Detailed Action
Notice of Pre-AIA or AIA Status
1. The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
2. Claims 1-20 are pending.
Claim Rejections - 35 USC § 101
3. 35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
4. Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more and thus is directed to non-patentable subject matter. Specifically, the claims are directed toward the judicial exception of an abstract idea without reciting additional elements that amount to significantly more than the judicial exception. The rationale for this determination is in accordance with the guidelines of USPTO, applies to all statutory categories, and is explained in detail below.
When considering subject matter eligibility under 35 U.S.C. 101, (1) it must be determined whether the claim is directed to one of the four statutory categories of invention, i.e., process, machine, manufacture, or composition of matter. If the claim does fall within one of the statutory categories, (2a) it must then be determined whether the claim is directed to a judicial exception (i.e., law of nature, natural phenomenon, and abstract idea), and if so (2b), it must additionally be determined whether the claim is a patent-eligible application of the exception. If an abstract idea is present in the claim, any element or combination of elements in the claim must be sufficient to ensure that the claim amounts to significantly more than the abstract idea itself. Examples of abstract ideas include certain methods of organizing human activities; a mental processes; and mathematical concepts.
STEP 1:
Per Step 1 of the two-step analysis, the claims are determined to include an method (independent claim 1), an apparatus (independent claim 8), and a non-transitory storage medium (independent claim 15) respectively and in the therefrom dependent claims. Therefore, the claims are directed to a statutory eligibility category.
Step 2A, Prong 1:
The independent claims recite:
“receiving training data comprising training inputs associated with ground truth labels corresponding to a plurality of variables, wherein the ground truth labels include a null value for a given variable of the plurality of variables” (A person can receive data comprising sent or inputted information which may be labels representing the values of variables; the values may can include zero or null value for a particular variable);
“providing the training inputs to a machine learning model that is configured to generate predictions corresponding to the plurality of variables” (a person can mentally and with pen and paper provide received data into an algorithm or set of mathematical and statistical calculations to generate a probabilities based on the algorithm and data information);
“receiving the predictions from the machine learning model in response to the training inputs” (a person can receive the generated probabilities);
“evaluating a loss function that compares the ground truth labels to the predictions and uses a masking value to disregard loss that corresponds to the given variable” (a person can mentally perform mathematical calculations to compute a loss function that removes certain values and includes others in the statistical calculations and can compare data values and label information);
“updating one or more parameters of the machine learning model based on the evaluating of the loss function” (a person can mentally assign values in the algorithm and mathematical/statistical calculation steps based on the result of the loss function calculation and data comparison).
Claims 8 and 15 recite the same features and the same analysis applies.
If claim limitations, under their broadest reasonable interpretation, covers performance of the limitations as a mental process but for the recitation of generic computer components, then it falls under the mental process grouping of abstract ideas. Accordingly, these claims recite an abstract idea.
Regarding dependent claim 2, in addition to that mentioned for claim 1, “wherein the evaluating of the loss function comprises: replacing the null value in the ground truth labels with the masking value, replacing a prediction in the predictions that corresponds to the given variable with the masking value, and computing a loss value based on the replacing of the null value in the ground truth labels with the masking value and the replacing of the prediction in the predictions with the masking value” (a person can perform the mathematical calculation replacing certain values and probabilities for particular variables, and calculate the loss function accordingly).
Regarding dependent claims 3, in addition to that mentioned for claim 2, “the computing of the loss value comprises, after the replacing of the null value in the ground truth labels with the masking value and the replacing of the prediction in the predictions with the masking value, determining differences between the ground truth labels and the predictions and dividing a sum of the differences by a total number of ground truth labels in the ground truth labels that do not comprise the masking value” (a person can perform the mathematical calculation of the loss value by replacing certain values and probabilities, comparing label data to count a number of labels which are not a particular value, and using this number in further mathematical calculations).
Regarding dependent claims 4, in addition to that mentioned for claim 3, “wherein the masking value comprises a negative number” (a person can use negative numbers to label data and make calculations).
