Lenovo (Singapore) Pte. Ltd.

Patent Trials and Appeals Board

Jun 5, 2020

2019006652 (P.T.A.B. Jun. 5, 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.
14/097,954 12/05/2013 Peter Hamilton Wetsel RPS920130154USNP(710.289) 4223
58127 7590 06/05/2020
FERENCE & ASSOCIATES LLC
409 BROAD STREET
PITTSBURGH, PA 15143
EXAMINER
COLUCCI, MICHAEL C
ART UNIT PAPER NUMBER
2656
MAIL DATE DELIVERY MODE
06/05/2020 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 PETER HAMILTON WETSEL, JONATHAN GAITHER KNOX,
SUZANNE MARION BEAUMONT, RUSSELL SPEIGHT VANBLON, and
ROD D. WALTERMANN
Appeal 2019-006652
Application 14/097,954
Technology Center 2600

Before JAMESON LEE, SALLY C. MEDLEY, and
JONI Y. CHANG, Administrative Patent Judges.

LEE, Administrative Patent Judge.

DECISION ON APPEAL
STATEMENT OF THE CASE
Pursuant to 35 U.S.C. § 134(a), Appellant1 appeals from the Examiner’s
decision to finally reject claims 1–6, 8–16, and 18–20, all of the claims now
pending in this Application.2 We have jurisdiction under 35 U.S.C. § 6(b).
We affirm.

1 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 Lenovo (Singapore) PTE.
LTD. Appeal Br. 3.
2 Claims 7 and 17 have been cancelled. Appeal Br. 24, 27.
Appeal 2019-006652
Application 14/097,954

2
CLAIMED SUBJECT MATTER
The invention generally relates to a method involving receiving user voice
input, identifying words in the user voice input, determining one of the identified
words as rendering ambiguous a command in the user voice input, accessing
context data, disambiguating the command based on context data, and committing
an action according to the command. Spec. ¶ 2. Claims 1, 11, and 20 are
independent. Each of claims 1, 11, and 20 recites that the command comprises
“one deictic word that renders the command ambiguous.” Further, each claim
requires the disambiguating to be done by “disambiguating the one deictic word.”
Claim 1 is illustrative and reproduced below:
1. A method, comprising:
receiving at an audio receiver of an information handling
device, user voice input;
identifying, using a processor, a command included in
the user voice input, wherein the command comprises at least
one deictic word that renders the command ambiguous;
accessing, using the processor, context data comprising a
list of recent actions on the information handling device;
disambiguating, using the processor, the command based
on the context data, said disambiguating comprising using the
list of recent actions on the information handling device to
disambiguate the at least one deictic word; and
automatically committing, based on the disambiguating,
a predetermined action according to the command.
Appeal Br. 23 (Claims App.).
Claim 11 is similar to claim 1 but is an apparatus claim reciting an
information handling device includinng a processor which performs functions
corresponding to the steps of claim 1. Id. at 25. Claim 11 also recites a memory
device for storing instructions executable by the processor. Id. Claim 20 recites a
Appeal 2019-006652
Application 14/097,954

3
product comprising a storage device having code stored therein, and the code is
operable to perform functions corresponding to the steps recited in claim 1. Id.
at 27–28.
REFERENCES
Phillips US 2011/0060587 A1 Mar. 10, 2011
Bangalore US 8,103,502 B1 Jan. 24, 2012

REJECTION
Claims 1–6, 8–16, and 18–20 were finally rejected under 35 U.S.C. § 103 as
obvious over Phillips and Bangalore. Final Act. 4.
OPINION
A. The Obviousness Rejection of Claims 1–6, 8–16,
and 18–20 over Phillips and Bangalore
1. Claim 1
Key is the limitation that “the command comprises at least one deictic word
that renders the command ambiguous,” and the associated limitation of
disambiguating that command by disambiguating the at least one deictic word.
Appeal Br. 23 (Claims App.). At the outset, we note that what the Examiner
regards as “the command” in Phillips is inconsistently expressed in the Final
Office Action and in the Examiner’s Answer. In the Final Office Action, for
Phillips’s phrase “Send SMS to joe cerra let’s meet at pete’s in harvard square at 7
am,” the Examiner refers to “Send SMS” as the command, and the remainder of
the phrase as context. Final Office Act. 4–5 (emphasis omitted). Similarly, for
Phillips’s phrase “call Joe Cerra,” the Examiner refers only to “call” as the
command. Id. at 5 (emphasis omitted). In the Examiner’s Answer, however, what
is regarded as the command is the entire sentence spoken by the user. Ans. 4, 6.
For instance, the Examiner states “[i]n Phillips a user speaks a full sentence
Appeal 2019-006652
Application 14/097,954

