Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
A preliminary examination of this application reveals that it includes terminology which is so different from that which is generally accepted in the art to which this invention pertains that a proper search of the prior art cannot be made. For example: see the rejections under 35 U.S.C. 112(b) below for a description of the indefinite and incomprehensible terms.
Applicant is required to provide a clarification of these matters or correlation with art-accepted terminology so that a proper comparison with the prior art can be made. Applicant should be careful not to introduce any new matter into the disclosure (i.e., matter which is not supported by the disclosure as originally filed).
A shortened statutory period for reply to this action is set to expire TWO (2) MONTHS from the mailing date of this letter.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-10 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Regarding Claim 1, it recites “divide an instance input space of each of a plurality of machine learning models into a plurality of regions, and assign a probability to each of the division regions.” It is unclear what is being divided, what “an instance space” is, and how the instance space relates to the machine learning models, and what it means for a division region to have a probability. Is an instance space a dataset (such as a training dataset for a machine learning model), a portion of a dataset, or something else? What does it mean for an instance space to be “of each of a plurality of machine learning models”? It is unclear how a space is part of a model, if the multiple models share a space . . . nothing about this limitation makes sense. As a consequence, it’s also completely unclear what the probability means, does, or represents because the instance space and the dividing is incomprehensible. The specification does nothing to explain these unusual terms or the operations being performed.
Further regarding claim 1, it recites “calculate a sampling probability on a predetermined instance belonging to the division region based on the probability assigned to the division region.” It is unclear what “the division region” is because there are multiple division regions (and because the division regions are not clearly defined or explained). It is unclear how “a sampling probability” differs from the probabilities of each of the division regions. It is unclear what a sampling probability of an instance means at all, and the claim does not recite how it is calculated or what the probability represents. And it cannot be understood how the predetermined instance relates to the division regions—how does it “belong” to a division region, and how do you determine which division region it belongs to? None of the terms or relationships in this limitation make sense, so they have no definable meaning.
Further regarding claim 1, it recites “select the predetermined instance based on the sampling probability on the predetermined instance.” What does it mean to select the predetermined instance? It has been predetermined, so why does it need to selected? What is it selected for? As described above, the sampling probability is incomprehensible, so it’s impossible to understand how the predetermined instance is selected “based on” the sampling probability.
As an additional note, an embodiment described in the specification states that the machine learning models are decision trees (such as those recited by claim 8). Most decision trees are described in the art as including weights that are learned in a training process, and decision trees are often constructed by generating splits. The applicant does not use these known terms at all. It is unclear how the input space, division regions, and probabilities of the present claim relate to the known properties and elements of decision trees or other machine learning models.
Regarding Claim 2, it recites “wherein the processor is further configured to execute the instructions to calculate the sampling probability on the predetermined instance based on the probability assigned to the division region set for each of the machine learning models different from each other.” The ending phrase “set for each of the machine learning models different from each other” is logically and substantially incomprehensible. There is no indication in claim 1 that a division region is set for a machine learning model, so it cannot be determined what this means. And it is unclear what entities are different from each other—is it the division regions, the machine learning models, or something else? Claim 2 just does not make any sense.
Regarding Claim 3, it recites “calculate the sampling probability on the predetermined instance based on the probabilities assigned to the division regions to which the identical predetermined instance belongs, the division regions being set for the respective machine learning models different from each other.” This too makes no sense. The probabilities assigned to the division regions are still confusing and undefined. Nothing in the claims defines which division region the predetermined instance belong to. The term “the identical predetermined instance” is undefined and lacks antecedent basis. And it is unclear what elements are “identical”; the present claim only recites a single predetermined instance, so it’s not even clear what it is being compared to in order to determine if it’s identical to something else. It is unclear what it means for a division region to be “set,” much less what it is being set to. What does it mean for a division region to be “set to” a machine learning model? And as for the phrase “the respective machine learning models different from each other,” nothing there is clear either. The phrase doesn’t make sense grammatically, and it is unclear what is different from each other—the machine learning models, the act of setting, the division region, or something else.
Regarding Claim 4, it recites “wherein the processor is further configured to execute the instructions to assign the probabilities to the division regions of the plurality of machine learning models based on a result of prediction on an input instance by another machine learning model that is different from the plurality of machine learning models and results of prediction on the input instance by the respective machine learning models.” It is still completely unclear what the division regions are and what the probabilities mean. Consequently, basing the probabilities on prediction of machine learning models is also unclear and indefinite.
Regarding Claim 5, it recites “assign the probabilities to the division regions of the plurality of machine learning models based on differences between a prediction probability in the division region set for the other machine learning model and prediction probabilities in the division regions set for the respective machine learning models.” As with claim 4, it remains unclear what the probabilities or the division regions are, so a limitation that recites how the probabilities are assigned is indefinite. It is further unclear what “a prediction probability in the division region set for the other machine learning model” means. What is a prediction probability, and how does it differ from the “prediction” in claim 4? What does it mean for a prediction probability to be “in the division region”? There is also no indication of what it means for a division region to be “set for” a machine learning model. Nothing in the recited process makes any sense because the terms are not clearly defined and the process is impossibly confusing.
Regarding Claim 6, it recites “set so that values of the probabilities assigned to the division regions of the plurality of machine learning models become larger as the differences become larger.” The claim does not explain what is being set. The probabilities and the division regions are still not clearly defined. It is unclear what “differences” are being referenced. And there is no indication that any differences become larger—there is no element of time described, nor any recitation of a change in any values or a difference between values. Consequently, “the differences becom[ing] larger” is undefined, and therefore the probabilities are even more undefined.
Regarding Claim 7, it recites “assign the probabilities to the division regions of the plurality of machine learning models based on a result of prediction on the input distance by the other machine learning model that is a new machine learning model generated based on the plurality of machine learning models and results of prediction on the input instance by the respective machine learning models.” As before, the probabilities and the division regions are still indefinite. The term “input distance” lacks antecedent basis. (Perhaps it was intended to recite “input instance”.) It is unclear what is being “generated”—is it the other machine learning model, the prediction on the input distance/instance, or something else? And how is this undefined entity generated “based on” both the respective machine learning models and the results of prediction by the machine learning models? Machine learning models and predictions are completely different entities that are defined and measured in completely different ways, so it is unclear how generating something is based on both models and predictions.
Regarding Claim 8, it recites “wherein the plurality of machine learning models are decision trees or decision lists.” This is just about the only limitation in the claims that is understandable. However, since claim 8 depends on claim 1, it includes all of the indefinite and incomprehensible limitations of claim 1.
Regarding Claims 9 and 10, they recite limitations substantially similar to those of claim 1, so they are indefinite and incomprehensible for the same reasons.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant’s disclosure. Karabadji, Nour El Islem, et al. (“A data sampling and attribute selection strategy for improving decision tree construction,” Expert Systems with Applications 129 (2019): 84-96) teaches data sampling in the construction of decision tree machine learning models. For all the reasons detailed above, it cannot be determined how similar this art is to the claimed invention or how its teachings could be mapped to the claim limitations.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to HAL W SCHNEE whose telephone number is (571) 270-1918. The examiner can normally be reached M-F 7:30 a.m. - 6:00 p.m.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Michael Huntley can be reached at 303-297-4307. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/HAL SCHNEE/Primary Examiner, Art Unit 2129
Prosecution Insights