Ex Parte Shu

Patent Trial and Appeal Board

Mar 17, 2017

12396933 (P.T.A.B. Mar. 17, 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.
12/396,933 03/03/2009 Chang-Qing Shu 0110243 7214
45191 7590
HERBERT L. AT .TEN
AT .TEN, DYER, DOPPELT, MILBRATH & GILCHRIST, P.A.
255 SOUTH ORANGE AVENUE, SUITE 1401
P. O. BOX 3791
ORLANDO, EL 32802-3791
EXAMINER
MOBIN, HASANUL
ART UNIT PAPER NUMBER
2168
NOTIFICATION DATE DELIVERY MODE
03/21/2017 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):
Skemraj @ addmg.com
jlong@addmg.com
nmacdonald @ addmg.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte CHANG-QING SHU1
Appeal 2017-000626
Application 12/396,933
Technology Center 2100
Before HUNG H. BUI, JON M. JURGOVAN, and
DAVID J. CUTITTAII, Administrative Patent Judges.
CUTITTA, Administrative Patent Judge.
DECISION ON APPEAL
This is an appeal under 35 U.S.C. § 134(a) from the Examiner’s
decision rejecting claims 1—5, 8—17, 19, 22, 25, and 26, all pending claims of
the application.2 We have jurisdiction under 35 U.S.C. § 6(b).
We affirm.
1 According to Appellant, the real party in interest is Adacel, Inc. Appeal
Br. 2.
2 Claims 6, 7, 18, 20, 21, 23, and 24 are cancelled. Final Act. 2.
Appeal 2017-000626
Application 12/396,933
STATEMENT OF THE CASE
According to Appellant, the application relates to tuning a user-
dependent language model for a speech recognition engine by reviewing
data files viewed or drafted by a user to determine the user’s preferred
vocabulary. Spec. 4, 15, 16.3 Claims 1, 17, and 22 are independent.
Claim 1 is representative and is reproduced below with disputed limitations
in italics:
1. A method of making a user dependent language model for
a speech recognition engine, the user dependent language model
being dependent on a particular user, the method comprising:
reviewing a plurality of data files to determine whether the
data files include text viewed by the particular user,
extracting the user-vie wed texts from the data files;
associating weighting factors with the extracted texts',
generating a sorted text element list based on the extracted
texts and the weighting factors; and
compiling the user dependent language model based on
the sorted text element list;
wherein associating weighting factors with the extracted
texts includes weighting user-generated extracted texts higher
than other extracted texts.
Appeal Br. 23 (Claims App’x).
3 Throughout this Decision, we refer to the following documents: (1)
Appellant’s Specification filed March 3, 2009 (Spec.); (2) the Final Office
Action (Final Act.) mailed June 4, 2015; (3) the Appeal Brief (Appeal Br.)
filed January 5, 2016; and (4) the Examiner’s Answer (Ans.) mailed July 15,
2016.
2
Appeal 2017-000626
Application 12/396,933
REFERENCES
The prior art relied upon by the Examiner in rejecting the claims on
appeal includes:
Smith US 6,308,151 B1
Nguyen et al. (“Nguyen”) US 2003/0050778 A1
Diao et al. (“Diao”) US 8,023,974 B1
Cheng et al. (“Cheng”) WO 8,495,144 B1
REJECTIONS
Claims 1, 8—11, and 16 stand rejected under 35 U.S.C. § 103(a) as
unpatentable over the combination of Smith and Nguyen. Final Act. 3—9.
Claims 2, 3, 4, 5, 17, and 19 stand rejected under 35 U.S.C. § 103(a)
as unpatentable over the combination of Smith, Nguyen, and Cheng. Final
Act. 9-16.
Claims 12, 13, 14, and 15 stand rejected under 35 U.S.C. § 103(a) as
unpatentable over the combination of Smith, Nguyen, and Diao. Final
Act. 16-19.
Claims 22, 25, and 26 stand rejected under 35 U.S.C. § 103(a) as
unpatentable over the combination of Smith, Cheng, Nguyen, and Diao.
Final Act. 19—24.
We review the appealed rejections for error based upon the issues
identified by Appellant, and in light of the arguments and evidence produced
thereon. See 37 C.F.R. § 41.37(c)(l)(iv) (2014).
