Ex Parte Doddi et al

Board of Patent Appeals and Interferences

Jan 8, 2009

10608300 (B.P.A.I. Jan. 8, 2009)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________

BEFORE THE BOARD OF PATENT APPEALS
AND INTERFERENCES
____________

Ex parte SRINIVAS DODDI, EMMANUEL DREGE,
NICKHIL JAKATDAR, and JUNWEI BAO
____________

Appeal 2008-4716
Application 10/608,3001
Technology Center 2100
____________

Decided: January 8, 2009
____________

Before LANCE LEONARD BARRY, ST. JOHN COURTENAY, III, and
CAROLYN D. THOMAS, Administrative Patent Judges.

THOMAS, Administrative Patent Judge.

DECISION ON APPEAL

1 Application filed June 27, 2003. The real party in interest is Timbre
Technologies, Inc.
Appeal 2008-4716
Application 10/608,300

2

I. STATEMENT OF THE CASE
Appellants appeal under 35 U.S.C. § 134 from a final rejection
of claims 1-29 mailed August 11, 2006, which are all the claims remaining
in the application. We have jurisdiction under 35 U.S.C. § 6(b).
We affirm.

A. INVENTION
Appellants invented a system, method, and computer readable storage
medium for examining a structure formed on a semiconductor wafer by
obtaining a first diffraction signal measured using a metrology device and
comparing it to a second diffraction signal generated using a machine
learning system. The machine learning system receives input parameters
that characterize a profile of the structure to generate the second diffraction
signal. When the first and second diffraction signals match within a
matching criterion, a feature of the structure is determined. (Spec., Abstract.)

B. ILLUSTRATIVE CLAIM
The appeal contains claims 1-29. Claims 1, 16, and 22 are
independent claims. Claim 1 is illustrative:
1. A method of examining a structure formed on a
semiconductor wafer, the method comprising:
obtaining a first diffraction signal measured using a
metrology device;
obtaining a second diffraction signal generated using a
machine learning system,
wherein the machine learning system receives as
an input one or more parameters that characterize a profile of the
Appeal 2008-4716
Application 10/608,300

3
structure to generate the second diffraction signal as an output of the
machine learning system;
comparing the first and second diffraction signals; and
when the first and second diffraction signals match
within a matching criterion, determining a feature of the structure
based on the one or more parameters of the profile used by the
machine learning system to generate the second diffraction signal.

C. REFERENCES
The references relied upon by the Examiner in rejecting the claims on
appeal are as follows:
Wormington US 6,192,103 B1 Feb. 20, 2001
Singh US 6,650,422 B2 Nov. 18, 2003
Kato US 6,665,446 B1 Dec. 16, 2003
Sirat EP 0 448 890 A1 Oct. 2, 1991

Mark Gahegan et al., Dataspaces as an organizational concept for the
neural classification of geographic datasets, Department of Geography, The
Pennsylvania State University (1999).

D. REJECTIONS
The Examiner entered the following rejections which are before us for
review:
(1) Claims 1-6, 11-14, and 16-29 are rejected under 35 U.S.C.
§ 103(a) as being unpatentable over Singh in view of Wormington.
(2) Claims 9, 10, and 15 are rejected under 35 U.S.C.
§ 103(a) as being unpatentable over Singh in view of Wormington and
further in view of Kato.
(3) Claim 7 is ejected under 35 U.S.C. § 103(a) as being unpatentable
over Singh in view of Wormington and further in view of Sirat.
Appeal 2008-4716
Application 10/608,300

4
(4) Claim 8 is ejected under 35 U.S.C. § 103(a) as being unpatentable
over Singh in view of Wormington and further in view of Gahegan.

II. PROSECUTION HISTORY
Appellants appealed from the Final Rejection and filed an Appeal
Brief (App. Br.) on February 12, 2007. The Examiner mailed an Examiner’s
Answer (Ans.) on June 12, 2007. Appellants filed a Reply Brief (Reply Br.)
on August 13, 2007. An Oral Hearing was held at the U.S. Patent and
Trademark Office on December 11, 2008.

III. FINDINGS OF FACT
The following findings of fact (FF) are supported by a preponderance
of the evidence.
Specification
1. The Specification does not provide a lexicographic definition for
the term “machine learning system.”
2. The Specification discloses that “diffraction signals used in a
library-based process and/or a regression-based process are generated using
a machine learning system 118 employing a machine learning algorithm
(Spec., ¶[0032]).
3. The Specification discloses that the “machine learning system 118
receives a profile as an input and generates a diffraction signal as an output”
(Spec., ¶[0033]).

