Ex Parte Chen et al

Board of Patent Appeals and Interferences

Mar 9, 2010

10600991 (B.P.A.I. Mar. 9, 2010)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________
BEFORE THE BOARD OF PATENT APPEALS
AND INTERFERENCES
____________
Ex parte DINGDING CHEN, LUIS E. SAN MARTIN, GULAMABBAS A.
MERCHANT, ROBERT W. STRICKLAND, and MARTIN T. HAGAN
____________
Appeal 2008-004508
Application 10/600,9911
Technology Center 2100
____________
Decided: March 9, 2010
____________

Before JOSEPH L. DIXON, JEAN R. HOMERE, and THU A. DANG,
Administrative Patent Judges.

HOMERE, Administrative Patent Judge.

DECISION ON APPEAL

1 Filed on June 19, 2003. The real party in interest is Halliburton Energy
Services, Inc. (Br. 3.)
Appeal 2008-004508
Application 10/600,991

2
I. STATEMENT OF THE CASE
Appellants appeal under 35 U.S.C. § 134(a) (2002) from the
Examiner’s final rejection of claims 1 through 3 and 5 through 9. (Ans. 2.)2
Claims 10 through 25 have been allowed. (Id. at 2-3.) Claim 4 is objected
to as being dependent upon a rejected based claim, but would be allowable if
rewritten in independent form including all of the limitations of the base
claim and any intervening claims. (Id.) We have jurisdiction under 35
U.S.C. § 6(b) (2008).
We affirm.
Appellants’ Invention
Appellants invented a method and apparatus for training a neural
network to process signals from a logging tool into a representation of
formation parameters. (Spec. 1, Para. [0002].) According to Appellants, the
claimed invention selects and scales earth formation models based on the
behavior of a particular well logging tool. (Spec. 12-13, Para. [0048].)
Further, the claimed invention utilizes the respective earth formation models
to train artificial neural networks (“ANNs”) which provide good inversion of
log data without geographical area limitation. (Id.)
Illustrative Claim
Independent claim 1 further illustrates the invention as follows:
1. Apparatus for converting output signals of a logging tool
into a log representing a parameter of earth formations
surrounding a borehole, comprising:
an artificial neural network trained with a set of synthetic
earth formation models selected to cover the

2 All references to the Examiner’s Answer are to the Examiner’s Answer
filed on June 29, 2007, which replaced the prior Examiner’s Answer filed on
May 9, 2007.
Appeal 2008-004508
Application 10/600,991

3
operating range of a selected logging tool based on
sensitivity and resolution limits of the logging tool
and based on realistic ranges of formation
parameters.

Prior Art Relied Upon
The Examiner relies on the following prior art as evidence of
unpatentability:
Anderson 3,954,006 May 4, 1976
Barber 5,184,079 Feb. 2, 1993
Mezzatesta 5,862,513 Jan. 19, 1999

Rejections on Appeal3
The Examiner rejects the claims on appeal as follows:
Claims 1 through 3 and 5 stand rejected under 35 U.S.C. § 102(b) as
being anticipated by Mezzatesta.
Claims 6 and 9 stand rejected under 35 U.S.C. § 103(a) as being
unpatentable over the combination of Mezzatesta and Anderson.
Claims 7 and 8 stand rejected under 35 U.S.C. § 103(a) as being
unpatentable over the combination of Mezzatesta and Barber.
Appellants’ Contentions
Appellants contend that Mezzatesta’s disclosure of deriving a training
set from well logging data amounts to a training set based on an actual
formation. (Br. 9-10.) Therefore, Appellants argue that Mezzatesta’s
training set does not teach a training set “selected to cover the operating
range of a selected logging tool,” as recited in independent claim 1. (Id.)
Further, Appellants allege that Mezzatesta’s disclosure is directed to