Oct. 23, 2001
March 13, 2003
Sept. 20, 2011
July 23, 2013
3
Appeal 2017-000626
Application 12/396,933
ISSUES
1. Does the Examiner err in finding the combination of Smith and
Nguyen teaches or suggests “reviewing a plurality of data files to determine
whether the data files include text viewed by the particular user,” as recited
in claim 1 ?
2. Does the Examiner err in finding the combination of Smith and
Nguyen teaches or suggests “associating weighting factors with the
extracted texts . . . wherein associating weighting factors with the extracted
texts includes weighting user-generated extracted texts higher than other
extracted texts,” as recited in claim 1 ?
DISCUSSION
We disagree with Appellant’s contentions, and we adopt as our own
(1) the Examiner’s findings and reasoning set forth in the Office Action
from which this appeal is taken (Final Act. 4—6) and (2) the Examiner’s
reasoning set forth in the Examiner’s Answer (Ans. 3—6). We highlight the
following points for emphasis.
Issue 1
Appellant argues the Examiner errs in rejecting claim 1 because Smith
neither teaches nor suggests reviewing a plurality of data files
to determine whether the data files include text viewed by the
particular user. Instead, Smith teaches a system and method for
the specific speech recognition application of ‘dictating a body
of text in response to an available body of text. ’
Appeal Br. 12 (citing Smith, Abstract).
4
Appeal 2017-000626
Application 12/396,933
We are not persuaded by Appellant’s argument. Smith discloses a
speech recognition system that “uses the content of [a] received E-mail
message to update the language model [] for the user’s dictation session
responding to the received E-mail message” to improve the recognition
accuracy of the user’s dictated E-mail response. Smith, col. 4,11. 12—17. As
such, we agree with the Examiner’s finding that “the particular user’s
interaction with the e-mail is the determination that the user has seen or
reviewed the e-mail message.” Ans. 4. That is, in dictating a response to a
particular email, we find that a user must determine whether the email
includes text viewed by the user, and if not, view the text in order to dictate
a response. See Smith, col. 4,11. 12—17.
Appellant further argues “the nature of the application-specific nature
of Smith renders such a determination unnecessary (i.e., if every text to be
incorporated would likely have been user-viewed, why is a preliminary
determination of that fact necessary or helpful?).” Appeal Br. 13.
In indicating every email would likely have been user-viewed,
Appellant acknowledges that there is a possibility an email would not have
been user-viewed. Appeal Br. 13. Thus, consistent with Appellant’s
acknowledgment, when a user is uncertain whether he or she has previously
viewed an email, we find the user will determine whether the email includes
text previously viewed by the user. Moreover, even if a user had viewed a
particular email in the past, the user might still wish to review the email to
prepare an appropriate response to the email based on further review.
Appellant next argues
claim 1 requires that a plurality of data files be reviewed to
determine whether they include user-viewed text, and extracted
texts from all such data files are used to make the language
5
Appeal 2017-000626
Application 12/396,933
model. The response dictation system and method of Smith
does not inherently require text to be extracted from more than
one data file (i.e., the original email, chat room or news group
conversation, or other body of text).
Appeal Br. 14.
We find this argument unpersuasive because we agree with the
Examiner’s finding that Smith does not limit a user to viewing a single
email. Ans. 4 (citing Smith col. 6,11. 45—55). For example, Smith discusses
a user “dictating a body of text in response to another body of text already
available to the user,” which we agree teaches or suggests reviewing a
plurality of emails. Smith, col. 6,11. 48—50. As another example, Smith
teaches “reviewing a plurality of data files,” as claimed, by suggesting a user
may review two emails in succession. See Smith Fig. 2, 102.
Accordingly, we agree with the Examiner’s finding that Smith teaches
or suggests “reviewing a plurality of data files to determine whether the data
files include text viewed by the particular user,” as recited in claim 1.
Issue 2
The Examiner relies on Nguyen to teach or suggest “associating
weighting factors with the extracted texts,” as recited in claim 1. Final
Act. 5—6; Ans. 3^4 (citing Nguyen || 20—22). Nguyen describes
automatically determining a topic from a body of a received email message
by weighting the words in the body of the email. Nguyen 121.
Appellant argues
weighting words within a text based on parts of speech has
absolutely nothing to do with assigning weighting factors based
on whether extracted texts are user-generated or not. Use of a
6
Appeal 2017-000626
Application 12/396,933
given part of speech - a la Nguyen - in no way distinguishes a
user generated text from one that is not user generated.