Regarding dependent claim 5, in addition to that mentioned for claim 1, “determining an accuracy of the machine learning model based on a number of instances in which both a prediction generated by the machine learning model and a corresponding ground truth label exceed a threshold” (a person can mentally or with pen and paper determine and calculate metrics based on comparing data values to a threshold value and counting those values which exceed it”
Regarding dependent claim 6, in addition to that mentioned for claim 5, “the determining of the accuracy of the machine learning model is based on using the masking value to disregard an accuracy determination that corresponds to a null ground truth label” (a person can mentally use different steps regarding how to include values in calculations, and this may be based on other associated data).
Regarding dependent claims 7, in addition to that mentioned for claim 1, “the receiving of the predictions from the machine learning model in response to the training inputs comprises receiving a plurality of normalized output values corresponding to the plurality of variables from an output layer of the machine learning model” (a person may receive data values that are normalized and/or may apply normalization calculations to the data, and this may be used to determine which statistical calculations to use).
Claims 8-14 show the same features as claims 1-7 respectively and are rejected for the same reasons.
Claims 15-20 show the same features as claims 1-6 respectively and are rejected for the same reasons.
All these claim features may be accomplished by applying particular calculations, groupings, inspection, and general manipulation of data. The invention is thus directed to mental process groupings of abstract ideas because they cover concepts performed in the human mind, including observation, evaluation, judgment, and opinion. See MPEP 2106.04(a)(2), subsection III.
Step 2A, Prong 2
This judicial exception is not integrated into a practical application. This part of the eligibility analysis evaluates whether the claim as a whole integrates the recited judicial exception into a practical application of the exception or whether the claim is “directed to” the judicial exception. This evaluation is performed by (1) identifying whether there are any additional elements recited in the claim beyond the judicial exception, and (2) evaluating those additional elements individually and in combination to determine whether the claim as a whole integrates the exception into a practical application. See MPEP 2106.04(d).
The processors and (computer) system in claims 8 and 12 carrying out the process steps, the memory in claim 8 comprising instructions executed by the processors to perform the steps, and non-transitory computer readable medium in claims 15-20 comprising instructions executed by the processors to perform the steps, are using a generic computer system to gather data and thus are mere instructions to apply the judicial exception using generic computer.
In addition, all uses of the recited judicial exceptions require such data gathering and output, and, as such, these limitations do not impose any meaningful limits on the claim. See MPEP 2106.05. It is used to perform an abstract idea, as discussed above in Step 2A, Prong One, such that it amounts to no more than mere instructions to apply the exception using any generic computer. See MPEP 2106.05(f). The limitations provide nothing more than mere instructions to implement an abstract idea on a generic computer. See MPEP 2106.05(f). MPEP 2106.05(f) provides the following considerations for determining whether a claim simply recites a judicial exception with the words “apply it” (or an equivalent), such as mere instructions to implement an abstract idea on a computer: (1) whether the claim recites only the idea of a solution or outcome i.e., the claim fails to recite details of how a solution to a problem is accomplished; (2) whether the claim invokes computers or other machinery merely as a tool to perform an existing process; and (3) the particularity or generality of the application of the judicial exception.
Thus, under Step 2A, the Examiner holds that the claims are directed to concepts identified as abstract ideas.
STEP 2B.
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception. This part of the eligibility analysis evaluates whether the claim as a whole amounts to significantly more than the recited exception i.e., whether any additional element, or combination of additional elements, adds an inventive concept to the claim. See MPEP 2106.05.
Regarding the processors and computer system carrying out the process steps, and the memory and non-transitory computer readable medium comprising instructions causing a processor and computer system to carry out the process steps, these insignificant extra solution activities are well understood routine and conventional activities. See Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362.
Considering the additional elements individually and in combination, and the claim as a whole, the additional elements do not provide significantly more than the abstract idea. Therefore, the claims are not patent eligible.
