DETAILED ACTION
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 .
Response to Amendment
The amendment filed 26 June 2026 has been entered. Applicant amended claims 1, 5-6, 10-11, 13-14, 18-19; added new claims 21-22; and canceled claims 8 and 16. Accordingly, claims 1-7, 9-15, and 17-22 remain pending.
Response to Arguments
Claim Rejection under 35 USC 112(a):
Applicant's arguments filed 26 June 2026 have been fully considered but they are not persuasive.
Applicant’s remarks:
Claims 1-20 were rejected under 35 U.S.C. § 112(a) as failing to comply with the written description requirement. Specifically, the Office alleged that the limitation "outputting the trained machine learning model configured to process input data indicative of healthcare information and consumer information related to the individual and to output an output value that is indicative of a probability that the individual takes a particular action" was too broad because the specification describes only one concrete species-purchasing likelihood for audience modeling-and does not demonstrate possession of a model configured to yield a probability for any action. Additionally, Claims 1-20 were also rejected under 35 U.S.C. § 112(b) as indefinite
because the limitation "determining a risk of disclosure is above a disclosure threshold" did not explain how the risk is computed or quantified. Applicant respectfully disagrees.
Without acquiescing to the rejections, and solely to advance prosecution, Applicant amends the claims as proposed by the Office, rendering the rejections moot. Accordingly, Applicant respectfully requests withdrawal of the rejections under 35 U.S.C. § 112.
Examiner’s remarks:
The claims recite a genus- machine learning model. A machine learning model entails representative number of species such as neural networks, random forest, classification algorithms, support vector machines, clusters etc.,. Machine learning model embodies a variety of machine learning algorithms with different structures and configuration. Applicant’s claims is claiming the entire scope of machine learning models. Applicant’s specification does not adequately disclose a representative number of species for the machine learning model. Applicant’s application describes one concrete species of a particular action which is purchasing likelihood for audience modeling with per entity scores and tiered outputs and it describes one pathway that combines synthetic health and consumer trends to produce cohort probability identifiers. It does not describe other action types or a general framework that demonstrates possession of the full breadth of a model configured to take any healthcare and any consumer inputs to yield an individual probability for any action. The written description requirement for a claimed genus may be satisfied through sufficient description of a representative number of species by actual reduction to practice, reduction to drawings, or by disclosure of relevant, identifying characteristics, see MPEP 2163.
Claim Rejection under 35 USC 112(b): Applicant’s remarks pertaining to the 35 USC 112(b) rejection of 17 February 2026 is not clear because the 35 USC 112(b) rejection do not recite “determining a risk of disclosure is above a disclosure threshold”.
Nonetheless, the amendment, filed 26 June 2026 overcomes the 35 USC 112(b) rejection. Therefore, the 35 USC 112(b) rejection is withdrawn.
Claim Rejection under 35 USC 101:
Applicant's arguments filed 26 June 2026 have been fully considered but they are not persuasive.
Applicant’s remarks 1:
The instant claims do not merely recite "mental processes" or "mathematical concepts," but instead recite several detailed steps and features that point to specific improvements in data processing and machine learning system capabilities. Accordingly, Applicant submits that the Office Action improperly "oversimplifies] the claims by looking at them generally and failing to account for the specific requirements of the claims." MPEP § 2106.05(a) (internal quotation marks and citations omitted).
Applicant’s support:
Amended Claim 1 recites "training, by a processor configured to access training data comprising the stored synthetic trends dataset from the storage device, a machine learning model configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data, wherein the training data comprises the combined sequences of embedding vectors that represent the healthcare dataset and the supplementary dataset in fewer dimensions than the healthcare dataset and the supplementary dataset." This training step specifies a concrete technical process for training a machine learning model on dimensionally- reduced embedding vectors that preserve the statistical relationships of the original datasets while requiring less storage. A human mind cannot practically train a machine learning model on combined sequences of embedding vectors representing healthcare and supplementary datasets in fewer dimensions. This is not a mental process or abstract mathematical concept it is a specific technical implementation of machine learning model training on privacy-preserving synthetic data.
