Microsoft Technology Licensing, LLC

Patent Trials and Appeals Board

Nov 23, 2021

2020004806 (P.T.A.B. Nov. 23, 2021)

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.
15/476,350 03/31/2017 Jason Wik 401538-US-NP (9020-US) 1338
143198 7590 11/23/2021
MICHAEL BEST & FRIEDRICH LLP (MS)
790 N WATER ST
SUITE 2500
MILWAUKEE, WI 53202
EXAMINER
SISON, JUNE Y
ART UNIT PAPER NUMBER
2443
NOTIFICATION DATE DELIVERY MODE
11/23/2021 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):
mkeipdocket@michaelbest.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte JASON WIK, SARAVANA KUMAR CHELLAPPAN, and
ERIC E. KNUTSEN

Appeal 2020-004806
Application 15/476,350
Technology Center 2400

Before JEAN R. HOMERE, AMBER L. HAGY, and
SCOTT RAEVSKY, Administrative Patent Judges.

HOMERE, Administrative Patent Judge.

DECISION ON APPEAL
I. STATEMENT OF THE CASE1
Pursuant to 35 U.S.C. § 134(a), Appellant appeals from the
Examiner’s rejection of claims 1–20, all of the pending claims.2 Claims App.
We have jurisdiction under 35 U.S.C. § 6(b).
We reverse.

1 We refer to the Specification filed Mar. 31, 2017 (“Spec.”); the Final
Office Action, mailed Aug. 15, 2019 (“Final Act.”); the Appeal Brief, filed
Jan. 15, 2020 (“Appeal Br.”); the Examiner’s Answer, mailed Apr. 9, 2020
(“Ans.”); and the Reply Brief, filed June 9, 2020 (“Reply Br.”).
2 “Appellant” refers to “applicant” as defined in 37 C.F.R. § 1.42(a).
Appellant identifies Microsoft Technology Licensing, LLC, as the real
party-in-interest. Appeal Br. 3.
Appeal 2020-004806
Application 15/476,350

2
II. CLAIMED SUBJECT MATTER
According to Appellant, the claimed subject matter relates to a
method and system for “predicting how many users of a computer system
have used, are currently using, or will use the computer system to detect and
plan for system outages.” Spec. ¶ 1. Fig. 1, reproduced and discussed below,
is useful for understanding the claimed subject matter:

Figure 1 above illustrates system 100 including prediction server 102
configured to predict computer system usage (e.g., network server 106
hosting application 112 for a plurality of users). Spec. ¶¶ 17, 18. In
particular, prediction server 102 includes electronic processor 202 coupled
to electronic database 104 storing historical windowed data points, current
windowed data points, user data, and user activity data. Id ¶ 23. Upon
receiving from network server 106 a data stream including a plurality of
Appeal 2020-004806
Application 15/476,350

3
historical activity indicators including unique user identifiers and
timestamps for the plurality of users, electronic processor 202 arranges the
historical activity indicators into a sequential time series based on the
plurality of timestamps. Id ¶ 28. Electronic processor 202 then groups the
historical activity indicators into data bins of a determined size to generate a
plurality of historical windowed data points based on the plurality of
timestamps. Id. Electronic processor 202 subsequently determines a
forecasted quantity of users for a forecast time window based on the
plurality of historical windowed data points and the plurality of unique user
identifiers to transmit the forecasted quantity of users to a user device. ¶¶ 32,
34.
Claims 1, 10, and 16 are independent. Claim 1, reproduced below
with disputed limitations emphasized, is illustrative:
1. A system comprising:
a server communicatively coupled to a database, the server
including an electronic processor configured to receive, from a
network server hosting at least one application for a plurality of
users, a data stream including a plurality of historical activity
indicators for the plurality of users, the plurality of historical activity
indicators including a plurality of unique user identifiers and a
plurality of timestamps;
arrange the historical activity indicators into a sequential time
series based on the plurality of timestamps;
group the plurality of historical activity indicators into data bins
of a determined size to generate a plurality of historical windowed
data points based on the plurality of timestamps;
determine a forecasted quantity of users for a forecast time
window based on the plurality of historical windowed data points and
the plurality of unique user identifiers; and
transmit, to a user device, the forecasted quantity of users.

