” KSR, 550 U.S. at 417.
We further note—without relying on—the fact that claim 1 recites
numerous conditional limitations (e.g., “if the confidence level is higher than
a threshold . . . ,” and “if the current voice print does not match any other
voice prints associated with any of the other customers and if the current
voice print is consistent with the voice print previously enrolled for the
customer . . .”). Should further prosecution ensue based on the presently-
pending claims, the Examiner might consider whether the conditional
limitations of claim 1 should be given patentable weight. See Ex parte
Schulhauser, No. 2013-007847, 2016 WL 6277792, at *3–6 (PTAB Apr. 28,
2016) (concluding the broadest reasonable interpretation of a claim
encompassed situations in which conditional method steps “need not be
reached”) (precedential).
CONCLUSION
We affirm the Examiner’s decision to reject claims 1, 3–17, and 19–
23 as obvious.
Appeal 2018-003595
Application 15/436,020
12
DECISION SUMMARY
Claims
Rejected
35 U.S.C. § Reference(s)/Basis Affirmed Reversed
1–23 Nonstatutory Double
Patenting
1–23
1, 4–8, 10–
12, 14, 17,
20–23
103(a) Wasserblat, GROSS 1, 4–8, 10–
12, 14, 17,
20–23

3, 13, 19 103(a) Wasserblat, GROSS,
Schmidt
3, 13, 19
9, 15, 16 103(a) Wasserblat, Davis,
GROSS
9, 15, 16
Overall
Outcome
1–23

TIME PERIOD FOR RESPONSE
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a). See 37 C.F.R.
§ 1.136(a)(1)(iv).
AFFIRMED