Claim Rejections - 35 USC § 103
5. The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
6. Claim(s) 1-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Padfield et al “Padfield” (US 11250552 B1) and Gupta et al “Gupta” (“Flexible Window Predictions on Electronic Health Records”).
(Please see the attached copy of Padfield that numbers paragraphs in the same manner as that used in this Action).
7. Regarding claim 1, Padfield shows a method for multi-head machine learning model training (para 32, 34, 38, 54 shows the machine learning model with multiple task heads) comprising receiving training data comprising training inputs associated with ground truth labels corresponding to a plurality of variables (para 22, 30, 36, 38 show receiving the training data that includes training inputs associated with ground truth labels; the labels correspond to a plurality of image “variables” such as those related to image defects), wherein the ground truth labels include a null value for a given variable of the plurality of variables (para 22, 38, 40, 45 show the ground truth labels include an unknown/null value for a given particular variable); providing the training inputs to a machine learning model that is configured to generate predictions corresponding to the plurality of variables (Figure 4B, para 45, 51, 82-84 show the machine learning model using the provided training input data to generate predictions corresponding to the image defect related variables); receiving the predictions from the machine learning model in response to the training inputs (Figure 6, para 92-93 show receiving the predictions from the machine learning model that are based on the training inputs; para 39-41 show receiving the predictions and using them to determine a loss function); evaluating a loss function that compares the ground truth labels to the predictions (para 38-41, 45, 80 show evaluating the loss function based on comparing the predictions to the ground truth labels) and uses an unknown label and indicator function to disregard loss that corresponds to the given variable (para 36, 38-39 show removing loss corresponding to a particular image defect variable using an unknown label and indicator function); and updating one or more parameters of the machine learning model based on the evaluating of the loss function (para 30, 36, 48, 53 show updating parameters of the machine learning model using the evaluation of the loss function). Padfield does not explicitly use a masking value per se to disregard loss that corresponds to the given variable. Gupta however does use a masking value per se to disregard loss that corresponds to the given variable (page 12512 column 2 para 5, page 12513 column 1 para 1 show using a masking value to disregard and minimize loss for a given variable – note Gupta replaces the unknown or missing label with the masking value). It would have been obvious to a person with ordinary skill in the art before the effective filing date of the claimed invention to use a masking value to disregard loss corresponding to the given variable, in the multi-head machine learning model training method of Padfield, because it would provide an efficient way to disregard and minimize loss for an unknown/null label.
8. Regarding claim 2, in addition to that mentioned for claim 1, although Padfield shows evaluating the loss function as explained for claim 1, Padfield does not explicitly show the evaluating of the loss function comprises: replacing the null value in the ground truth labels with the masking value; replacing a prediction in the predictions that corresponds to the given variable with the masking value; and computing a loss value based on the replacing of the null value in the ground truth labels with the masking value and the replacing of the prediction in the predictions with the masking value, but Padfield para 38-41 does show checking if the label was unknown/null and if so ignore the class’s loss contribution when evaluating the loss function so as to minimize it. Gupta however shows the evaluating of the loss function comprises: replacing the null value in the ground truth labels with the masking value (page 12512 column 2 para 3, 5 show replacing the missing/null value in the ground truth labels with the masking value); replacing a prediction in the predictions that corresponds to the given variable with the masking value (page 12512 column 1 para 2 and column 2 para 3-5 show replacing the prediction corresponding to the particular variable with the masking value); and computing a loss value based on the replacing of the null value in the ground truth labels with the masking value and the replacing of the prediction in the predictions with the masking value (page 12512 column 2 para 5, page 12513 column 1 para 1 show calculating the loss value based on the masking value which replaced the null value and the masking value which replaced the prediction). It would have been obvious to a person with ordinary skill in the art before the effective filing date of the claimed invention to have this in the multi-head machine learning model training method of Padfield, because it would provide an efficient way to disregard unknown label values and predictions corresponding to unknown labels when evaluating the loss function, in order to minimize it.