Appeal Br. 15 (Claims App.) (emphasis added).
Appeal 2020-004806
Application 15/476,350

4
III. REFERENCES
The Examiner relies upon the following references.3
Name Reference Date
Anderholm US 2005/0183143 A1 Aug. 18, 2005
Ferreira US 2010/0295856 A1 Nov. 25, 2010
Siciliano US 2015/0278061 A1 Oct. 1, 2015
Wang US 2016/0092774 A1 Mar. 31, 2016
Dean US 10,169,711 B1 Jan. 1, 2019
IV. REJECTIONS
The Examiner rejects the claims on appeal as follows:
Claims 1, 2, 8, 9–12, and 15–17 stand rejected under 35 U.S.C.
§ 103 as being unpatentable over the combined teachings of Wang, Ferreira,
and Anderholm. Final Act. 4–13.
Claims 3, 13, and 18 stand rejected under 35 U.S.C. § 103 as being
unpatentable over the combined teachings of Wang, Ferreira, Anderholm,
and Sicialon. Final Act. 13–14.
Claims 4–7, 14, 15, 19, and 20 stand rejected under 35 U.S.C. § 103
as being unpatentable over the combined teachings of Wang, Ferreira,
Anderholm, and Dean. Final Act. 14–19.

3 All reference citations are to the first named inventor only.
Appeal 2020-004806
Application 15/476,350

5
V. ANALYSIS
We consider Appellant’s arguments, as they are presented in the
Appeal Brief, pages 10–13 and the Reply Brief, pages 2–5.4 We are
persuaded by Appellant’s contentions as discussed below.
Appellant argues, inter alia, that the Examiner errs in finding that the
combination of Wang, Ferreira, and Anderholm teaches or suggests
“determin[ing] a forecasted quantity of users for a forecast time window
based on the plurality of historical windowed data points and the plurality of
unique user identifiers, as recited in independent claim 1. Appeal Br. 11. In
particular, Appellant argues that the Examiner erroneously finds Wang’s
disclosure of comparing the behavior of a single user to that of a community
of users during a same time period to determine whether the user’s behavior
is anomalous teaches or suggests the disputed limitations. Id. at 11 (citing
Wang ¶ 4, 7, 8, 21). More particularly, Appellant states the following:
Wang teaches the use of graphs including nodes (users) an
edges (relationships between users). Id. at ¶ [21]. Feature values are
determined based on the relationships between the nodes (users).
Id. at ¶ [23]. Wang discloses that “feature values” are characteristics
of user nodes in a unipartite graph. Id. at ¶ [45]. Wang’s user features
include the degree of the user node, the local clustering coefficient of
the user node, the closeness of the user node, the “betweenness” of the
user node, and the eigenvector centrality of the user node. Id. at
¶¶ [46–50]. While each of these is a mathematical value, none are
either explicitly or implicitly a “quantity of the users.” When a
community of users is considered in Wang, the aggregate values for
the features are determined and compared, not the quantity of the
users. Furthermore, even assuming for the pure sake of argument that
Wang’s determinations of anomalous behavior were equivalent to

4 We have considered in this Decision only those arguments Appellant
actually raised in the Briefs. Arguments not made are waived. See
37 C.F.R. § 41.37(c)(1)(iv).
Appeal 2020-004806
Application 15/476,350

6
determining and comparing quantities of users, Wang still would not
disclose the claimed subject matter. Comparing a single user to
single user to detect anomalous behavior always has a quantity of
one. Similarly, comparing a fixed user community to itself as
disclosed in Wang also fails to teach the claimed subject matter.
Finally, for a comparison between a single user and a community,
there would naturally be a difference in a number of users, but this
does not provide any insight as to whether the single user's behavior is
anomalous.
Id. at 11−12 (emphases added).
In response, the Examiner finds the following:
[D]isclosures by Wang of aggregating single users (i.e. claimed
quantity of users) and comparing a fixed user(s) community to itself
are analogous to examples given in appellant's own specification.
That is, per appellant’s own specification, for example, systems and
methods use historical activity indicators to indicate which user(s)
perform which activities and when - and these particular subsets of
historical data points from past instances are selected to predict a
particular quantity of users (i.e. comparing a fixed user community
to itself) (see appellant’s specification [0006]). Per appellant’s own
specification, for example, such "comparing a fixed user community
to itself" may be used to predict how many particular users perform
particular activities may be affected by an outage - such as a planned
outage - so that system administrator(s) can plan/ upgrade a particular
service with minimal impact on particular users who use that
particular service at particular times (i.e. comparing a fixed user
community to itself) (see appellant’s specification [0036])
(appellant’s specification [006; 36] given below - emphasis added
by examiner - bold and/or underline).
Ans. 6.
Appellant’s arguments are persuasive of reversible Examiner error.
As noted above, the disputed claim limitations recite “determin[ing] a
forecasted quantity of users for a forecast time window based on the
plurality of historical windowed data points and the plurality of unique user
identifiers.” As an initial matter, we note that the cited claim language must
Appeal 2020-004806
Application 15/476,350