Examiner’s remarks:
The examiner’s maintains that the claims continue to recite an abstract idea. The mental process abstract idea are the steps of “anonymizing the data stored in the databases, wherein the anonymized data from the databases is stored in the corresponding segregated databases; generating, for the segregated data environments, a numerical representation of the data feature of a plurality of data features of the anonymized data stored in the corresponding segregated data environment, wherein the numerical representation comprises a first sequence of embedding vectors…; determining, for the segregated data environments, a second sequence of embedding vectors, wherein the second sequence of embedding vectors is a transformation of the corresponding first sequence of embedding vectors, the transformation comprising a reduction of information from the first sequence of embedding vectors; generating… the synthetic trends dataset by combining the second sequence of embedding vectors associated with the segregated data environments….; training a machine learning model configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data, wherein the training data comprises the combined sequences of embedding vectors that represents the healthcare dataset and the supplementary dataset in fewer dimensions than the healthcare dataset and the supplementary dataset, and deploying the trained machine learning model to process input data indicative of healthcare information and consumer information related to the individual and output an output value that is indicative of a probability that the individual purchases a healthcare-related product, wherein the output value enables privacy-preserving targeted communications related to the healthcare-related product”.
It has been determined that the limitations above recite an abstract idea of mental process and mathematical concepts. A machine learning model is a genus limitation, wherein applying the broadest reasonable interpretation, training the machine learning model can be achieved manually in the human mind using pencil and paper via gathering data, calculating outputs, and adjusting weights.
The claims and specification have not provided nor shown evidence that there is an improvement in how the machine learning model operates internally (improvement how the machine learning model is trained for better results).
Applicant’s remarks 2:
The claims are patent eligible because they integrate any alleged judicial exception into a practical application. As the MPEP explains, "the claimed invention may integrate the judicial exception into a practical application by demonstrating that it improves the relevant existing technology although it may not be an improvement over well- understood, routine, conventional activity." MPEP § 2106.04(d)(1).
Applicant’s support:
Amended Claims 1, 11, and 19 recite that the training data "comprises the combined sequences of embedding vectors that represent the healthcare dataset and the supplementary dataset in fewer dimensions than the healthcare dataset and the supplementary dataset," and that the stored synthetic trends dataset "requires less storage and has less PII pertaining to the individual than a combination of the received healthcare dataset and the received supplementary dataset." Like Desjardins, which credited reduced storage and reduced system complexity as technological improvements, the present claims specify how the machine learning model is trained on data that requires less storage (embedding vectors in fewer dimensions) while preserving the statistical relationships needed for accurate predictions. The claims further recite "deploying the trained machine learning model to process input data indicative of healthcare information and consumer information related to the individual and output an output value that is indicative of a probability that the individual purchases a healthcare-related product, wherein the output value enables privacy-preserving targeted communications related to the healthcare-related product." Together, these limitations specify both how the ML model is trained on dimensionally reduced, privacy- preserving data and how its output enables a concrete technological result-privacy-preserving targeted communications. This is an improvement to how the data processing system itself operates, not just applying generic machine learning to new data, which distinguishes the present claims from Recentive Analytics, Inc. v. Fox Corp., 134 F.4th 1205 (Fed. Cir. 2025)…. Accordingly, the claimed technology solves these issues as described in the claim limitations reciting "training, by a processor configured to access training data comprising the stored synthetic trends dataset from the storage device, a machine learning model configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data, wherein the training data comprises the combined sequences of embedding vectors that represent the healthcare dataset and the supplementary dataset in fewer dimensions than the healthcare dataset and the supplementary dataset" and "deploying the trained machine learning model to process input data indicative of healthcare information and consumer information related to the individual and output an output value that is indicative of a probability that the individual purchases a healthcare-related product, wherein the output value enables privacy-preserving targeted communications related to the healthcare-related product." The amended claims address these technical problems by specifying both how the ML model is trained on dimensionally reduced embedding vectors and how its output enables privacy-preserving targeted communications.
Like Desjardins, these improvements are "tantamount to how the machine learning model itself would function in operation" and are not merely applying generic machine learning to new data.
For at least the above reasons, Claims 1-20 are patent eligible at the Step 2A analysis.
Examiner’s remarks:
Training the data and deploying the training machine learning model to output a probability are limitations, as disclosed above, of an abstract idea-mental process and mathematical concepts. The judicial exception alone cannot provide the improvement. The improvement can be provided by one or more additional elements. See Diamond v. Diehr, 450 U.S. 175, 187 and 191-92, 209 USPQ 1, 10 (1981)). In addition, the improvement can be provided by an additional element(s) in combination with the recited judicial exception.