4
command and the system disambiguates words that can refer to multiple items” (id.
at 4) (emphasis added) and “[a] user speaks a full sentence command and the
system disambiguates words that can refer to multiple items” (id. at 6) (emphasis
added). Because the Examiner failed to identify anything in “call” or in “Send
SMS” that is even arguably ambiguous, the reliance on “call” and “Send SMS” as
the command in Phillips is deficient. Hereinafter, we treat and discuss the entire
spoken sentence as the command.
Further, the Final Office Action and the Examiner’s Answer are inconsistent
with regard to whether the command in Phillips is rendered ambiguous by a deictic
word. In the Final Office Action, the Examiner states: “Phillips while expressly
teaching ambiguity, context extraction, and user history for disambiguation, fails to
teach the at least [one] deictic word [in the command].” Final Act. 7. Thus, in the
Final Office Action, the Examiner relies on the teachings of Bangalore, in
combination of Phillips, to account for the requirement of a deictic word in the
command, which renders the command ambiguous. Id. at 7–9. In the Examiner’s
Answer, however, and contrary to the determination in the Final Office Action, the
Examiner finds a deictic word in the command of Phillips and explains as follows:
However, deictic is not only in reference to pronouns. In the context of
Phillips we look to fig. 16 where there are two “Mike Jones” contacts,
using the historical and frequent usage in Phillips the system would be
able to disambiguate which Mike Jones to call e.g. the user calls
MOBILE 123 90% of the time therefore disambiguation would take
place so that Mike Jones in a call example of “Call Mike Jones” would
refer to Mobile thereof.
Ans. 7. Indeed, in the Examiner’s Answer, it is determined that in Phillips,
“disambiguation of a deictic word is inherently present.” Id. at 7–8. Hereinafter,
we will address both positions of the Examiner, treating them as alternative
positions.
Appeal 2019-006652
Application 14/097,954

5
We reject the Examiner’s position regarding “Mike Jones” in the command
“Call Mike Jones” as a deictic word rendering the command ambiguous. Phillips
does not describe that there are two people both named Mike Jones, as the
Examiner suggests. Rather, Figure 16 of Phillips shows a home phone number for
Mike Jones and a mobile number for Mike Jones in the contact list maintained by
the system, as shown below:

Figure 16 illustrates a sample navigating browser screen for a user interface.
Phillips ¶ 42. The command “Call Mike Jones” can correctly be executed by using
either number. Breadth of command does not equal ambiguity. Treating “Mike
Jones” as a deictic word rendering the command “Call Mike Jones” ambiguous is
an unreasonably broad interpretation.
The Examiner determined that “deictic” is not limited to pronouns, e.g.,
“him,” “her,” etc., and that “deictic” words are “words which cannot be understood
without further context analysis.” Ans. 5. Generally, we do not disagree. But the
devil is in the detail and the line has to be drawn reasonably, or else almost every
word can be deemed a deictic word on the basis that more detail possibly can be
provided in the command. For instance, a hypothetical command “display a
cartoon character image” is not ambiguous just because no specific cartoon
character is specified; a hypothetical command “print screen” is not ambiguous
simply because it does not specify how many copies or which printer to use; and a
Appeal 2019-006652
Application 14/097,954

6
hypothetical command “purchase a loaf of bread” is not ambiguous merely because
it does not specify the brand or payment method. To the extent the Examiner
regards a command ambiguous simply because there is more than one way the
command may be executed and satisfied, we disagree.
None of the other full sentence commands identified by the Examiner, e.g.,
“call joe cerra,” “navigate to 17 dunster street, Cambridge Mass.,” and “send SMS
to joe cerra let’s meet at pete’s in harvard square at 7 am” (Ans. 4) is ambiguous,
much less rendered ambiguous by a deictic word in the command. The Examiner
regards “call joe cerra” as ambiguous because the phone number is not provided in
the command. Ans. 4. We disagree. The command is clear and unambiguous.
Which number to use is an execution or implementation detail and Phillips does
not describe the system contains information on two people both named Joe Cerra.
The Examiner regards “navigate to 17 dunster street, Cambridge Mass.” as
ambiguous because the geographical address and step by step directions are not
provided by the command. Id. We disagree. The command is clear and
unambiguous. The geographical address and step by step navigation are what the
system is supposed to find while executing the unambiguous command.
The Examiner regards “send SMS to joe cerra let’s meet at pete’s in harvard
square at 7 am” as ambiguous because it does not provide the actual entry for the
“To” field of the message but relies on the system to generate the “To” field. Id.
We disagree. The command is clear and unambiguous. Providing the actual entry
for the “To” field of the SMS message constitutes implementation detail and is a
part of the execution of the command.
As further evidence of ambiguity in a command, the Examiner cites to
Figure 7b of Phillips (Final Act. 6), reproduced below:
Appeal 2019-006652
Application 14/097,954

