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 .
Amendments
This action is in response to amendments filed December 2nd, 2025, in which Claims 1, 2, 10, 11, 12, 13, and 19 have been amended. Claims 8, 9, 17, and 18 are cancelled. The amendments have been entered, and Claims 1-3, 5-7, 10-16, 19, and 20 are currently pending.
Claim Objections
Claim 1 is objected to because of the following informalities: The claim recites the phrase classification labels for the one mor more test data sets … This appears to be a typographical error, and the claim will be interpreted as if it had read classification labels for the one or more test data sets …
Claim 13 is objected to because of the following informalities: The claim recites the phrase classification labels for the one mor more test data sets … This appears to be a typographical error, and the claim will be interpreted as if it had read classification labels for the one or more test data sets …
Appropriate correction is required.
Claim Interpretation
The following is a quotation of 35 U.S.C. 112(f):
(f) Element in Claim for a Combination. – An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The following is a quotation of pre-AIA 35 U.S.C. 112, sixth paragraph:
An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The claims in this application are given their broadest reasonable interpretation using the plain meaning of the claim language in light of the specification as it would be understood by one of ordinary skill in the art. The broadest reasonable interpretation of a claim element (also commonly referred to as a claim limitation) is limited by the description in the specification when 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is invoked.
As explained in MPEP § 2181, subsection I, claim limitations that meet the following three-prong test will be interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph:
(A) the claim limitation uses the term “means” or “step” or a term used as a substitute for “means” that is a generic placeholder (also called a nonce term or a non-structural term having no specific structural meaning) for performing the claimed function;
(B) the term “means” or “step” or the generic placeholder is modified by functional language, typically, but not always linked by the transition word “for” (e.g., “means for”) or another linking word or phrase, such as “configured to” or “so that”; and
(C) the term “means” or “step” or the generic placeholder is not modified by sufficient structure, material, or acts for performing the claimed function.
Use of the word “means” (or “step”) in a claim with functional language creates a rebuttable presumption that the claim limitation is to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites sufficient structure, material, or acts to entirely perform the recited function.
Absence of the word “means” (or “step”) in a claim creates a rebuttable presumption that the claim limitation is not to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is not interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites function without reciting sufficient structure, material or acts to entirely perform the recited function.
Claim limitations in this application that use the word “means” (or “step”) are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action. Conversely, claim limitations in this application that do not use the word “means” (or “step”) are not being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action.
This application includes one or more claim limitations that do not use the word “means,” but are nonetheless being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, because the claim limitation(s) uses a generic placeholder that is coupled with functional language without reciting sufficient structure to perform the recited function and the generic placeholder is not preceded by a structural modifier. Such claim limitation(s) is/are: an input module configured to receive, a model generator configured to generate, a pre-processer configured to process, an extractor configured to parse, a training module configured to train, and a validator configured to validate in Claim 13, as well as an application module configure to apply in Claim 20.
Because this/these claim limitation(s) is/are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, it/they is/are being interpreted to cover the corresponding structure described in the specification as performing the claimed function, and equivalents thereof.
If applicant does not intend to have this/these limitation(s) interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, applicant may: (1) amend the claim limitation(s) to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph (e.g., by reciting sufficient structure to perform the claimed function); or (2) present a sufficient showing that the claim limitation(s) recite(s) sufficient structure to perform the claimed function so as to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph.
Claim Rejections - 35 USC § 101
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.
Claims 1-3, 5-7, 10-16, 19, and 20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Regarding Claim 1:
Subject Matter Eligibility Analysis Step 1:
Claim 1 recites a method, thus a process, one of the four statutory categories of patentable subject matter.