The additional element of stored synthetic trend dataset resulting in less storage and has less PII pertaining to the individual than a combination of the received healthcare dataset and the received supplementary dataset is determined to be storage of data that is post solution activity that is well-understood, routine, and conventional.
The claims and specification have not provided nor shown evidence that there is an improvement in how the machine learning model operates internally (improvement how the machine learning model is trained for better results). The claims merely recite that training by a processor is performed on the machine learning model. Using the training data, the machine learning model is configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data. The claim recites that the machine learning model is configured to; however, this is not a positive limitation. “Configured to” can merely imply that the machine learning model is set up or arranged to achieve a function, but the function is not required to execute. Therefore, the claims are applying a machine established methods of machine learning to a new data environment. Claims that do not more than apply established methods of machine learning to a new data environment are ineligible under 35 USC 101 (Recentive Analytics, Inc. v. Fox Corp., (April 18, 2025).
Applicant’s remarks 3:
[The claims] are patent eligible because they recite an inventive concept that is significantly more than the alleged abstract idea.
Applicant’s support:
Accordingly, the present claims recite an inventive concept for at least the same reasons as discussed above with respect to a practical application. Amended Claims 1, 11, and 19 recite "training, by a processor configured to access training data comprising the stored synthetic trends dataset from the storage device, a machine learning model configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data, wherein the training data comprises the combined sequences of embedding vectors that represent the healthcare dataset and the supplementary dataset in fewer dimensions than the healthcare dataset and the supplementary dataset" and "deploying the trained machine learning model to process input data indicative of healthcare information and consumer information related to the individual and output an output value that is indicative of a probability that the individual purchases a healthcare-related product, wherein the output value enables privacy- preserving targeted communications related to the healthcare-related product."
These elements are significantly more than an abstract idea because they specify a particular machine learning implementation where the model is trained on combined embedding vectors representing healthcare and supplementary datasets in fewer dimensions, and the model's output enables privacy-preserving targeted communications. The combination of training on dimensionally reduced, privacy-preserving synthetic data and deploying the model to enable targeted communications without exposing the underlying PII is not well-understood, routine, or conventional activity, and represents a specific technical improvement to how data processing and machine learning systems operate. Claims that specify how a machine learning model is trained on privacy-preserving embedding vectors and how its output enables privacy-preserving communications reflect improvements to how the system itself operates and are not merely well- understood, routine, or conventional activity.
Examiner’s remarks:
Training the data and deploying the training machine learning model to output a probability are limitations, as disclosed above, of an abstract idea-mental process and mathematical concepts. The judicial exception alone cannot provide the improvement. The improvement can be provided by one or more additional elements. See Diamond v. Diehr, 450 U.S. 175, 187 and 191-92, 209 USPQ 1, 10 (1981)). In addition, the improvement can be provided by an additional element(s) in combination with the recited judicial exception.
The additional element of stored synthetic trend dataset resulting in less storage and has less PII pertaining to the individual than a combination of the received healthcare dataset and the received supplementary dataset is determined to be storage of data that is post solution activity that is well-understood, routine, and conventional.
The claims and specification have not provided nor shown evidence that there is an improvement in how the machine learning model operates internally (improvement how the machine learning model is trained for better results). The claims merely recite that training by a processor is performed on the machine learning model. Using the training data, the machine learning model is configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data. The claim recites that the machine learning model is configured to; however, this is not a positive limitation. “Configured to” can merely imply that the machine learning model is set up or arranged to achieve a function, but the function is not required to execute. Therefore, the claims are applying a machine established methods of machine learning to a new data environment. Claims that do not more than apply established methods of machine learning to a new data environment are ineligible under 35 USC 101 (Recentive Analytics, Inc. v. Fox Corp., (April 18, 2025).
Claim Rejection under 35 USC 103:
Applicant’s arguments with respect to the independent claims limitations of “deploying the trained machine learning model to process input data indicative of healthcare information and consumer information related to the individual and output an output value that is indicative of a probability that the individual purchases a healthcare-related product, wherein the output value enables privacy-preserving targeted communications related to the healthcare-related product” have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Claim Rejections - 35 USC § 112
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
Claims 1-20 are rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the written description requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor at the time the application was filed, had possession of the claimed invention.
