Ex Parte Veksler et al

Patent Trial and Appeal Board

Apr 10, 2017

14667134 (P.T.A.B. Apr. 10, 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.
11123.1112US 2739
EXAMINER
ROBERTS, SHAUN A
ART UNIT PAPER NUMBER
2657
MAIL DATE DELIVERY MODE
14/667,134 03/24/2015
52558 7590 04/11/2017
PANASONIC AUTOMOTIVE SYSTEM
COMPANY OF AMERICA
776 HWY 74 SOUTH
c/o Panasonic Legal
PEACHTREE CITY, GA 30269
ILYA VEKSLER
04/11/2017 PAPER
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.
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte ILYA VEKSLER, AMBUJ KUMAR,
and NAVEEN REDDY KORUPOL
Appeal 2016-007936
Application 14/667,1341
Technology Center 2600
Before ROBERT E. NAPPI, SCOTT B. HOWARD, and JOYCE CRAIG,
Administrative Patent Judges.
HOWARD, Administrative Patent Judge.
DECISION ON APPEAL
Appellants appeal under 35 U.S.C. § 134(a) from the Final Rejection
of claims 1—20, which constitute all of the claims pending in this application.
We have jurisdiction under 35 U.S.C. § 6(b).
We reverse.
1 Appellants identify Panasonic Automotive Systems Company of America,
Division of Panasonic Corporation of North America, as the real party in
interest. App. Br. 3.
Appeal 2016-007936
Application 14/667,134
THE INVENTION
The disclosed and claimed invention is directed to the training of a
voice recognition system. Spec. 12.
Claim 1, reproduced below, is illustrative of the claimed subject
matter:
1. A method of operating a speech recognition system,
comprising the steps of:
converting a spoken utterance by a user into an electrical
voice signal by use of a local microphone associated with a local
electronic device;
transmitting the electrical voice signal to a remote voice
recognizer;
using the remote voice recognizer to transcribe the
electrical voice signal and to produce a confidence score, the
confidence score indicating a level of confidence that the
transcription of the electrical voice signal substantially matches
the words of the spoken utterance;
transmitting the transcription of the electrical voice signal
and the confidence score from the remote voice recognizer to the
local electronic device; and
using the electrical voice signal, the transcription of the
electrical voice signal and the confidence score at the local
electronic device to train a local voice recognizer.
REFERENCE
The prior art relied upon by the Examiner as evidence in rejecting the
claims on appeal is:
Talwar US 2012/0245934 A1 Sept. 27,2012
REJECTION
Claims 1—20 stand rejected under 35 U.S.C. § 102(a)(1) as being
anticipated by Talwar. Final Act. 2—\.
2
Appeal 2016-007936
Application 14/667,134
ANALYSIS
We have reviewed the Examiner’s rejection in light of Appellants’
arguments that the Examiner erred. In reaching this decision, we have
considered all evidence presented and all arguments made by Appellants.
We are persuaded by Appellants’ arguments that, on the record before us,
the Examiner erred.
Appellants argue the Examiner erred in finding Talwar discloses
“using the electrical voice signal, the transcription of the electrical voice
signal and the confidence score at the local electronic device to train a local
voice recognizer,” as recited in claim 1. App. Br. 12—13; Reply Br. 5—6.
Specifically, Appellants contend confidence scores are not used at a client
device to train a voice recognition system:
Transcriptions and confidence scores are not used at a client
device in Talwar to train a voice recognizer. Rather, as disclosed
in [0101] of Talwar, models 427, 429 are sent from the server to
the client device to merely update models 407, 409. That is, no
training occurs at the client device, and certainly no training
using transcriptions and confidence values occurs at the client
device.
App. Br. 13; see also Reply Br. 6.
The Examiner finds Talwar discloses “communicating models from
the server to the client device to update the models stored on the client
device - where the reference teaches distributed recognition/transcription,
and updating the local device,” and maps that to the using limitation recited
in claim 1. Final Act. 9 (citing Talwar 1101). The Examiner further finds
“[t]he confidence is incorporated in the training as the thresholds are needed
to obtain the best/most likely results for proper training (where for training, a
3
Appeal 2016-007936
Application 14/667,134
most likely result must first be obtained, which is usually performed using
thresholds or confirmation, where Talwar teaches both).” Id. at 4.
“A claim is anticipated only if each and every element as set forth in
the claim is found, either expressly or inherently described, in a single prior
art reference.” Verdegaal Bros. v. Union Oil Co. of California, 814 F.2d
628, 631 (Fed. Cir. 1987). Moreover, the reference must also “disclose[]
within the four comers of the document not only all of the limitations
claimed but also all of the limitations arranged or combined in the same way
as recited in the claim.” Net Money IN, Inc. v. Verisign, Inc., 545 F.3d 1359,
1371 (Fed. Cir. 2008). However, “the reference need not satisfy an
ipsissimis verbis test.” In re Cleave, 560 F.3d 1331, 1334 (Fed. Cir. 2009)
(internal quotation marks omitted). Moreover, “it is proper to take into
account not only specific teachings of the reference but also the inferences
which one skilled in the art would reasonably be expected to draw
therefrom.” In re Preda, 401 F.2d 825, 826 (CCPA 1968); In re Paulsen, 30
F.3d 1475, 1480 (Fed. Cir. 1994) (holding that prior art references must be
“considered together with the knowledge of one of ordinary skill in the
pertinent art”).
We are persuaded by Appellants’ argument that the Examiner erred in
finding Talwar discloses “using the electrical voice signal, the transcription
of the electrical voice signal and the confidence score at the local electronic
device to train a local voice recognizer.” The Examiner identifies Talwar
Paragraph 101 as disclosing training a voice recognizer. Final Act. 9; Ans.
5. However, as shown in Figure 4, that training (step 430) takes place at the
server and not the local electronic device. Talwar Fig. 4.
4
Appeal 2016-007936
Application 14/667,134
Additionally, the Examiner finds, the confidence value “is sent and
used to train the local models to further improve recognition at the local
device.” Ans. 6; see also Final Act. 4. However, the Examiner does not
provide a citation supporting that finding and a rejection “may not resort to
speculation, unfounded assumptions or hindsight reconstruction to supply
deficiencies in its factual basis.” In re Warner, 379 F.2d 1011, 1017 (CCPA
1967).
Because we agree with at least one of the dispositive arguments
advanced by Appellants, we need not reach the merits of Appellants’ other
arguments.
Accordingly, we are constrained on this record to reverse the
Examiner’s rejection of claim 1, along with the rejections of claims 8 and
15, which recite limitations commensurate in scope to the disputed
limitations discussed above, and dependent claims 2—7, 9-14, and 16—20.
DECISION
For the above reasons, we reverse the Examiner’s decisions rejecting
claims 1—20.
REVERSED
5