Subject Matter Eligibility Analysis Step2A Prong 1:
Claim 1 recites steps of dividing the plurality of prospectuses into one or more training sets and one or more testing sets using a k-fold cross validation method (a mental process), generating one or more modified training sets based on the one or more training sets which consists of the following mental process steps: extracting one or more data of interest from each prospectus in the one or more training sets, wherein the one or more data of interest includes at least one of an entity name, an objective, and a principle investment strategy (a mental process); combining the one or more data of interest into a single text string; (a mental process), extracting, using a natural language processing technique, a plurality of text features from each test string and generating weights corresponding to the plurality of text features of each test string, wherein the plurality of text features are correlated to respective ones of a plurality of classification labels; (a mental process) to determine mappings between the classification labels and the text features; (a mental process), and validating the trained machine learning model by creating a confusion matrix to determine where a mismatch occurs between the trained machine learning model and the one or more testing data sets, the confusion matrix reveals a degree of match between predicted classification labels using the trained machine learning model and actual classification labels for the one or more test data sets (a mental process of comparing predicted labels and true labels).. Claim 1 therefore recites an abstract idea of extracting features from a generated set of documents, learning mappings between the features and classification labels, and scoring the learned mappings.
Subject Matter Eligibility Analysis Step2A Prong 2:
Claim 1 further recites additional elements of
receiving a plurality of prospectuses
training the machine learning model
a computing device (to perform the abstract idea steps identified above)
storing one or more training items corresponding to the one or more modified training sets, in a database, wherein the each training items includes at least one of the text string, the plurality of features, and the weights that correspond to the plurality of features
These additional elements do not integrate the abstract idea into a practical application because (a) recites an insignificant extra-solution activity of data gathering (MPEP 2106.05(g)); (b,c) merely recite using a computer or other machinery, i.e. a machine learning model, as a tool to perform the mental processes of mapping and predicting, which by MPEP 2106.05(f) cannot provide a practical application; (d) merely recites a generic computer (MPEP 2106.05(f)), and (e) recites insignificant extra-solution activity of storing information. The claim is thus directed to the abstract idea of extracting features from a generated set of documents and learning mappings between the features and classification labels.
Subject Matter Eligibility Analysis Step 2B:
The additional elements of Claim 1 do not provide significantly more than the abstract idea because (a) receiving a plurality of prospectuses is well understood, routine, and conventional activity of “transmitting or receiving data over a network” (MPEP 2106.05(d); and (b-d) merely recite using a computer or other machinery as a tool, (i.e. "apply it") and cannot provide significantly more (MPEP 2106.05(f)); and (e) storing data in memory is well-understood, routine, and conventional activity (MPEP 2106.05(d)). Therefore, Claim 1 is subject-matter ineligible.
Claim 2, dependent upon Claim 1, recites additional elements to receive input data from a third party vendor (insignificant extra-solution activity of data gathering, by MPEP 2106.05(g), which is further well-understood, routine, and conventional by MPEP 2106.05(d), receiving data over a network) and the fact that the data requires classification, which merely specifies the particular field of use, i.e. the data used in the mental processes of the abstract idea (see MPEP 2106.05(h)), neither of which integrates the abstract idea into a practical application nor provides significantly more than the abstract idea itself (taken alone or in combination)
Claims 3 and 5-7, dependent on Claim 1, further recite details of the data used in the abstract idea, i.e., specifies a particular field of use, which, by MPEP 2106.05(h), cannot integrate the abstract idea into a practical application nor provide significantly more than the abstract idea itself.
Claim 10, dependent on Claim 2, further recites additional mental process steps to predicts a classification label for the input data and generate a confidence level for the prediction as well as assigning the predicted classification label. Claim 10, however, does not recite any new additional elements which could integrate the abstract idea into a practical application nor provide significantly more than the abstract idea itself because the only additional element is applying the trained machine learning model to make the prediction, which is merely using a computer or other machinery as a tool to perform the mental process, which by MPEP 2105.05(f)(2) cannot do so.
Claim 11, dependent on Claim 10, further recites details of the data used in the abstract idea, i.e., specifies a particular field of use, which, by MPEP 2106.05(h), cannot integrate the abstract idea into a practical application nor provide significantly more than the abstract idea itself.