Regarding claims 1, 11, and 19 and the limitations of “training by a processor configured to access training data comprising the stored synthetic trends dataset from the storage device, a machine learning model using the training data”, the specification does not provide adequate written description for the claimed training such that a person of ordinary skill in the art would understand that the inventor possessed the claimed invention at the time of tiling. The written description fails to provide: specific, concrete embodiments that teach a practical way to implement the training of the machine learning model using the training data, including equations and algorithmic steps. The written description does not provide concrete, non-ambiguous examples with actual datasets illustrating the training of the machine learning model using the training data set.
Claims 1, 11, and 19 is also rejected for reciting new matter. The examiner does not find in the original disclosure “the output value enables privacy preserving targeted communications related to the healthcare related product”.
Claim 21 is rejected for reciting new matter. The examiner does not find in the original disclosure “generate the output value in real time”.
Claim 22 is rejected for reciting new matter. The examiner does not find in the original disclosure “online identifier linked to the individual, and wherein the online identifier enables delivery of the targeted communications”.
Claims 2-10 are rejected because said claims depend upon rejected claim 1.
Claims 12-18 are rejected because said claims depend upon rejected claim 11.
Claim 20 is rejected because said claims depend upon rejected claim 19.
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 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea of a mental process without significantly more. The claim(s) recite(s) “anonymizing the data stored in the databases, wherein the anonymized data from the databases is stored in the corresponding segregated databases; generating, for the segregated data environments, a numerical representation of the data features of a plurality of data features of the anonymized data stored in the corresponding segregated data environment, wherein the numerical representation comprises a first sequence of embedding vectors…; determining, for the segregated data environments, a second sequence of embedding vectors, wherein the second sequence of embedding vectors is a transformation of the corresponding first sequence of embedding vectors, the transformation comprising a reduction of information from the first sequence of embedding vectors; generating… the synthetic trends dataset that comprises the second sequence of embedding vectors associated with the segregated data environments….; training…a machine learning model; process input data indicative of healthcare information and consumer information related to the individual and output an output value that is indicative of a probability that the individual purchases healthcare product…”.
The limitations above pertaining to a method for generating a synthetic trends dataset are directed to steps that under its broadest reasonable interpretation covers performance of the limitations being a mental process abstract idea and mathematical concepts abstract idea. The steps can manually be performed by a human using pencil and paper. The output is a probability that an individual takes a particular action which is a targeted behavior prediction-organized human activity abstract idea. The claims fall within the mental process grouping of abstract idea, mathematical concept grouping of abstract idea, and organized human activity grouping of abstract idea. Accordingly, claims 1, 11, and 19 pertain to an abstract idea.
This judicial exception is not integrated into a practical application. Claims 1, 11, and 19 recite additional elements which are “receiving the healthcare dataset from a database in a first segregated data environment of a federated data cleanroom, the healthcare dataset comprising personally identifiable information (PII) pertaining to the individual; receiving the supplementary dataset from a database in a second segregated data environment of the federated data cleanroom”, “storing the synthetic trends dataset in a storage device, wherein the stored synthetic trends dataset comprising the combined sequence of embedding vectors requires less storage and has less PII….”, “training, by a processor configured to access training data comprising the stored synthetic trends dataset from the storage device, a machine learning module using the training data; and outputting the trained machine learning module configured to process input data indicative of healthcare information and consumer information related to the individual”. Claims 1, 11, and 19 recite limitations pertaining to computer components wherein the method is computer implemented.
The additional elements listed above of “receiving the healthcare dataset….; receiving the supplementary dataset…” are merely transmission and receiving of data which is insignificant extra-solution activity and activity that is well understood and conventional. Thus, the additional elements do not integrate the abstract idea into a practical application.
The additional element of ““storing the synthetic trends dataset in a storage device, wherein the stored synthetic trends dataset comprising the combined sequence of embedding vectors requires less storage and has less PII…” is merely storage of data that is post solution activity that is routine, conventional, and well-known and understood. The additional element of the “embedding vectors requiring less storage and less PII” is an intended benefit and field of use and an element that is considered insufficient to be more than the abstract idea. Thus, the additional elements do not integrate the abstract idea into a practical application.
