Wednesday, May 11, 2011

BPAI reverses anticipation based on multiple distinct embodiments


Takeaway: In Ex parte Cucerzan, the BPAI reversed an anticipation rejection when the Examiner relied on multiple distinct embodiments within the single reference. "Thus, it is not enough [in an anticipation rejection] that the prior art reference discloses part of the claimed invention, which an ordinary artisan might supplement to make the whole, or that it includes multiple, distinct teachings that the artisan might somehow combine to achieve the claimed invention. Net MoneyIn, Inc. v. Verisign, Inc., 545 F.3d 1359, 1369 (Fed. Cir. 2008)."

Details:

Ex parte Cucerzan
Appeal 2009008190; Appl. No. 11/094,078; Tech. Center 2100
Decided  May 2, 2011

The technology for this patent application related to spell checking. A representative claim on appeal read:
1. A method of forming a target error model to facilitate spell checking input text related to a target data collection comprising steps of:
a) providing a source query log containing user queries to at least one source data collection;
b) generating target relational data based on the source query log including corrective substring suggestions that relate to the target data collection and corresponding misspelled substrings for the
corrective substring suggestions extracted from the source query log;
c) building a target error model using the target relational data including target statistical occurrence data for the substrings of the target relational data derived from the source query log; and
d) storing the target error model on a computer readable medium.

All independent claims were rejected as being anticipated by Shazeen. On appeal, the Applicant argued that Shazeen did not disclose the claimed relationships between the elements (e.g., generating X based on Y and building Z using X). The Applicant pointed out that these relationships were missing because the Examiner relied on separate and distinct embodiments in Shazeen. For example, the Applicant argued :
[T]here is no disclosed relationship between the cited elements of the spell-checking embodiments C and E and the cited "target error model" (i.e., database of contexts of the spell-checking embodiment B [col. 2, ll. 22-44]) of Shazeer, as the Examiner contends. Thus, there is no disclosure of the spell-checking embodiments C and E forming the database or corpus of the spellchecking embodiment B (i.e., cited target error model), as the Examiner contends.

The Examiner did not address this argument in the Answer.

The Board reversed the anticipation rejection because "the Examiner reaches a finding of anticipation by combining separate, distinct embodiments of Shazeer." The Board noted this is prohibited by the case law of anticipation, which requires elements that are both identically shown in the reference and  arranged as they are in the claims, as explained by NetMoneyIn:
Thus, it is not enough that the prior art reference discloses part of the claimed invention, which an ordinary artisan might supplement to make the whole, or that it includes multiple, distinct teachings that the artisan might somehow combine to achieve the claimed invention. Net MoneyIn, Inc. v. Verisign, Inc., 545 F.3d 1359, 1369 (Fed. Cir. 2008).

The Board found that Shazeer clearly referred to multiple embodiments:
In particular, the respective portions of Shazeer’s columns 2, 3, and 10 that are relied on by the Examiner refer to different disclosed embodiments, as clearly indicated by the first sentence of each paragraph in Shazeer’s columns 2 and 3, and the first sentence of each of the first two paragraphs of column 10: e.g., “According to another general example embodiment of the present invention . . . .” (col. 2, ll. 22-23), “In another example embodiment of the present invention . . . .” (col. 3, l. 1), “According to a further example embodiment of the present invention . . . .” (col. 10, ll. 1-2).

The Board concluded:
Because of the Examiner’s reliance on multiple distinct embodiments in Shazeer, we are in accord with the Appellant that the elements of the claimed invention are not identically shown in the reference, arranged as they are in the claims.

Two dependent claims were rejected as being obvious over the anticipatory reference. The Board affirmed this obviousness rejection, noting that anticipation arguments relating to multiple embodiments are not relevant for obviousness. In a footnote, the Board declined to reject the independent claims under § 103, instead stating that "we leave this question of obviousness to the Examiner in the event of further prosecution."

My two cents: Patents are full of the word "embodiment," so the next time you see an anticipation rejection, it may be worth investigating whether the Examiner is relying on what the reference describes as different embodiments. Now, if you convince the Examiner that anticipation is inappropriate under Net MoneyIn, you'll probably get an obviousness rejection. So you should also consider whether you have good arguments against combining the multiple embodiments of the single reference.