7
be given its broadest reasonable interpretation consistent with Appellant’s
disclosure, as explained in Morris:
[T]he PTO applies to the verbiage of the proposed claims the
broadest reasonable meaning of the words in their ordinary usage
as they would be understood by one of ordinary skill in the art,
taking into account whatever enlightenment by way of
definitions or otherwise that may be afforded by the written
description contained in the applicant’s specification.
In re Morris, 127 F.3d 1048, 1054 (Fed. Cir. 1997); see also In re Zletz, 893
F.2d 319, 321 (Fed. Cir. 1989) (stating that “claims must be interpreted as
broadly as their terms reasonably allow”). Our reviewing court further states,
“the proper . . . construction is not just the broadest construction, but rather
the broadest reasonable construction in light of the specification.” In re Man
Mach. Interface Techs. LLC, 822 F.3d 1282, 1287 (Fed. Cir. 2016).
As noted by the Examiner, and undisputed by Appellant, the
Specification refers to aggregating into time windows and comparing
“historical activity indicators” and current activity indicators stored in the
database to generate a “forecasted quantity of users” during a particular time
window. Ans. 4; Reply 4 (citing Spec. ¶¶ 39, 40). Accordingly, we agree
with Appellant that the broadest reasonable interpretation of “forecasted
quantity of users for a forecast time window” refers to a predicted amount of
users during the forecast time window. Reply Br. 3, 4 (citing Spec. ¶¶ 33,
36). Therefore, the disputed claim limitations require comparing aggregated
historical activity indicators with aggregated current indicators to forecast
the amount of users (of an application) for that a particular time window.
Wang discloses aggregating a plurality of window points based on a
plurality of user identifiers in a community of users, and comparing window
point of a single user to an aggregated window points of the community
Appeal 2020-004806
Application 15/476,350

8
users to determine whether the behavior of a single user is anomalous. Wang
¶¶ 21, 32–36. More particularly, Wang discloses the following:
The community identification engine 238 receives
user graph data 225 and determines one or more communities
within the user graph data 225, e.g., by assigning each node in
the user graph to one of multiple communities. The community
identification engine 238 can then provide the community
data to the prediction engine 136 and the feature engine
234.

Spec. ¶ 33.

The feature engine 234 generates user feature values
for distinct users within a particular time interval as well as
aggregate community feature values for a particular community
within a particular time interval. The feature engine 234
can provide the computed features both to the anomaly detection
engine 232 for detecting anomalies and to the prediction
engine 236 for building user and community prediction models.

Id. ¶ 34.

The prediction engine 136 generates a user model
for each distinct user that occurs in the user graph data 225.
The prediction engine also generates a community model for
each distinct community identified in the user graph data 225.
The user and community models can predict the user feature
values or community feature values of the user or community
within a subsequent time interval.

Id. ¶ 35.

The anomaly detection engine 232 receives observed user and
community feature values 233 for a user or a community within a
particular time interval. The anomaly detection engine 232 also
receives predicted user and community feature values 237. The
anomaly detection engine can then determine whether a user or a
community anomaly is present in the user graph data 225 for a
Appeal 2020-004806
Application 15/476,350

9
particular time interval and provide an appropriate notification to the
master node 220.

Id. ¶ 36.

Although Wang’s aggregation of users in a community indicates the
number of users within that community for a particular time period, we
agree with Appellant that the disclosed comparison of the aggregated
community users with a particular user within the same community does not
result in forecasting the amount of users within that particular time window.
As persuasively argued by Appellant, the disclosed comparison at best
teaches forecasting whether the user’s behavior is anomalous, as opposed to
forecasting the amount of users for that particular time window. Reply Br.
4–5. Accordingly, we are persuaded that the cited portions of Wang do not
teach or suggest the disputed limitations.
Because Appellant shows at least one reversible error in the
Examiner’s obviousness rejection of independent claim 1, we do not reach
Appellant’s remaining arguments. Accordingly, we do not sustain the
Examiner’s obviousness rejection of independent claim 1 obvious over the
combination of Wang, Ferreira, and Anderholm. Similarly, we do not sustain
the rejections of claims 2–20, which also recite the disputed limitations.
VI. CONCLUSION
We reverse the Examiner’s rejections of claims 1–20.
Appeal 2020-004806
Application 15/476,350

10
VII. DECISION SUMMARY
In summary:
Claims
Rejected
35
U.S.C. § Reference(s)/Basis Affirmed Reversed
1, 2, 8, 9−12,
15–17 103
Wang, Ferreira,
Anderholm
1, 2, 8, 9–
12, 15–17
3, 13, 18 103
Wang, Ferreira,
Anderholm, and
Sicialon
3, 13, 18
4–7, 14, 15, 19,
20 103
Wang, Ferreira,
Anderholm, Dean
4–7, 14,
15, 19, 20
Overall
Outcome 1–20

REVERSED