3 The Examiner withdrew the 35 U.S.C. § 101 rejection of claims 1 through
4, 10 through 12, and 20. (Ans. 3.)

Appeal 2008-004508
Application 10/600,991

4
utilizing trained neural networks, whereas the claimed invention utilizes data
to train neural networks. (Id. at 10.) Additionally, Appellants argue that
Mezzatesta’s disclosure of generating synthetic models based on data from a
particular tool in various earth formations does not teach a model that falls
within the operation range of a logging tool, as claimed. (Id. at 11.)
Examiner’s Findings and Conclusions
The Examiner finds that Mezzatesta’s disclosure of obtaining raw data
from a particular tool in actual formations in order to produce an input earth
model that includes synthetic data teaches synthetically developing data for
both an actual tool and formation. (Ans. 13.) In particular, the Examiner
finds that Mezzatesta’s input models include synthetically developed data
specifically formulated for a particular tool and, further, based on realistic
ranges of formation parameters. (Id.) Therefore, the Examiner finds that
Mezzatesta’s disclosure teaches input models that include synthetic data
“selected to cover the operating range of a selected logging tool,” as recited
in independent claim 1. (Id. at 14.) Further, the Examiner finds that
Mezzatesta’s disclosure of introducing an input earth model to an ANN
teaches training neural networks, as claimed. (Id.)

II. ISSUE
Have Appellants shown that the Examiner erred in finding that
Mezzatesta anticipates independent claim 1? In particular, the issue turns on
whether Mezzatesta teaches “an artificial neural network trained with a set
of synthetic earth formation models selected to cover the operating range of
a selected logging tool,” as recited in independent claim 1.

Appeal 2008-004508
Application 10/600,991

5
III. FINDINGS OF FACT
The following Findings of Fact (“FF”) are shown by a preponderance
of the evidence.
Mezzatesta
1. Mezzatesta is generally related to well logging data
interpretation systems and methods which use ANNs and, in particular, a
system and method that is usable at a well site to predict well logging tool
responses, unknown formation parameters, and reservoir descriptions. (Col.
1, ll. 20-25.)
2. Mezzatesta discloses utilizing “an ANN to receive an input
earth model or models (based on real data actually acquired by a real tool)
and output an appropriate tool response.” (Col. 3, ll. 38-41.) “A separate
ANN…is created for each type of well-logging tool. Each such ANN may
have its own training dataset of well-log tool responses (these training
datasets contain measured actual raw data for a real formation or synthetic
responses derived based on such data).” (Id. at ll. 43-48.)
3. Mezzatesta discloses that “[a] well logging tool acquires data
from a formation for an initial data set which is used to produce an earth
model (which may be a group of models)…”. (Col. 7, ll. 35-38.) “[B]ased
on known raw data obtained by a particular tool in various actual formations,
an input earth model is produced that includes synthetic rather than raw
data.” (Id. at ll. 38-41.)

Appeal 2008-004508
Application 10/600,991

6
IV. PRINCIPLES OF LAW
Anticipation
In rejecting claims under 35 U.S.C. § 102, “[a] single prior art
reference that discloses, either expressly or inherently, each limitation of a
claim invalidates that claim by anticipation.” Perricone v. Medicis Pharm.
Corp., 432 F.3d 1368, 1375 (Fed. Cir. 2005) (citing Minn. Mining & Mfg.
Co. v. Johnson & Johnson Orthopaedics, Inc., 976 F.2d 1559, 1565 (Fed.
Cir. 1992)).
Anticipation of a patent claim requires a finding that the claim
at issue “reads on” a prior art reference. In other words, if
granting patent protection on the disputed claim would allow
the patentee to exclude the public from practicing the prior art,
then that claim is anticipated, regardless of whether it also
covers subject matter not in the prior art.

Atlas Powder Co. v. IRECO, Inc., 190 F.3d 1342, 1346 (Fed Cir. 1999)
(internal citations omitted).
Obviousness
“On appeal to the Board, an applicant can overcome a rejection [under
§ 103] by showing insufficient evidence of prima facie obviousness or by
rebutting the prima facie case with evidence of secondary indicia of
nonobviousness.” In re Rouffet, 149 F.3d 1350, 1355 (Fed. Cir. 1998))
(citation omitted).

