Ex Parte Fan et al

Board of Patent Appeals and Interferences

May 25, 2012

10737123 (B.P.A.I. May. 25, 2012)

UNITED STATES PATENT AND TRADEMARK OFFICE
_____________

BEFORE THE BOARD OF PATENT APPEALS
AND INTERFERENCES
_____________

Ex parte WEI FAN, HAIXUN WANG, and PHILIP S. YU
_____________

Appeal 2010-000577
Application 10/737,123
Technology Center 2100
______________

Before ROBERT E. NAPPI, DEBRA K. STEPHENS, and JASON V.
MORGAN, Administrative Patent Judges.

Per Curiam.

DECISION ON APPEAL
Appeal 2010-000577
Application 10/737,123
2

This is a decision on appeal under 35 U.S.C. § 134(a) of the rejection
of claims 1 through 35.

We reverse.
INVENTION
The invention is directed to a method of searching for data in
databases using an ensemble of models. See pages 1 and 2 of Appellants’
Specification. Claim 1 is representative of the invention and reproduced
below:
Claim 1. A method of searching data in databases using an ensemble of
models, said method comprising:
ordering models within said ensemble in order of prediction accuracy,
with the most accurate model being first in said order;
selecting a sub-ensemble of said models that meets a given level of
confidence, wherein models are joined together in said sub-ensemble in said
order of prediction accuracy; and
applying said sub-ensemble, in place of said ensemble, to an example
to make a prediction.

REFERENCES

Rosen US 6,513,025 B1 Jan. 28, 2003
Venkayala US 2003/0212679 A1 Nov. 13, 2003

Appeal 2010-000577
Application 10/737,123
3

REJECTION AT ISSUE
The Examiner has rejected claims 1 through 35 under 35 U.S.C.
§ 103(a) as unpatentable over Venkayla in view of Rosen. Answer 3-14.
1

ISSUES
Appellants argue on pages 27 through 31 Appeal Brief,
2
that the
Examiner’s rejection of independent claims 1, 8, 15, 21, 28, and 35 is in
error. Appellants’ arguments present us with the issue: did the Examiner err
in finding that the combination of the references teaches applying a sub-
ensemble of models in place of an ensemble of models to make a prediction
as recited in the independent claims?
3

ANALYSIS
We have reviewed Appellants’ arguments in the Brief, the Examiner’s
rejection and the Examiner’s response to the Appellants’ arguments. We
concur with Appellants’ conclusion that the Examiner erred in finding that
the combination of the references teaches applying a sub-ensemble of
models in place of an ensemble of models to make a prediction as recited in
the independent claims.

1
Throughout this opinion we refer to the Examiner’s Answer mailed on
July 8, 2009.
2
Throughout this opinion we refer to Appellants’ Appeal Brief dated March
19, 2009.
3
We note Appellants’ arguments present additional issues with respect to
both the independent claims and dependent claims, however as this issue is
dispositive we do not reach the other issues presented.
Appeal 2010-000577
Application 10/737,123
4

The Examiner cites to Venkayala’s teaching of building models using
seed models as teaching using more than one model. Answer 19. Further,
the Examiner equates Venkayala’s step of selection of criterion of the
selected class values with the claimed applying sub-ensembles. Answer 20.
Appellants argue that this does not meet the claims. Appeal Brief 28-29.
We concur with the Appellants. The claims identify that the ensembles and
sub-ensembles are of models, we do not find that the selection of criterion of
the selected class values are sub-ensembles of models. While we concur
with the Examiner that Venkayala teaches using more than one model in that
a model is generated using seed models, the Examiner has not shown how
this teaching meets the step of applying sub-ensemble of models as claimed.
Thus, we are persuaded of error in the Examiner’s obviousness rejection and
we will not sustain the Examiner’s rejection.

ORDER
The decision of the Examiner to reject claims 1 through 35 is
reversed.

REVERSED

dw