The additional elements of ““training, by a processor configured to access training data comprising the stored synthetic trends dataset from the storage device, a machine learning module using the training data; and outputting the trained machine learning module configured to process input data indicative of healthcare information and consumer information related to the individual” are activities that are well understood, routine, and conventional. The additional element of the machine learning model merely applies the machine learning module without disclosing any technological advances to the underlying machine learning technique. Applying a machine learning model without disclosing any technological advances to the underlying machine learning technique is not sufficient for patent eligibility under 35 USC 101.
The additional elements recited in claims 11 and 19 pertaining to the computer components and the shared data environment wherein the method are executed by computer via instructions and claim 1 recitation of the processor entails the steps being computer implemented and are recited at a high level of generality such that the generic computer amounts to no more than mere instructions to apply the exception. Implementing an abstract idea on generic computer does not integrate the abstract idea into a practical application nor add significantly more.
Thus, the independent claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception. Accordingly, claims 1, 11, and 19 are not eligible under 35 USC 101.
The limitations recited in claims 2-4, 12, and 20, under the broadest reasonable interpretation, further narrow the transformation step that is recited in the independent claims. Claims 2-4, 12, and 20 do not provide additional elements that integrate the abstract idea into a practical application. Thus claims 2-4, 12, and 20 are not eligible under 35 USC 101.
The limitations recited in claims 5, 9, 13, and 17 under the broadest reasonable interpretation, further narrow the generation of the first sequence of embedding vectors that is recited in the independent claims. Claims 5, 9, 13, and 17do not provide additional elements that integrate the abstract idea into a practical application. Thus claims 5, 9, 13, and 17 are not eligible under 35 USC 101.
The limitations recited in claims 6 and 14, under the broadest reasonable interpretation, provide an additional limitation that can be implemented in the human mind using pencil and paper. Claims 6 and 14 do not provide additional elements that integrate the abstract idea into a practical application. For the same reasons as recited in the independent claims, claims 6 and 14 are not eligible under 35 USC 101.
The limitations recited in claims 7 and 15, under the broadest reasonable interpretation, provide additional limitations that can be implemented in the human mind using pencil and paper. Claims 7 and 15 do not provide additional elements that integrate the abstract idea into a practical application. For the same reasons as recited in the independent claims, claims 7 and 15 are not eligible under 35 USC 101.
The limitations recited in claims 8 and 16, under the broadest reasonable interpretation, further narrow the generation of the risk of disclosure that is recited in the independent claims, wherein said limitations can be performed in the human mind using pencil and paper. Claims 8 and 16 do not provide additional elements that integrate the abstract idea into a practical application. Thus claims 8 and 16 are not eligible under 35 USC 101.
The limitations recited in claims 10 and 18 under the broadest reasonable interpretation, further narrow the transformation step that is recited in the independent claims. Claims 10 and 18 do not provide additional elements that integrate the abstract idea into a practical application. Thus claims 10 and 18are not eligible under 35 USC 101.
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.
Claim(s) 1-7, 9-15, and 17-21 is/are rejected under 35 U.S.C. 103 as being unpatentable over Chakraborty et al US 20190156061 (hereinafter Chakraborty), in further view of Lesh et al US 20230010686 (hereinafter Lesh), in further view of Aptekar et al US 11977550 (hereinafter Aptekar), in further view of Pauly US 20200035365 (hereinafter Pauly), in further view of Woytarowicz et al US 20230298058 (hereinafter Woytarowicz), and in further view of Choudhury et al US 20210150269 (hereinafter Choudhury).