Appeal 2008-004508
Application 10/600,991

7
V. ANALYSIS
35 U.S.C. § 102(b) Rejection
Claim 1
Independent claim 1 recites, in relevant part, “an artificial neural
network trained with a set of synthetic earth formation models selected to
cover the operating range of a selected logging tool.”
As detailed in the Findings of Fact section, Mezzatesta discloses
utilizing an ANN to predict the responses of well logging tools. (FF 1.) In
particular, Mezzatesta discloses training an ANN using one or more earth
models that contain synthetic response data. (FF 2.) Further, Mezzatesta
discloses deriving the synthetic response data from raw data obtained by a
particular well logging tool in various formations. (FF 3.)
We find that Mezzatesta’s disclosure of obtaining raw data by
utilizing a well logging tool in a specific formation amounts to creating a
range of synthetic response data particular to the operations of a well
logging tool. Consequently, by utilizing a well logging tool in multiple
formations to obtain raw data, Mezzatesta teaches creating synthetic
response data for a wider range of the well logging tool’s operations. We
thus find that such a wider range reasonably approaches the limits of the
operating range of a selected well logging tool. Thus, we find that
Mezzatesta’s cited disclosure teaches the disputed limitation. It follows that
Appellants have not shown that the Examiner erred in finding that
Mezzatesta anticipates independent claim 1.
Claims 2 and 3
Appellants do not provide separate arguments for patentability with
respect to dependent claims 2 and 3. Therefore, we select independent claim
Appeal 2008-004508
Application 10/600,991

8
1 as representative of the cited claims. Consequently, Appellants have not
shown error in the Examiner’s rejection of dependent claims 2 and 3 for the
reasons set forth in our discussion of independent claim 1. See 37 C.F.R. §
41.37(c)(1)(vii) (2008).
Claim 5
Appellants contend that Mezzatesta’s disclosure teaches the use of a
neural network for forward modeling, whereas the claimed invention recites
a neural network for direct inversion of tool logging signals. (Br. 12.)
Therefore, Appellants argue that Mezzatesta does not teach “train[ing] an
artificial neural network to generate representations of the formation models
in response to the synthetic responses,” as recited in independent claim 5.
(Id.) We do not agree.
First, Appellants’ assertion that Mezzatesta’s disclosure teaches the
use of a neural network for forward modeling, whereas the claimed
invention recites a neural network with direct inversion of tool logging
signals (id), is not commensurate with the scope of independent claim 5.
Nowhere does independent claim 5 recite “a neural network with direct
inversion of tool logging signals.” Second, as set forth above, we find that
Mezzatesta’s disclosure teaches training an ANN by using one or more earth
models containing synthetic response data. (FF 2.) In particular, we find
that Mezzatesta’s disclosure of training an ANN utilizing synthetic response
data to produce one or more earth models amounts to “train[ing] an artificial
neural network to generate representations of the formation models in
response to the synthetic responses,” as recited in independent claim 5. It
follows that Appellants have not shown that the Examiner erred in finding
that Mezzatesta anticipates independent claim 5.
Appeal 2008-004508
Application 10/600,991

9
35 U.S.C. § 103(a) Rejections
Claims 6 through 9
Appellants offer the same argument set forth in response to the
anticipation rejection to rebut the obviousness rejection of claims 6 through
9. (Br. 14.) We have already addressed this argument in our discussion of
claim 5 above, and we found it unpersuasive. Consequently, Appellants
have not shown error in the Examiner’s conclusion that: 1) claims 6 and 9
are unpatentable over the combination of Mezzatesta and Anderson; and 2)
claims 7 and 8 are unpatentable over the combination of Mezzatesta and
Barber.

VI. CONCLUSIONS OF LAW
Appellants have not shown that the Examiner erred in rejecting claims
1 through 3 and 5 as being anticipated under 35 U.S.C. § 102(b).
Appellants have not shown that the Examiner erred in rejecting claims
6 through 9 as being unpatentable under 35 U.S.C. § 103(a).

VII. DECISION
We affirm the Examiner’s decision to reject claims 1 through 3 and 5
through 9.

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

AFFIRMED

Appeal 2008-004508
Application 10/600,991

10

nhl

CONLEY ROSE, P.C.
David A. Rose
PO BOX 3267
HOUSTON, TX 77253-3267