During Nodal, Session 13, we did not discuss the substance of the Office Action. We since have received a question about the Office Action. Further, the Applicant’s response is now publicly available, so our comments now are moot (as desired).
When considering the Office Action, the Applicant initially should read the various objections and rejections enough to understand the basic issues. For example, it is not important to understand fully how the Examiner constructed the eligibility rejection under 35 USC 101. It is sufficient to understand the eligibility rejection is substantial, as opposed to being overcome by a trivial amendment (e.g., adding the word “non-transitory” to a medium claim).
Having gotten a sense for these objections and rejections, the Applicant’s substantive consideration should begin with the anticipation rejection. As discussed in the video, the Examiner premised the anticipation rejection on a reference describing the extraction of a server name from an information technology (IT) service ticket. Although this technology is similar from the viewpoint of machine learning performed on IT data, the technology differs with respect how the training data is constructed.
Claim 1’s recitation of the broad term “feature information” (i.e., the IT data) is precisely what permits this rejection. Thus, one way to overcome the rejection is by better defining this feature information. We believe such an amendment is consistent with the Applicant’s original intent.
But how much clarity should the amendment provide for the “feature information?” A limited amendment reciting only a “control code” would permit another broad interpretation. Similarly, an amendment reciting only a “relative positional relationship between a string and a control code” also might be interpreted broadly.
What about an “arrangement pattern” of the control code and the relative positional relationship? This element is also broad. However, this element suggests a specific association between the control codes and the positional relationship.
Therefore, the proposed amendment included “a feature input unit that outputs feature information indicating an arrangement pattern of (i) control codes and (ii) a relative positional relationship between a string and a control code.”
From our view, the specific association is what gives this amendment merit. Again, our goal was not to draft an allowable claim. The goal was to produce a more meaningful conversation with the Examiner about eligibility and patentability, if another Office Action issues. By including these structures and this specific association in the claim, any resulting rejection would give the Applicant a better sense of how to advance prosecution in reply to the next Office Action.
To address a few minor issues, the proposed Nodal amendment also considers a potential rejection under 35 USC 112(a) as reciting a single means. MPEP 2181 V. The proposed Nodal amendment does not address the interpretation of the recited units as invoking 35 USC 112(f). Further, the proposed Nodal amendment also deletes about 35 words from the end of the original claim. These words were redundant or unclear.
After our 5am (EST) recording, the Applicant filed an amendment to incorporate Claim 2 into Claim 1. So, the Applicant expects infringement to occur by a device that both trains and uses the model. In contrast, the Nodal amendment could be infringed by a device that only trains the model.
If Modal PLLC can assist you with analyzing USPTO Office Actions, please contact us.