As to claim 1, Chakraborty teaches a method for generating a synthetic trends dataset (abstract discloses generating anonymized synthetic data by adding noise to a first set of data from one or more data stores. Paragraph 78 reveals the data can come from private databases. Claim 19 discloses aggregating/combining the plurality of databases (see Figure 10, reference number 1006). Paragraphs 77-78 reveals the data record are health records from hospital private database. Paragraphs 77-78, 32, 52, and 68 disclose additional data records from private databases that include secondary information such as education level, zip code, and/or salary data) for training a machine learning model (paragraphs 31, 60-62 and steps 608 and 612 reveal the output candidate for anonymization is retrained in the embedding vector model. The data record strings is sent through the word embedding vector model to obtain a candidate output for anonymization/synthetic data), the method comprising:
receiving a healthcare dataset from a database in a first segregated data environment of a federated data cleanroom, the healthcare dataset comprising personally identifiable information (PII) pertaining to the individual (paragraphs 4 and 52 reveal receiving raw data/first data associated with a data stores. The raw data comes from database records, unstructured data, and/or any other set of data. Paragraphs 23 and 68 discloses the datastore record can include information such as disease. Paragraphs 77- 78 reveal the data can come from private databases such as health data from hospital; therefore, the databases are federated. See also Figure 9, reference number 910 of Figure 10, reference number 1008);
receiving a supplementary dataset from a database in a second segregated data environment of the federated data cleanroom, the supplementary data comprising PII pertaining to the individual (paragraphs 4 and 52 reveal receiving raw data/first data associated with a data stores. The raw data comes from database records, unstructured data, and/or any other set of data. Paragraphs 32, 52, and 68 disclose the datastore record can include information such as education level, zip code, and/or salary. The data can also include name of a person, SSN. Paragraphs 77- 78 reveal the data can come from private databases; therefore, the databases are federated. See also Figure 9, reference number 912 or Figure 10, reference number 1010);
anonymizing the data stored in the database, wherein the anonymized data from the databases is stored in the corresponding segregated databases (paragraph 52 discloses anonymizing received data values by deleting, masking, changing, and/or encrypting the dataset such that the data are not discernable to a human reader. Paragraph 79 further discloses each of the private databases are anonymized to a corresponding anonymized database. See also Figure 9, reference numbers 904, 906, and 908);
generating, for the segregated data environments, a numerical representation of the data features of a plurality of data features of the anonymized data stored in the corresponding segregated data environment, wherein the numerical representation comprises a first sequence of embedding vectors (paragraphs 30 and 53-54 disclose for the plurality of real strings and real number values in the datastores of the anonymized datastore, identifying quasi-identifiers/data features and sensitive identifier values for both string data and numeral data; normalizing the data (if the data is a numerical dataset) and generating one or more tensors values for the strings data and the normalized numerical data. Tensors are numerical representation of the data record token that are fed through an embedding vector model (for the word ) or normalize vector. Vector embedding is converting words/numbers into number sequence), wherein the first sequence of embedding vectors for a segregated data environment is a compressed representation of the corresponding data feature (paragraphs 30-31 disclose the data tensor are representations of data in different form than the data itself, such as a token representing the data. Text strings that are fed through the word embedding vector model, the output is the data tensors which may be represented as integers or other real numbers. The data tensor are mapped from its high dimensionality space to a lower dimensionality space, the data is compressed);
determining, for the segregated data environments, a second sequence of embedding vectors, wherein the second sequence of embedding vectors is a transformation of the corresponding first sequence of embedding vectors, the transformation comprising a reduction of information from the first sequence of embedding vectors (paragraph 56 discloses adding noise to each tensor vector of the anonymized databases. Figure 4 and paragraphs 45-50 illustrate the noise propagation module which include input layer, a hidden layer, an anonymizer layer, and an output layer. The tensors are represented as a dimensional vector. An encoder maps each data point to a representation point h (e.g., h.sub.1, h.sub.2, or h.sub.3) (h=f(d)). The mapping of the entire data set includes extracting relevant features or weights associated with the data points and projecting the features from a first dimension (e.g., a higher dimension) to another dimension (e.g., a lower dimension). For example, a 300 dimensional vector represented as a table of token records (d.sub.1, −d.sub.4) can be mapped to a 100 dimensional vector (h.sub.1, −h.sub.3). This can be useful for principles such as data compression, which reduces dimensionality by removing static or non-important features of data. Each data point within the hidden layer is then mapped to another representation of the same or analogous dimension within the anonymizer layer. The processing for this mapping includes first using a Mondrian algorithm);
generating the synthetic trends dataset (paragraph 48 discloses the output of the Mondrian algorithm is a set of equivalence classes, where each equivalence class is of size at least k to meet k-anonymity. The equivalence class members at this point can be codes (derived from the data vectors using the encoder). Accordingly, at the output of the anonymizer layer (e.g., the anonymizer of FIG. 1), the c1, c2, c3, and c4 are indistinguishable. They are all represented by noisy code C. In some embodiments, different equivalence classes have different centroids and thus different outputs. The output layer is the synthetic trends dataset that comprises the second sequence of embedding vectors/noise associated with the anonymized databases);