9. Regarding claim 3, in addition to that mentioned for claim 2, Padfield para 38-41 show the computing of the loss value comprises determining differences between the ground truth labels and the predictions and dividing a sum of the differences by a total number of ground truth labels in the ground truth labels. Note in para 40 how the values corresponding to the unknown/null labels are ignored, which means the ground truth labels that are counted are the ones without the unknown/null labels. Gupta shows as explained for claim 2 that the values corresponding to the unknown/null labels and corresponding predictions are replaced with the masking value, and so those ground truth labels that are counted are thus the ones that do not comprise the masking value.
10. Regarding claim 4, in addition to that mentioned for claim 3, Padfield para 21, 38 show assigning a label of -1 to the unknown/null labels, but Padfield does not explicitly say the masking value itself per se comprises a negative number. Given the combination of Padfield with Gupta as explained for claim 3, it would have been obvious to a person with ordinary skill in the art before the effective filing date of the claimed invention to use the assigning number of -1 as the masking value, because it would provide a convenient label which is out of normal counting range and which can be used to indicate skipping of the class’s loss from calculation.
11. Regarding claim 5, in addition to that mentioned for claim 1, Padfield shows determining an accuracy of the machine learning model based on a number of instances in which both a prediction generated by the machine learning model and a corresponding ground truth label exceed a threshold (Padfield para 25, 34, 83, 86 show an accuracy is determined based on amount of predictions and corresponding labels are over a threshold. Note that Gupta page 12513 column 2 para 2, 4 also show a performance metric determined based on predictions and associated labels exceeding a baseline standard. This metric is based on loss and page 12512 column 2 shows the loss derives from the difference in predictions from the ground truth. Therefore the metric is an accuracy metric).
12. Regarding claim 6, in addition to that mentioned for claim 5, the determining of the accuracy of the machine learning model is based on using the masking value to disregard an accuracy determination that corresponds to a null ground truth label (Gupta 12513 column 2 para 4 shows the performance metric [which per the explanation for claim 5 is thus an accuracy metric] is based on loss, and page 12512 column 2 para 2-4 and page 12513 column 1 para 1 shows a masking value is used to disregard the loss contributions corresponding to a null/missing ground truth label. Thus a masking value is used to disregard an accuracy determination corresponding to a null/missing ground truth label).
13. Regarding claim 7, the receiving of the predictions from the machine learning model in response to the training inputs comprises receiving a plurality of normalized output values corresponding to the plurality of variables from an output layer of the machine learning model (Padfield para 45, 54 show the received output predictions are normalized output values of the variables from the output layer of the machine learning model. Gupta page 12514 column 2 para 1 also shows normalized scores from the model to predict output).
14. Claims 8-14 show the same features as claims 1-7 and are rejected for the same reasons. In addition, Padfield para 68-69 show the computer system with processors and the memory with instructions which when executed by the processors cause the computer system to carry out the method steps.
15. Claims 15-20 show the same features as claims 1-6 and are rejected for the same reasons. In addition, Padfield para 68-69 and 100 show the computer system with processors and the non-transitory computer readable medium (hardware memory) with instructions which when executed by the processors cause the computer system to carry out the method steps.
Conclusion
16. The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
a) Durand (US 2020/0160177 A1) trains a multi-output neural network when some labels are missing, using masking values to exclude loss from unknown label portions (and calculate loss only from known labels). The partial loss is normalized according to the number of known labels.
b) Hassani (US 2020/0160177 A1) is a multi-task neural network in which some tasks have sparse or limited training datasets. When a task has no ground truth for an image, the output is set to a null value and the loss is set to zero.
17. Any inquiry concerning this communication or earlier communications from the examiner should be directed to STEVEN PAUL SAX whose telephone number is (571)272-4072. The examiner can normally be reached Monday - Friday, 9:30 - 6:00 Est.
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, Usmaan Saeed, can be reached at 571-272-4046. 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.
/STEVEN P SAX/ Primary Examiner, Art Unit 2146
Prosecution Insights