7

Figure 7b of Phillips depicts a search landing page. Phillips ¶ 32. The depiction
shows how the system asks the user what was said, among several choices. We
agree with Appellant (Appeal Br. 18–19) that that is not resolving an ambiguity of
a properly identified command. Properly identifying the words in a command has
little meaningful significance in meeting the claim.
For the foregoing reasons, what the Examiner first stated in the Final Office
Action is correct, i.e., that Phillips fails to teach “wherein the command comprises
at least one deictic word that renders the command ambiguous.” Final Act. 7–8.
Further, because Phillips does not teach a deictic word in the command that
renders the command ambiguous, it follows that Phillips does not teach
disambiguating that command by disambiguating the deictic word.
Following that route, the Examiner combines the teachings of Phillips and
Bangalore. Final Act. 7–9. In that regard, the Examiner explains that Phillips
describes using context data and user history to assist in providing specifics for
performing a user command, and that Bangalore teaches user commands which are
ambiguous because of the presence of a deictic word such as “this” or “that” in the
Appeal 2019-006652
Application 14/097,954

8
command but which can be disambiguated by parallel modes of user input, such as
gesture. Id. at 7–8 (citing Bangalore, 11:29–45, 12:44–55, 15:14–35, 16:16–40).
The assertions are supported by the cited evidence, notwithstanding Appellant’s
argument to the contrary. Based on such disclosures, the Examiner further
explains that it would have been obvious to one with ordinary skill in the art to
modify Phillips to support a command including a deictic word which renders the
command ambiguous, by adapting Phillips’s existing disambiguation, usage
history, and user models to resolve the ambiguity. Id. at 8. As an example, the
Examiner indicates that Phillips’s system would be improved to support a
command “Send SMS to ‘HIM’ lets meet at Harvard Square.” Id. at 8–9. The
Examiner also explains that Phillips describes using the user’s frequent and recent
actions, in context, to help with determining the intended user message. Ans. 4
(citing Phillips ¶¶ 48, 88, 99). The assertion is supported by the cited evidence.
To the extent Appellant argues that Phillips does not disclose using a list of recent
user actions to assist in determining user input, the argument is unpersuasive.
Phillips specifically describes “keeping track of the most frequent usages of user
speech” to inform the speech recognition facility. Phillips ¶ 48. Such tracking of
usages ranked by frequency reasonably constitutes a list.
Appellant acknowledges that Bangalore is directed to a system that receives
multi-modal input, e.g., user voice input and user gesture input. Appeal Br. 20.
However, Appellant argues that Bangalore teaches only using simultaneously
provided user voice input to disambiguate user gesture input, and not anything to
disambiguate a user voice command. Id. According to Appellant, Bangalore
discloses using a deictic word to disambiguate a deictic gesture command, whereas
the claim requires a deictic word which causes ambiguity and which is
Appeal 2019-006652
Application 14/097,954