training, by a processor configured to access training data comprising the ….synthetic dataset, a machine learning model …using the training data (Figure 1 and paragraphs 31 and 62 disclose the word embedding vector model(s) are retrained using the anonymized data. After the decoder turns the data into the anonymized data, the word embedding vector model(s) is retrained, which is described in more detail below. This is indicated by the arrow from the decoder to the word embedding vector model(s)), wherein the training data comprises the combined sequences of embedding vectors that represents [the datasets] in fewer dimensions than the healthcare dataset and the supplementary dataset (paragraph 37 discloses the training data includes the word embedding vector model. Paragraph 56 discloses adding noise to each tensor vector of the anonymized databases. Figure 4 and paragraphs 45-50 illustrate the noise propagation module which include input layer, a hidden layer, an anonymizer layer, and an output layer. The tensors are represented as a dimensional vector. An encoder maps each data point to a representation point h (e.g., h.sub.1, h.sub.2, or h.sub.3) (h=f(d)). The mapping of the entire data set includes extracting relevant features or weights associated with the data points and projecting the features from a first dimension (e.g., a higher dimension) to another dimension (e.g., a lower dimension). For example, a 300 dimensional vector represented as a table of token records (d.sub.1, −d.sub.4) can be mapped to a 100 dimensional vector (h.sub.1, −h.sub.3). This can be useful for principles such as data compression, which reduces dimensionality by removing static or non-important features of data. Each data point within the hidden layer is then mapped to another representation of the same or analogous dimension within the anonymizer layer. The processing for this mapping includes first using a Mondrian algorithm; and
…the trained machine learning model to process input data indicative of healthcare information and consumer information related to the individual (Figures 1 and 6 and paragraphs 31, 58, and 62 reveal the retrain embedded vector model processes data record strings/input data. Paragraph 78 reveals the data can come from private databases. Claim 19 discloses aggregating/combining the plurality of databases (see Figure 10, reference number 1006). Paragraphs 77-78 reveals the data record are health records from hospital private database. Paragraphs 77-78, 32, 52, and 68 disclose additional data records from private databases that include secondary information such as education level, zip code, and/or salary data).
Chakraborty does not teach generating, in a shared data environment, the synthetic trends dataset by combining the second sequence of embedding vectors associated with the segregated data environments; storing the synthetic trends dataset in a storage device, wherein the stored synthetic trends dataset comprising the combined sequences of embedding vectors requires less storage and has less PII pertaining to the individual than a combination of the received healthcare dataset and the received supplementary dataset; training a machine learning model configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data, wherein the training data comprises the combined sequences of embedding vectors that represents the healthcare dataset and the supplementary dataset; and deploying the trained machine learning model to process input data and to output an output value that is indicative of a probability that the individual purchases a healthcare related product, wherein the output value enables privacy preserving targeted communication related to the healthcare related product.
Lesh teaches generating, in a shared data environment, the synthetic trends dataset by combining the second sequence of embedding vectors associated with the segregated data environments (paragraphs 64-65 and Figure 5B disclose the medical records (from databases (see paragraph 48) for a patient is transformed and for each record, a numerical vector/embedded vector is generated. The medical records are transformed into concatenated vector using the embedding vectors and are inputted into a generative model (shared data environment, per paragraph 48 the synthetic data is generated from a single or multiple source database ) to generate synthetic dataset), wherein the synthetic trends dataset has [threshold metric difference] with the healthcare dataset and the supplementary dataset (paragraph 66 further discloses a discriminator can determine a distribution over the state of the same based on the received input of the synthetic dataset and the authentic medical datasets (which can be the combined healthcare dataset and the dosage dataset), detecting whether the sample is synthetic. Paragraphs 80-83 also discloses computing distance metrics that compare the real records to the synthetic records. After comparing, if any of the pairs have a distance value that is too low or within a threshold distance, the system may highlight that pair for manual review by a human to determine if any authentic patient medical data has leaked into the generated synthetic dataset. The synthetic dataset may be assumed validated if a certain threshold of confidence between a relevant metric of predictive performance is achieved. For statistical validation, the generative system may utilize statistical measures over a population to demonstrate that the generated data has similar characteristics to the original/authentic data. For example, the system may utilize a distribution visualization called a violin plot to determine whether the synthetic data is comparable to the authentic data within an acceptable threshold. The system may also utilize other statistical validation such as a plot of a univariate correlation between each pair of variables in the original authentic electronic health record dataset and compare this against the same correlation plot in the synthetic electronic health record dataset).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify Chakraborty’s method of generating synthetic trends with Lesh’s teachings of combining the embedded vectors to generate the synthetic dataset using machine learning models such that the system can provide quality synthetic datasets that provide a threshold distance with the authentic datasets, and quickly update the synthetic dataset using the machine learning model, such that the gradient/distance in distribution between the synthetic dataset and the authentic dataset meets a threshold value (paragraph 95 of Lesh).