Claim 12, dependent on Claim 11, further recites an additional mental process step of determining whether the predicted classification label agrees with the new classification label. Claim 12, however, does not recite any new additional elements which could integrate the abstract idea into a practical application nor provide significantly more than the abstract idea itself, because the only additional element is to stop processing input data in response to a condition occurring, which only is a description of the environment in which the abstract idea takes place (i.e., not during that condition), i.e. falls under 2106.05(h).
Claims 13-16 and 20 recite a computer-implemented system comprising components configured to perform precisely the methods of Claims 1, 3, 5, 6, and 12, respectively. As performance of an abstract idea on generic computer components cannot integrate the abstract idea into a practical application or provide significantly more than the abstract idea itself (MPEP 2106.05(f)), Claims 13-17 are rejected for reasons set forth in the rejections of Claims 1, 3, 5, 6, 8, and 10-12, respectively.
Claim 19, dependent upon Claim 13, recites additional elements to receive input data from a third party vendor (insignificant extra-solution activity of data gathering, by MPEP 2106.05(g), which is further well-understood, routine, and conventional by MPEP 2106.05(d), receiving data over a network) and the fact that the data requires classification and includes financial information and a change to a new classification label, which merely specifies the particular field of use, i.e. the data used in the mental processes of the abstract idea (see MPEP 2106.05(h)), neither of which integrates the abstract idea into a practical application nor provides significantly more than the abstract idea itself (taken alone or in combination).
Claim Rejections - 35 USC § 103
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-3, 5-7, and 13-16 are rejected under 35 U.S.C. 103 as being unpatentable over Tao et al., “Analyzing forward-looking statements in initial public offering prospectuses: a text analytics approach,” in view of Tamilchelvan, US PG Pub 2018/0293231.
Regarding Claim 1, Tao teaches a computer-implemented for training a machine learning model (Tao, Abstract, “evaluating the relationship between features extracted from FLSs and IPO valuation … we propose an analytical pipeline for identifying FLSs using machine learning techniques”), the method comprising: … by a computing device (Tao, pg. 64, 2nd column, 2nd-to-last paragraph, “we use a computing system with dual Intel Xeon CPUs”) … receiving a plurality of prospectuses from a plurality of third-party entities (Tao, pg. 58, 1st column, last paragraph, “The data in this study include the final IPO prospectuses (Form 424B4) in successful IPOs of common stocks”); dividing the plurality of prospectuses into one or more training sets and one or more testing sets using a k-fold cross validation method (Tao, pg. 64, 1st column, last paragraph, “we employ stratified 10-fold cross validation method to avoid overfitting in the training process”); generating … one or more modified training sets based on the one or more training sets (Tao, pg. 64, 1st column, last paragraph, “9 out of the 10 folds are used to train the predictive models, while the remaining fold is used to evaluate” where the steps of turning the prospectuses into features for the models are generating modified training sets) generating the one or more [modified] training sets comprising, generating the trained machine learning model comprising:
… extracting … using a natural language processing technique, a plurality of text features from each [prospectus] and generating weights corresponding to the plurality of text features of each [prospectus], wherein the plurality of text features are correlated to respective ones of a plurality of classification labels (Tao, pg. 61, 2nd column, 3rd paragraph, “we use classic discrete BoWs along with the TF-IDF approach (bi-gram representations), which is consistent with the extant text classification studies” in order to determine features with which to “train our FLS classifier” i.e. use the TF-IDF weights along with the bi-gram text features to predict the classification labels), storing … one or more training items corresponding to the one or more modified training sets, in a database, where each training item includes at least one of the text string, the plurality of text features, and the weights corresponding to the plurality of text features in the database (Tao, pg. 64, 2nd column, last paragraph, “all traditional learning models are implemented in Python using Scikit-Learn” denoting that all the required data, i.e. text, features, and weights are stored in memory in a computer, i.e. a database), training … the machine learning model using the [one or more] modified trained set[s] and the plurality of classification labels, that determine mappings between the classification labels and text features, to generate a trained machine learning model (Tao, pg. 63, 2nd column, 1st paragraph, “To assess the value of analyzing FLS from MD&A sections in IPO prospectuses, we conduct a predictive analysis of IPO valuation using machine learning models …