9
disambiguated by context data. Id. at 20–21. Appellant’s argument is
unpersuasive, for reasons discussed below.
Bangalore discloses a multi-modal parallel user input system, wherein the
recognition results of any one mode can be used to constrain the recognition
process of a different mode of user input, and is not limited to only using user
gesture input to limit or disambiguate user voice input. Bangalore, 3:20–24, 6:53–
56. In particular, Bangalore discloses a specific embodiment in which the speech
recognition system uses the output of the gesture recognition system to compensate
for uncertainties in speech recognition. Bangalore, 6:29–38. Bangalore
specifically states that “the output of the gesture recognition system 200 can be
used to provide compensation to the automatic speech recognition system 100.”
Id. at 6:49–51.
Bangalore explains that its system generates a new grammar or language
model lattice for the utterance recognition subsystem 170 of the automatic speech
recognition system 100 from the gesture recognition lattice 255. Id. at 7:11–17.
This new grammar or language model lattice represents all the possible spoken
strings that can successfully combine with the particular sequence of gestures input
by the user through the gesture input portion 410. Id. at 7:22–25. Bangalore
indicates that this approach improves speech recognition based on the user’s
gestures. Id. at 7:25–30. Thus, Appellant is incorrect in its assertion that
Bangalore describes using voice input to disambiguate user’s gesture, not using
gesture input to disambiguate user voice input. Appeal Br. 20.
Finally, Appellant argues that Phillips does not teach the claimed
“automatically committing” step, because “Phillips requires additional actions
from the user to execute the provided command.” Appeal Br. 18–19 (emphasis
omitted). Appellant asserts the following:
Appeal 2019-006652
Application 14/097,954

10
As further elaborated by Phillips, “the determined action 720 is shown
in a box which allows the user to click on the down arrow or other icon
to see other action choices . . . The search button 724 allows the user
to carry out the search based on a portion of the text in the text box.”
See Phillips at [0104]. More particularly, Phillips requires the user to
press a search button to carry out the command.
Appeal Br. 19 (emphasis omitted).
The argument is misplaced, because it only refers to the one example in
Phillips that relates to searching the web and where the system requires
confirmation by the user through pressing of the search button. Phillips ¶ 104.
Appellant identifies no evidence that similar user confirmation and intervention
exists for these commands relied on by the Examiner: (1) “Send SMS to joe cerra
let’s meet at pete’s in Harvard square at 7 am”; (2) “call joe cerra”; (3) “navigate to
17 dunster street Cambridge Mass.”; and (4) “Call Mike Jones.” Ans. 4–9. Also,
the user confirmation and pressing of the search button constitutes manual
disambiguating. Thus, it is not a part of the “automatically committing” relied on
by the Examiner. Claim 1 recites “automatically committing, based on the
disambiguating, a predetermined action according to the command.” Appeal
Br. 23 (Claim App.). In that regard, the Examiner states:
An important aspect of Phillips is that the system will not
automatically commit until disambiguation takes place including
seamlessly disambiguation via at least contextual, most frequent, and
recent usage (and manual disambiguating if necessitated).
In other words, the operation of properly opening and populating
an application will not automatically proceed until disambiguation is
performed in some capacity. Similarly the system disambiguates then
automatically executes the user command to produce fig. 7c for
example.
Ans. 4–5, 7, 9, 11–12. Appellant has articulated no persuasive reason why the
above assessment by the Examiner is incorrect.
Appeal 2019-006652
Application 14/097,954

11
For the foregoing reasons, Appellant has not shown error in the Examiner’s
rejection of claim 1 as obvious over Phillips and Bangalore.
2. Claims 11 and 20
Appellant has not separately argued the merits of claims 11 and 20 from that
of claim 1. Accordingly, claims 11 and 20 stand or fall with claim 1. 37 C.F.R.
§ 41.37(c)(1)(iv). Appellant has not shown error in the Examiner’s rejection of
claims 11 and 20 as obvious over Phillips and Bangalore.
3. Claims 2–6, 8–10, 12–16, 18, and 19
Claims 2–6 and 8–10 each depend, directly or indirectly, from claim 1.
Appellant has not separately argued the merits of claims 2–6 and 8–10 from that of
claim 1. Accordingly, claims 2–6 and 8–10 stand or fall with claim 1. 37 C.F.R.
§ 41.37(c)(1)(iv).
Claims 12–16, 18, and 19 each depend, directly or indirectly, from claim 11.
Appellant has not separately argued the merits of claims 12–16, 18, and 19 from
that of claim 11. Accordingly, claims 12–16, 18, and 19 stand or fall with
claim 11. 37 C.F.R. § 41.37(c)(1)(iv).
Appellant has not shown error in the Examiner’s rejection of claims 2–6, 8–
10, 12–16, 18, and 19 as obvious over Phillips and Bangalore.

CONCLUSION
In summary:
Claims Rejected 35
U.S.C.
§
References Affirmed Reversed
1–6, 8–16, 18–20 103 Phillips,
Bangalore
1–6, 8–16,
18–20

Appeal 2019-006652
Application 14/097,954

12
No time period for taking any subsequent action in connection with this
appeal may be extended under 37 C.F.R. § 1.136(a)(1)(iv).

AFFIRMED