Appeal Br. 15 (emphasis added).
We find this argument unpersuasive because claim 1 does not limit
the weighting to whether the extracted text is user-generated or not user­
generated. Rather, the claim recites “weighting user-generated extracted
texts higher than other extracted texts.” Hence, Appellant argues for
patentability on the basis of limitations that are not recited in the claim. See
In re Self, 671 F.2d 1344, 1348 (CCPA 1982).
The Examiner finds, and we agree, that Nguyen’s received message is
user-generated text. Ans. 5. We also agree with the Examiner’s finding that
Nguyen assigns a low weight to certain words of the received message such
as articles (“‘the.’ ‘an’” etc.) and assigns a higher weight to other words such
as nouns. Ans. 5; see also Nguyen 122.
Thus, we find that Nguyen teaches weighting user-generated extracted
texts such as nouns higher than other extracted texts such as articles (which,
commensurate with the scope of claim 1, may also be user generated).
Appellant argues “Nguyen’s teachings regarding weighting of words
within a text is not even performed in connection with the inclusion of such
words into a language model.” Appeal Br. 16.
The Examiner, however, relies on Smith, not Nguyen, to teach a user
dependent language model. See Final Act. 4. We conclude, therefore, that
Appellant’s argument does not address the actual reasoning of the
Examiner’s rejections. Instead, Appellant attacks the references singly for
lacking teachings that the Examiner relies on a combination of references to
show. It is well established that one cannot show nonobviousness by
7
Appeal 2017-000626
Application 12/396,933
attacking references individually where the rejections are based on
combinations of references. See In re Merck & Co., 800 F.2d 1091 (Fed.
Cir. 1986). This form of argument is inherently unpersuasive to show
Examiner error. Our reviewing court requires that references must be read,
not in isolation, but for what they fairly teach in combination with the prior
art as a whole. Merck, 800 F.2d at 1097.
Appellant argues “even if paragraph [0022] suggested weighting user­
generated texts higher than other texts, there would still be no clear reason
for incorporating such a teaching into a method of making a language
model.” Appeal Br. 17.
Appellant’s argument that the combination lacks the required
motivation is not persuasive of error because Appellant does not address the
motivation identified by the Examiner. See Final Act. 6. That is, the
Examiner has found actual teachings in the prior art and has additionally
provided a rationale for the combination:
it would have been obvious to one ordinary skill in the art at the
time of invention was made having the teachings of Smith and
Nguyen before him/her, to modify Smith with the teaching of
Nguyen’s focused language models for improved speech input
of structured documents. One would have been motivated to do
so for the benefit of providing Smith with a focused language
model for generating e-mail and text message (i.e., user
dependent language model) to improve speech recognition.
Final Act. 6.
Further, we find the teachings of Smith and Nguyen suggest that the
combination involves the predictable use of prior art elements according to
their established functions. “The combination of familiar elements
according to known methods is likely to be obvious when it does no more
8
Appeal 2017-000626
Application 12/396,933
than yield predictable results,” KSR Int’l Co. v. Teleflex, Inc., 550 U.S. 398,
416 (2007), especially if the combination would not be “uniquely
challenging or difficult for one of ordinary skill in the art,” Leapfrog Enters.,
Inc. v. Fisher-Price, Inc., 485 F.3d 1157, 1162 (Fed. Cir. 2007) (citing KSR,
550 U.S. at 420). We consequently find the Examiner has provided
sufficient motivation for combining Smith and Nguyen.
Accordingly, we sustain the Examiner’s obviousness rejection of
claim 1. We also sustain the Examiner’s obviousness rejection of
independent claims 17 and 22, which are nominally argued separately, and
are consequently rejected with independent claim 1 for similar reasons.
Appeal Br. 19-22. Dependent claims 2—5, 8—16, 19, 25, and 26, are either
nominally argued separately, or are not argued separately, and thus are
rejected with their respective independent claims. Appeal Br. 19—22.
DECISION
We affirm the Examiner’s decision rejecting claims 1—5, 8—17, 19, 22,
25, and 26 under 35 U.S.C. § 103(a).4
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
4 In the event of further prosecution of this application, the Examiner should
review and consider rejecting claim 1 under 35 U.S.C. § 101 in light of the
Supreme Court Decision in Alice Corporation Pty. Ltd. v. CLS Bank Int’l,
134 S. Ct. 2347 (2014) and subsequent agency guidance.
9