Y
1
is defined as a binary target variable … In the second analysis,
Y
2
is defined as a binary target variable ” with 4th paragraph, “In the fourth model configuration, we use the classic text analytics approach based on the BoW and TF-IDF word representation” & last paragraph, “a variety of machine learning techniques” are used to train the model to predict the correct classification labels
Y
1
and
Y
1
); validating … the trained machine learning model using test data corresponding to the one or mor testing sets (Tao, pg. 64, 1st column, last paragraph, “10-fold cross-validation … the remaining fold is used to evaluate the modeling results”) by creating a confusion matrix to determine where a mismatch occurs between the trained machine learning model and the one or more testing data sets tests, the confusion matrix reveals a degree of match between predicted classification labels using the trained machine learning model and actual classification labels for the one or more test data sets (Tao, pg. 64, 1st column, 2nd-3rd paragraphs, “In order to evaluate the performance, we select different performance measures” including computing each of “TP” “FP” “TN” “FN” which is creating a 2x2 confusion matrix of values counting the degree of match between predicted and actual labels, i.e. how many “true positive” and “true negatives” matches vs “false positive”/”false negative” mismatches, exactly corresponding to Fig. 4 of the disclosed invention).
Tao does not clearly teach the combination of extracting … one or more data of interest from each prospectus in the one or more training sets and combining the one or more data of interest, that is extracted from each prospectus, into a single text string, but Tamilchelvan, also in the art of natural language processing, teaches extracting one or more data of interest from each [document of interest] and combining the one or more data of interest, that is extracted from each [document], into a single text string (Tamilchelvan, [0007], “the LVLI resource may extract the contents from each source document and save the resulting collection of contents as a single text string (STS) document comprising a plurality of textual components”). It would have been obvious to one or ordinary skill in the art before the effective filing date of the claimed invention to extract the contents of each prospectus document of Tao into a single text string, as does Tamilchelvan. The motivation to do so is that then allows the system to then “parse or split” the document, via known means, into data to be analyzed (Tamilchelvan, [0007-0009]). Further, extracted the text of the prospectuses of Tao, in generating the single string as does Tamilchelvan, thus teaches extracting one or more data of interest from each prospectus in the one or more training sets, wherein the one or more data of interest includes at least one of an entity name, an objective, and a principle investment strategy, because the data of interest/prospectus text of Tao includes each of these: an entity name (Tao, pg. 30, 1st column, 2nd paragraph, “Subject-Verb-Object” & “a reference of a future time (year)” where subjects and objects and years are names of entities , also see pg. 63, “Top 5 Topics in FLSs” are words extracted from the prospectuses, including “Inventory … SFAS, dollar, …, currency, … award … store, restaurant” are all entity names), an objective (Tao, pg. 60, 1st column, last paragraph, “Example FLS From Step 2: ‘We intend to use the net proceeds from this offering to repay indebtedness’” includes the objective “we intend”), and a principle investment strategy (Tao, pg. 60, 1st column, last paragraph, “Example FLS From Step 2: ‘We intend to use the net proceeds from this offering to repay indebtedness’” where “to repay indebtedness” is a principle strategy of what to do with the investment “net proceeds from this offering”, thus a principle investment strategy).
Regarding Claim 2, the Tao/Tamilchelvan combination of Claim 1 teaches the computer implemented method of Claim 1 (and thus the rejection of Claim 1 is incorporated). Tao further teaches receiving input data from a third-party vendor (Tao, pg. 58, 1st column, last paragraph, “The data in this study include the final IPO prospectuses (Form 424B4) in successful IPOs of common stocks”) wherein the input data requires classification using the trained machine learning model (Tao, pg. 58, Fig. 1, the input data has not yet been classified by the predictive model, thus it requires classification).