Though the Applicant pointed out that the Examiner relied on multiple embodiments, the Applicant didn't directly make the argument that anticipation doesn't allow this. But the Board got the point and filled in the blanks a bit. I would definitely cite Net MoneyIn or In re Arkley when making this argument.

9 comments:

  1. Applicant didn't cite New MoneyIn since it was not decided at the time the Appeal Brief was filed -- pretty sad considering that New MoneyIn was decided in 2008.

    Also, this concept is (or should be) pretty well known. Search "picking" and "choosing" and you'll get some hits for other Federal Circuit cases on point.

    ReplyDelete
  2. >was not decided at the time the Appeal Brief
    >was filed

    Ah, you're right. The Applicant cited State Street rather than In Re Bilski for the same reason. Sad that it took so long to get to appeal.

    >Search "picking" and "choosing" and you'll
    >get some hits for other Federal Circuit
    > cases on point.

    Right. I usually use In re Arkley.

    ReplyDelete
  3. I am amazed at the number of practitioners that don't hammer the examiners when the "anticipation" rejections are clearly based on mixing and matching embodiments of the reference. The standard of anticipation is identity of invention, not "the reference has everything you're claiming in it."

    ReplyDelete
  4. The practitioners probably realize that there is a plausible 103 rejection and basically argue against that in response to the erroneous 102 rejection. It likely isn't worth the time or money to fight and win against the 102 rejection only to turn around and have to deal with basically the same rejection under 103. In that case its better to just argue or amend against the art and advance the case even if it is technically an improper rejection under the wrong section.

    ReplyDelete
  5. Make the argument against the 102, let the examiner change it to a 103 if the examiner wants to, but if there's no amendment the new 103 rejection has to be non-final, and then argue it. Arguing against rejections that are not of record is a bad idea.

    ReplyDelete
  6. I'd take a pragmatic approach, somewhere in between "always argue" and "don't bother."

    First, I'd evaluate the rationale for combining the various embodiments in the single reference. If that looks weak, I'd probably argue without amending. With a fallback position of adding a dependent claim that distingishes.

    If the rationale for combining looks strong, then I'd consider amending around. But first I'd see whether or not I need to make amendments for other reasons -- 101, 112, etc. -- such that the Examiner is gonna take me final. If so, probably amend. If not, maybe I'd argue rather than amend -- then the sure-to-come 103 rejection should be non-final.

    ReplyDelete
  7. "The Examiner did not address this argument in the Answer."

    So did the Board apply the old "silence equals consent" case law against the examiner? I highly doubt it.

    ReplyDelete
  8. "If the rationale for combining looks strong, then I'd consider amending around. But first I'd see whether or not I need to make amendments for other reasons -- 101, 112, etc. -- such that the Examiner is gonna take me final. If so, probably amend. If not, maybe I'd argue rather than amend -- then the sure-to-come 103 rejection should be non-final."

    Bad practice. Anytime you amend, you open yourself up to a final rejection next time around, which means your opportunity to amend (without filing a RCE) is very limited.

    If you can overcome the prior art without amendment, this is what you should always attempt to do. Besides the prosecution history estoppel issue, you allow yourself more flexibility to subsequently amend claims prior to the action going final.

    Even if there are 112/101 issues that could be addressed by amendment, I would rather wait. Those can always be corrected at a later time. Besdies most 112/102 amendments are to placate examiners, not because the claims NEED the amendments.

    ReplyDelete
  9. >Even if there are 112/101 issues that could be
    >addressed by amendment, I would rather wait.
    >Those can always be corrected at a later time.

    Well, gee, any amendment to overcome any statutory requirement can be made at a later time :-)

    I wasn't thinking of fixes to antecedent basis to placate the Examiner. I was thinking of non-prior-art rejections that had merit. I do sometimes delay amending to fix 101 problems until I have other amendments to make, as long as I have a colorable argument.

    But you've convinced me to revise my position to be:
    "If the rationale for combining looks strong, then I'd consider amending around. But first I'd see whether or not I *need* to make amendments for *112 1st or indefiniteness * -- such that the Examiner is gonna take me final. If so, probably amend. If not, maybe I'd argue rather than amend -- then the sure-to-come 103 rejection should be non-final."

    That is, I'd probably still amend to deal with a *valid* 112 1st or indefiniteness rejection. And if I was gonna amend for that reason, I'd probably go ahead amend *at least one claim* to head off a probable 103.

    ReplyDelete