Appeal 2008-4716
Application 10/608,300

5
Wormington
4. Wormington discloses that “[e]volutionary algorithms are used to
find a global solution to the fitting of experimental X-ray scattering data to
simulated models” (Abstract).
5. Wormington discloses that “it is desirable to fit measured, or
experimental, data to a simulated model to determine characterizing
parameters of a structure” (col. 4, ll. 65-67).
6. Wormington discloses:
Referring to the figure, the fitting procedure begins by
measuring the X-ray scattering data for a specimen being tested,
at step 30. A model for that specimen is then estimated at step
32, . . . Once the model has been estimated, the X-ray scattering
for that model is simulated at step 34, using known methods . . .
to produce a characteristic curve. The differences between the
values for that curve and the data which was measured at step
30 are determined at step 36, . . . At step 38, a determination is
made whether the error value is less than a threshold value T. If
it is not, one or more parameters of the model are adjusted at
step 40. . . . After the model parameters have been adjusted,
steps 34, 36 and 38 are repeated.

(Col. 6, ll. 5-32.)

IV. PRINCIPLES OF LAW
“What matters is the objective reach of the claim. If the claim extends
to what is obvious, it is invalid under § 103.” KSR Int’l Co. v. Teleflex, Inc.,
127 S. Ct. 1727, 1742 (2007). To be nonobvious, an improvement must be
“more than the predictable use of prior art elements according to their
established functions.” Id. at 1740.
Appeal 2008-4716
Application 10/608,300

6
Appellants have the burden on appeal to the Board to demonstrate
error in the Examiner’s position. See In re Kahn, 441 F.3d 977, 985-86
(Fed. Cir. 2006) (“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.”) (quoting In re Rouffet, 149 F.3d 1350, 1355
(Fed. Cir. 1998)). Therefore, we look to Appellants’ Brief to show error in
the proffered prima facie case. Only those arguments actually made by
Appellants have been considered in this decision. Arguments which
Appellants could have made but chose not to make in the Brief has not been
considered and are deemed to be waived. See 37 C.F.R. § 41.37(c)(1)(vii).

V. ANALYSIS
Grouping of Claims
In the Brief, Appellants argue claims 1-29 as a group (App. Br. 5-7).
For claims 2-29, Appellants repeat the same argument made for claim 1. We
will, therefore, treat claims 2-29 as standing or falling with claim 1. See 37
C.F.R. § 41.37(c)(1)(vii). See also In re Young, 927 F.2d 588, 590 (Fed. Cir.
1991).

The Board's Claim Construction
"Our analysis begins with construing the claim limitations at issue."
Ex Parte Filatov, No. 2006-1160, 2007 WL 1317144, at *2 (BPAI 2007).
Claims are given their broadest reasonable construction “in light of the
specification as it would be interpreted by one of ordinary skill in the art.”
In re Am. Acad. of Sci. Tech. Ctr., 367 F.3d 1359, 1364 (Fed. Cir. 2004).
Appeal 2008-4716
Application 10/608,300

7
“Without evidence in the patent specification of an express intent to
impart a novel meaning to a claim term, the term takes on its ordinary
meaning.” Optical Disc Corp. v. Del Mar Avionics, 208 F.3d 1324, 1334
(Fed. Cir. 2000). Here, Appellants have not presented any evidence to show
that the phrase “machine learning system” includes a novel meaning in the
specification. Therefore, we shall adopt the ordinary meaning of “machine
learning system” which includes a system that receives a profile as an input
and through employing a machine learning algorithm generates desired
outputs (FF 1-3).

The Obviousness Rejection
We now consider the Examiner’s rejection of claims 1-29 under 35
U.S.C. § 103(a).
Appellants contend:
Note, claims 1, 16, and 22 do not merely recite that the second
simulated diffraction signal is generated as an output, or that the
machine learning system is used in generating the second
simulated diffraction signal. Instead, claims 1, 16, and 22
expressly recite that the output of the machine learning system
is the second simulated diffraction signal.

(App. Br. 5.)
Appellants further contend that “the Examiner must establish that the
X-ray scattering disclosed in the Wormington reference is generated as an
output of the genetic and evolutionary algorithms” (App. Br. 6).
Appellants also contend that “the simulated x-ray scattering data is the
output of the ‘calculating’ step rather than the ‘modifying the model’ step in
which the evolutionary algorithm is used” (Reply Br. 6). Further,
Appeal 2008-4716
Application 10/608,300

8
Appellants contend that “[r]egardless the number of times that the loop is
completed, the Wormington reference does not disclose the use of genetic
algorithms at steps 32 or 34 to output X-ray scattering” (Reply Br. 9).

