Ex parte Sutton

Board of Patent Appeals and Interferences

Dec 11, 1998

07979139 (B.P.A.I. Dec. 11, 1998)

Application for patent filed November 20, 1992. 1
According to appellant, the application is a continuation in
part of Application 07/947,213, filed September 18, 1992.
1
THIS OPINION WAS NOT WRITTEN FOR PUBLICATION
The opinion in support of the decision being entered today (1)
was not written for publication in a law journal and (2) is
not binding precedent of the Board.
Paper No. 14
UNITED STATES PATENT AND TRADEMARK OFFICE
__________
BEFORE THE BOARD OF PATENT APPEALS
AND INTERFERENCES
__________
Ex parte RICHARD S. SUTTON
__________
Appeal No. 94-4200
Application 07/979,1391
___________
ON BRIEF
___________
Before HAIRSTON, JERRY SMITH, and FLEMING, Administrative
Patent Judges.
FLEMING, Administrative Patent Judge.
DECISION ON APPEAL
This is a decision on appeal from the final rejection of
claim 1.
The invention relates to a machine for learning a pattern
Appeal No. 94-4200
Application 07/979,139
2
sequence utilizing an incrementally adjustable gain parameter.
The independent claim 1 is reproduced as follows:
1. A computer system for machine learning of a time
dependent pattern sequence y(t) comprising:
input means for receiving a plurality, indexed by i, of time
dependent inputs x (t) and a meta-step-size parameter 2;i
calculation means for calculating from said time dependent
inputs a predicted value, y , of said pattern sequence;*
a computer memory associated with the said means for
calculating;
said calculating means further including a learning rate, k ,i
exponentially related to an incremental gain $ (t) and ai
derivation means for deriving the incremental gain $ (t) fromi
previous values of B (t) and having means for 1
Initializing h , a per input memory parameter, to O, andi
weight coefficients, w , and $ , the incremental gaini i
parameter, to chosen values, i=1,...,n,
Repeating for each new inputs (x ,...,x , y )the steps of:1 n *
calculating,
n
y = E w xi i
i = 1
calculating,
* = y - y*
Repeating for i = 1,...,n where k is an input learning i
rate and 2 is a positive constant denoted the
meta-learning rate:
Appeal No. 94-4200
Application 07/979,139
The Examiner responded to the brief with an Examiner's2
answer, mailed June 14, 1994. The Examiner mailed a
supplemental Examiner's answer on May 20, 1996. The Examiner
mailed a second supplemental Examiner's answer on July 30,
1996.
3
calculating,
$ = $ +2*x hi i i i
n
k (t) = x (t)e /(R + E x (t) j )i i j$i(t) 2 $ (t)
j=1
w (t+1) = w (t)+k (t)*(t)x (t)i i i i
h (t+1) = [h (t) + k (t)*(t)][l - k (t)x (t)] .i i i i i t
The Examiner does not rely on any references.
Claim 1 stands rejected under 35 U.S.C. 101 as being non-
statutory subject matter. Claim 1 is also rejected under 35
U.S.C. § 112, second paragraph.
Rather than reiterate the arguments of Appellant and the
Examiner, reference is made to the brief and answers for the2
respective details thereof.
OPINION
After a careful consideration of the record before us, we
will not sustain the 35 U.S.C. § 101 rejection of claim 1.
With respect to the mathematical algorithm exception, the
Federal Circuit in State Street Bank v. Signature Financial,
Appeal No. 94-4200
Application 07/979,139
4
149 F.3d 1368, 47 USPQ2d 1596 (Fed. Cir. 1998), first
identified the three categories that are not patentable--laws
of nature, natural phenomena and abstract ideas. The opinion
went on to note "the mathematical algorithm is unpatentable
only to the extent that it represents an abstract idea" and is
thus not "useful." Id. at 1373 n.4, 47 USPQ2d at 1600-01 n.4.
Later in its opinion, the
court returned to this issue: "[T]he mere fact that a claimed
invention involves inputting numbers, calculating numbers,
outputting numbers, and storing numbers, in and of itself,
would not render it nonstatutory subject matter, unless, of
course, its operation does not produce a ‘useful, concrete and
tangible result.’" Id. at 1374, 47 USPQ2d at 1602. In this
case, the court stated that "the transformation of data,
representing discrete dollar amounts, by a machine through a
series of mathematical calculations into a final share price,
constitutes a practical application of a mathematical
algorithm . . . because it produces ’a useful, concrete and
Appeal No. 94-4200
Application 07/979,139
5
tangible result’ . . . . " Id. at 1373, 47 USPQ2d at
1601.
Significantly, the court concluded its analysis of the
mathematical algorithm issue as follows: "The question of
whether a claim encompasses statutory subject matter should
not focus on which of the four categories of subject matter a
claim is directed to . . .but rather on the essential
characteristics of the subject matter, in particular, its
practical utility." Id. at 1375, 47 USPQ2d at 1602.
With respect to the Freeman-Walter-Abele test, the
Federal Circuit held the district court erred in applying it.
According
to the court, after Diehr [602 F.2d 982, 203 USPQ 44 (CCPA
1979)] and Chakrabarty [571 F.2d 40, 197 USPQ 72 (CCPA 1978)]
were decided by the Supreme Court, the test had "little, if
any, applicability to determining the presence of statutory