Regarding Claim 3, the Tao/Tamilchelvan combination of Claim 1 teaches the computer implemented method of Claim 1 (and thus the rejection of Claim 1 is incorporated). Tao further teaches wherein each classification label indicates an investment category (Tao, pg. 59, 1st column, 1st paragraph, “(
Y
2
): If the IPO offering price is lower than the closing prices of the first trading day, then 1” where
Y
2
was previously identified as the classification label to be predicted and investments higher/lower than offering price is an investment category).
Regarding Claim 5, the Tao/Tamilchelvan combination of Claim 1 teaches the computer implemented method of Claim 1 (and thus the rejection of Claim 1 is incorporated). Tao has already been shown to teach wherein the natural language technique is a term-frequency, inverse document frequency (TD-IDF) technique (Tao, pg. 61, 2nd column, 3rd paragraph, “we use classic discrete BoWs along with the TF-IDF approach”).
Regarding Claim 6, Tao teaches the computer implemented method of Claim 1 (and thus the rejection of Claim 1 is incorporated). Tao has already been shown to teach wherein each text feature comprises a term in the corresponding text string and the weight of the text feature quantifies importance in mapping the text feature to a classification label (Tao, pg. 61, 2nd column, 3rd paragraph, “we use classic discrete BoWs along with the TF-IDF approach (bi-gram representations), which is consistent with the extant text classification studies” in order to determine features with which to “train our FLS classifier” i.e. use the TF-IDF weights as text features to predict the classification labels)
Regarding Claim 7, the Tao/Tamilchelvan combination of Claim 1 teaches the computer implemented method of Claim 1 (and thus the rejection of Claim 1 is incorporated). Tao further teaches wherein the machine learning model is a multi-class classification model (Tao, pg. 59, 1st column, 1st paragraph, “(
Y
2
): If the IPO offering price is lower than the closing prices of the first trading day, then 1, otherwise zero” denotes two classes predicted by the machine learning model).
Claims 13-16 recites a computer-implemented system comprising components configured to perform precisely the methods of Claims 1, 3, 5, and 6, respectively. As Tao performs their method on a computer (Tao, pg. 64, 2nd column, 2nd-to-last paragraph, “we use a computing system with dual Intel Xeon CPUs”), Claims 13-16 are rejected for reasons set forth in the rejections of Claims 1, 3, 5, and 6, respectively.
Claim 10 is rejected under 35 U.S.C. 103 as being unpatentable over Tao, in view of Tamilchelvan, and further in view of Nugent, US PG Pub 20200387675.
Regarding Claim 10, Tao teaches the computer implemented method of Claim 2 (and thus the rejection of Claim 2 is incorporated). Tao further teaches applying the trained machine learning model on the input data to predict a classification table for the input data and generate a confidence level for the prediction (Tao, pg. 64, 1st column, 4th paragraph, “
y
p
is the prediction probability as emitted by the classifier” is a confidence and 2nd paragraph, “Accuracy” based on “True Positives” denotes that a binary prediction for labels
Y
1
and
Y
2
are made, see Tables 4 &5).
Tao does not teach, but Nugent teaches assigning the predicted classification label to the input data when the confidence level is above a predetermined threshold. (Nugent, Fig. 4, [0068] “confidence threshold may indicate that the respective probability that the content relates to corresponding topic may be high but not sufficiently high to make an automatic determination that the classification is correct”). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use a confidence threshold. The motivation to do so is “that the content relates to corresponding topic may be high but not sufficiently high to make an automatic determination that the classification is correct” (Nugent, [0068]).
Claims 11 and 12 are rejected under 35 U.S.C. 103 as being unpatentable over Tao, in view of Tamilchelvan and Nugent, and further in view of Mizutani, US PG Pub 2021/0263954.