The Examiner found that “step 34 is functionally part of the genetic
algorithm. Thus, the data representing two (or more) X-ray scatterings (or
diffraction signals) is clearly a result of computations inside the genetic
algorithm of Wormington’s invention and could be indicted to be the output
of the genetic algorithm with the appropriate flowchart symbol” (Ans. 12).

Issue: Whether Appellants have shown that the Examiner erred in
finding that Wormington’s step 34 which outputs that X-ray scattering curve
is functionally apart of the genetic algorithm, i.e., included in the machine
learning system?

Initially, we note that the Examiner has found, and Appellants do not
dispute, that Wormington’s genetic and evolutionary algorithms (e.g., Fig. 4,
step 40) are machine learning algorithms (Ans. 3). However, Appellants
contend that because the Examiner has equated Wormington’s evolutionary
algorithm at step 40 as a learning algorithm, the Examiner must further
establish that the X-ray scattering disclosed in the Wormington reference is
generated as an output of step 40, instead of step 34.
In other words, Appellants argue that because the Examiner has
construed Wormington’s step 40 as a machine learning algorithm, to read on
the claimed language, the X-ray scattering signal must come directly from
Wormington’s step 40. We disagree.
Appeal 2008-4716
Application 10/608,300

9
Firstly, we are unaware of any authority holding that we are bound by
an Examiner’s determination as to claim construction. As noted supra, we
construe a “machine learning system” as including any system that receives
a profile as an input and through employing a machine learning algorithm
generates desired outputs. We further note that such a “system” could
embody a single unit or multiple units. Appellants’ claim 1 does not limit its
“machine learning system” to a particular embodiment. All that is required is
that a “learning algorithm” by used therein. Further, the term “system” in
itself suggests “a group of interacting bodies under the influence of related
forces.” Merriam-Webster’s Collegiate Dictionary, p. 1269 (11th ed. 2003).
Secondly, the Examiner found that step 34 could be functionally apart
of the genetic algorithm (Ans. 12) and that the boundary of the genetic
algorithm does not end at step 40 (Ans. 11). We agree.
Appellants claim “a machine learning system” not merely a learning
algorithm. Thus, we note that although Appellants argue that the X-ray
scattering data must be an output of Wormington’s step 40, Appellants have
chosen to draft the claims, claim 1 in particular, far more broadly.
Wormington’s system uses an evolutionary algorithm to find X-ray
scattering data (FF 4) which is thereafter used to determine characterizing
parameters of a structure (FF 5). In other words, Wormington’s Fig. 4
discloses measuring the X-ray scattering data for a tested specimen,
estimating a model for that specimen, simulating the X-ray scattering using
known methods, determining the difference between the simulated and
measured curves, determining whether an error value is less than a threshold
value, and if not, adjusting one or more parameters of the model and
repeating the steps (FF 6). As such, Wormington discloses a system that
Appeal 2008-4716
Application 10/608,300

10
receives a profile as an input and through employing a machine learning
algorithm generates desired outputs. Therefore, we find that the claimed
“the second diffraction signal as an output of the machine learning system”
reads on Wormington’s Fig. 4.
Even though, in some instances, we sustain the examiner's rejections
for different reasons than those advanced by the examiner, our position is
still based upon the collective teachings of the references and does not
constitute a new ground of rejection. In re Bush, 296 F.2d 491, 496 (CCPA
1961); In re Boyer, 363 F.2d 455, 458 n.2 (CCPA 1966).
Therefore, we do not find that Appellants have shown error in the
Examiner’s rejection of illustrative claim 1. Instead, we find the Examiner
has set forth a sufficient initial showing of obviousness. Therefore, we
affirm the rejection of independent claim 1 and of claims 2-29, which fall
therewith.

VI. CONCLUSIONS
We conclude that Appellants have not shown that the Examiner erred
in rejecting claims 1-29.
Thus, claims 1-29 are not patentable.

VII. DECISION
In view of the foregoing discussion, we affirm the Examiner’s
rejection of claims 1-29.
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a). See 37 C.F.R.
§ 1.136(a)(1)(iv) (2007).
Appeal 2008-4716
Application 10/608,300

11

AFFIRMED

msc

PETER J. YIM
MORRISON & FOERSTER LLP
425 MARKET STREET
SAN FRANCISCO CA 94105-2482