subject matter." Id. at 1374, 47 USPQ2d at 1601.
Appellant's claim 1 recites a "machine" claim having
Appeal No. 94-4200
Application 07/979,139
6
"means" clauses. Machine claims having means clauses may only
be reasonably viewed as process claims if there is no
supporting structure in the written description that
corresponds to the claimed "means" element. See State Street
Bank,149 F.3d at 1371, 47 USPQ2d at 1599 citing In re Alappat,
33 F.3d 1526, 1540-41, 31 USPQ2d 1545, 1554 (Fed. Cir.
1994)(in banc). This is not the case now before us.
When claim 1 is properly construed in accordance with 35
U.S.C. § 112, sixth paragraph, it is directed to a specific
machine. As demonstrated below, the relevant part of claim 1
is set forth with the brackets stating the structure the
written description discloses as corresponding to the
respective "means" recited in the claim.
1. A computer system for machine learning of a time
dependent pattern sequence y(t) comprising:
input means [input ports 115 shown in Figure 1 and
described on page 5 of the specification] for receiving a
plurality, indexed by i, of time dependent inputs x (t) and ai
meta-step-size parameter 2;
Appeal No. 94-4200
Application 07/979,139
7
calculation means [processor 105 shown in Figure 1 and
described on page 5 of the specification] for calculating from
said time dependent inputs a predicted value, y*, of said
pattern sequence;
a computer memory [memory 110 shown in Figure 1 and
described on page 5 of the specification] associated with said
means for calculating . . . .
Thus, when properly construed, claim 1 claims a machine for
receiving a plurality of time dependent inputs and for
learning of a time dependent pattern sequence based upon these
inputs having the specific structures disclosed in the written
description and corresponding to the means-plus-function
elements recited in the claim.
We agree with the Examiner that the claim recites a
mathematical algorithm. The mere fact that a claimed
invention involves inputting numbers, calculating numbers,
outputting numbers, and storing numbers, in and of itself,
would not render it nonstatutory subject matter, unless, of
course, its operation does not produce a ‘useful, concrete and
tangible result.' In this case, we find that the claim
language recites a specific machine for learning of a time
dependent pattern sequence
Appeal No. 94-4200
Application 07/979,139
8
comprising an input port for receiving a plurality of time
dependent inputs, a processor for calculating from the time
dependent input a predicted value and a computer memory
associated with the processor. Furthermore, we find that
Appellant's claim 1 recites subject matter that constitutes a
practical application of a mathematical algorithm because it
produces a useful, concrete and tangible result by using a
specific machine to learn a time dependent pattern sequence
from receiving a plurality of time dependent inputs.
Therefore, we find statutory subject matter.
Claim 1 stands rejected under 35 U.S.C. § 112, second
paragraph. The Examiner argues that h is not clearly definedi
and the language "a per input memory parameter" is unclear.
Analysis of 35 U.S.C. § 112, second paragraph, should
begin with the determination of whether claims set out and
circumscribe the particular area with a reasonable degree of
precision and particularity; it is here where definiteness of
the language must be analyzed, not in a vacuum, but always in
light of teachings of the disclosure as it would be
interpreted by one possessing ordinary skill in the art. In
Appeal No. 94-4200
Application 07/979,139
9
re Johnson, 558 F.2d 1008, 1015, 194 USPQ 187, 193 (CCPA
1977), citing In re Moore, 439 F.2d 1232,
1235, 169 USPQ 236, 238 (1971). Furthermore, our reviewing
court points out that a claim which is of such breadth that it
reads on subject matter disclosed in the prior art is rejected
under 35 U.S.C. § 102 rather than under 35 U.S.C. § 112,
second paragraph. See In re Hyatt, 708 F.2d 712, 715, 218
USPQ 195, 197 (Fed. Cir. 1983) and In re Borkowski, 422 F.2d
904, 909, 164 USPQ 642, 645-46 (CCPA 1970).
In light of the specification, h is a parameter which isi
influenced by previous sample inputs. Thus, the language in
the claim, "h , a per input memory parameter," is simply seti
forth this parameter that is influenced by previous sample
inputs. Therefore, we find the claim language sets out and
circumscribes the particular area with a reasonable degree of
precision and particularity when read in light of teachings of
the disclosure as it would be interpreted by one possessing
ordinary skill in the art.
Appeal No. 94-4200
Application 07/979,139
10
We have not sustained the rejection of claim 1 under 35
U.S.C. § 101 or 112. Accordingly, the Examiner's decision is
reversed.
REVERSED
)
KENNETH W. HAIRSTON )
Administrative Patent Judge )
)
)
) BOARD OF PATENT
JERRY SMITH )
Administrative Patent Judge ) APPEALS AND
)
) INTERFERENCES
)
MICHAEL R. FLEMING )
Administrative Patent Judge )
Appeal No. 94-4200
Application 07/979,139
11
MRF/pgg
Leonard Charles Suchyta
GTE Service Corporation
40 Sylvan Road - MS31
Waltham, MA 02254