Regarding Claim 11, the Tao/Tamilchelvan/Nugent combination of Claim 10 teaches the computer implemented method of Claim 1 (and thus the rejection of Claim 10 is incorporated). Tao further teaches wherein the input data includes financial information (Tao, pg. 59, 1st column, 1st paragraph, “(
Y
2
): If the IPO offering price is lower than the closing prices of the first trading day, then 1”). Tao/Tamilchelvan/Nugent does not teach but Mizutani teaches a change to a new classification label for the financial information supplied by the third-party vendor (Mizutani, [0028] “there is often a request to change/add a classification label later in addition to a classification label first assumed for data.”) It would have been obvious to a person of ordinary skill in the art before the effective filing date of the invention to incorporate the teachings of Mizutani to the method of Tao. A person of ordinary skill in the art would have been motivated to do so because “it may be necessary for adding/changing a classification label for updating” (Mizutani, [0007]).
Regarding Claim 12, the Tao/Tamilchelvan/Nugent/Mizutani combination of Claim 11 teaches the method of Claim 11 (and thus the rejection of Claim 11 is incorporated). The combination has not yet been shown to teach, but Nugent teaches, determining whether the predicted classification label agrees with the new classification label; and stop processing the input data when there is at least one of (i) no agreement between the predicted and new classification label, or (ii) the confidence level for the prediction does not satisfy a predefined threshold. (Nugent. Fig 4. “illustrates an example of confidence thresholds used to determine whether to automatically classify content, flag content for verification, or determine that the content does not relate to a topic of interest.”). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to stop processing the input data/flag the content for verification when the confidence level does not satisfy a threshold, as does Nugent, in the current Tao/Tamilchelvan/Nugent/Mizutani combination. The motivation to do so is “that the content relates to corresponding topic may be high but not sufficiently high to make an automatic determination that the classification is correct” (Nugent, [0068]), that is, to verify that the model is making predictions in a manner that the user desires.
Claim 19 is rejected under 35 U.S.C. 103 as being unpatentable over Tao, in view of Tamilchelvan, and further in view of Mizutani, US PG Pub 2021/0263954.
Regarding Claim 19, the Tao/Tamilchelvan combination of Claim 13 teaches the computer-implemented system of Claim 13 (and thus the rejection of Claim 13 is incorporated). Tao further teaches to receive input data from a third-party vendor (Tao, pg. 58, 1st column, last paragraph, “The data in this study include the final IPO prospectuses (Form 424B4) in successful IPOs of common stocks”) that requires classification using the trained machine learning model (Tao, pg. 58, Fig. 1, the input data has not yet been classified by the predictive model, thus it requires classification) and wherein the input data includes financial information (Tao, pg. 59, 1st column, 1st paragraph, “(
Y
2
): If the IPO offering price is lower than the closing prices of the first trading day, then 1”). Tao/Tamilchelvan does not teach but Mizutani teaches a change to a new classification label for the financial information supplied by the third-party vendor (Mizutani, [0028] “there is often a request to change/add a classification label later in addition to a classification label first assumed for data.”) It would have been obvious to a person of ordinary skill in the art before the effective filing date of the invention to incorporate the teachings of Mizutani to the method of Tao. A person of ordinary skill in the art would have been motivated to do so because “it may be necessary for adding/changing a classification label for updating” (Mizutani, [0007]).
Claim 20 is rejected under 35 U.S.C. 103 as being unpatentable over Tao, in view of Tamilchelvan and Mizutani, and further in view of Nugent.
Regarding Claim 20, the Tao/Tamilchelvan/Mizutani combination of Claim 19 teaches the computer-implemented system of Claim 19 (and thus the rejection of Claim 19 is incorporated). Tao further teaches to apply the trained machine learning model on the input data to predict a classification label for the input data and generate a confidence level for the prediction (Tao, pg. 64, 1st column, 4th paragraph, “
y
p
is the prediction probability as emitted by the classifier” is a confidence and 2nd paragraph, “Accuracy” based on “True Positives” denotes that a binary prediction for labels
Y
1
and
Y
2
are made, see Tables 4 &5).
The combination does not teach, but Nugent teaches, to determine whether the predicted classification label agrees with the new classification label; and stop processing the input data when there is at least one of (i) no agreement between the predicted and new classification label, or (ii) the confidence level for the prediction does not satisfy a predefined threshold. (Nugent. Fig 4. “illustrates an example of confidence thresholds used to determine whether to automatically classify content, flag content for verification, or determine that the content does not relate to a topic of interest.”). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to stop processing the input data/flag the content for verification when the confidence level does not satisfy a threshold, as does Nugent, in the Tao/Tamilchelvan/Mizutani combination. The motivation to do so is “that the content relates to corresponding topic may be high but not sufficiently high to make an automatic determination that the classification is correct” (Nugent, [0068]), that is, to verify that the model is making predictions in a manner that the user desires.
Response to Arguments
Applicant’s arguments filed December 2nd, 2025, have been fully considered, but are not fully persuasive.
Applicant’s arguments regarding the 35 U.S.C. 101 rejections of the previous office action, as being directed towards an abstract idea without significantly more, have been fully considered, but are unpersuasive.
Applicant first argues that, like Example 39, Claim 22 does not recite an abstract idea. However, Claim 1 clearly recites steps of extracting one or more data of interest from each prospectus, combining the data of interest into a single text string, extracting a plurality of text features, to determine mapping between the classification labels and text features, and creating a confusion matrix to determine where a mismatch occurs, all which are all actions performable in the human mind. Applicant argues that validating, by the computing device, the trained machine learning model by creating a confusion matrix … is not an abstract idea; however, as evidence by Fig. 4 of the instant specification, validating consists of no more than counting a number of outcomes correctly predicted by the model vs the number of outcomes predicted to be in a different category from the true category of the item. This is clearly a mental process.
Applicant next argues that “the claims are integrated into a practical application because the features of the claims provide an improvement in technology, namely the faster training of a machine learning model.” However, this argument is unpersuasive, because a) there is nothing in the specification to support this assertion that “improving training speed” would be a result of the recited steps, and b) there are no specific limitations in the claims that achieve a goal of faster training – applicant appears to focus on the word “extracting” which does not imply “that the full prospectus is not used to train the machine learning model,” as asserted by the applicant. “Extraction” is interpreted, under the broadest reasonable interpretation and in light of the specification, as identifying these features from the text of the prospectus. There is no support in the specification for any improvement in machine learning models, only for using a machine learning model to learn correlations between features and labels.
In additional, validating a model is not an improvement in a technology of machine learning model. It is merely an analysis of a model. The claimed invention merely uses existing machine learning technology in furtherance of the purpose of classifying prospectuses, and recites no improvements in machine learning, or any other, technology.
Applicant’s arguments regarding the prior art rejections of the previous office action have been fully considered, but are alternatively moot or unpersuasive.
Applicant asserts, without argument, that Tao does not disclose training, by the computing device, the machine learning model using the one or more modified training sets and the plurality of classification labels, that determines mappings between the classification labels and the text features, to generate a trained machine learning model, asserting that “nowhere does Tao teach or suggest using machine learning model to map FLS to a plurality of classification labels to determine relationships between FLS and classification labels.” However, this is the entire point of Tao – to map the prospectuses, including forward looking statements which have been extracted, to the labels
Y
1
and
Y
2
which predict future prices. See the previous and current rejections, which map the limitation to pg. 63, 2nd column, of Tao.
Applicant’s arguments regarding the art rejections of the dependent claims and other independent claims rely upon features argued with respect to Claim 1, and are thus also unpersuasive.
Conclusion
Applicant’s amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to BRIAN M SMITH whose telephone number is (469)295-9104. The examiner can normally be reached Monday - Friday, 8:00am - 4pm Pacific.
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, Kakali Chaki can be reached on (571) 272-3719. 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.
/BRIAN M SMITH/Primary Examiner, Art Unit 2122