The combination of Chakraborty in view of Lesh does not teach storing the synthetic trends dataset in a storage device, wherein the stored synthetic trends dataset comprising the combined sequences of embedding vectors requires less storage and has less PII pertaining to the individual than a combination of the received healthcare dataset and the received supplementary dataset; training a machine learning model configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data, wherein the training data comprises the combined sequences of embedding vectors that represents the healthcare dataset and the supplementary dataset; and deploying the trained machine learning model to process input data and to output an output value that is indicative of a probability that the individual purchases a healthcare related product, wherein the output value enables privacy preserving targeted communication related to the healthcare related product.
Aptekar teaches storing the synthetic trends dataset in a storage device (column 3, lines 62-65 disclose outputting the synthetic dataset to a storage system), wherein the stored synthetic trends dataset comprising the combined sequences of embedding vectors requires less storage (column 11, lines 11-40 disclose the synthetic dataset is generated based on combined sequences/subsequence patterns of feature vector vectors. The feature vectors are embedded in a lower dimension space).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify Chakraborty’s method of generating synthetic trends in view of Lesh’s teachings of combining the embedded vectors to generate the synthetic dataset using machine learning models with Aptekar’s teachings of storing the synthetic dataset to provide improve data fidelity and data privacy by obfuscating the data in a way that preserves its statistical properties while rendering it unidentifiable to unauthorized users. This may allow researchers to more widely access and analyze such data at an earlier stage of clinical trials, which can accelerate the development of new drugs and medical treatments (column 2, lines 2-9 of Aptekar).
The combination of Chakraborty in view of Kwatra, Lesh, and Aptekar does not teach but wherein the synthetic trends dataset has less PII pertaining to the individual than a combination of the received healthcare dataset and the received supplementary dataset; training a machine learning model configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data, wherein the training data comprises the combined sequences of embedding vectors that represents the healthcare dataset and the supplementary dataset; and deploying the trained machine learning model to process input data and to output an output value that is indicative of a probability that the individual purchases a healthcare related product, wherein the output value enables privacy preserving targeted communication related to the healthcare related product.
Pauly teaches wherein the synthetic trends dataset has less PII pertaining to the individual than a combination of the received healthcare dataset and the received supplementary dataset (paragraphs 102 and 147 disclose the encoded output data/synthetic data has a dimension that is lower than the medical data. The length of a vector representing medical data of the patient may be longer than the length of a vector that is an encoded output matrix encoded from the vector of medical data by an encoding module. Paragraph 4 reveals the data can include imaging, laboratory test measurements, clinical and family history data. Family history can be supplementary dataset. Paragraph 207 reveals the encoded data set has a higher degree of privacy than the raw dataset).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify Chakraborty’s method of generating synthetic trends in view of Lesh’s teachings of combining the embedded vectors to generate the synthetic dataset using machine learning models and Aptekar’s teachings of storing the synthetic dataset with Pauly’s teachings of the encoded data set having less PII such that the data can be sent to remote computing environments that have lower privacy standard without compromising the data (paragraph 207 of Pauly).
The combination of Chakraborty in view of Lesh, Aptekar, and Pauly does not teach training a machine learning model configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data, wherein the training data comprises the combined sequences of embedding vectors that represents the healthcare dataset and the supplementary dataset; and deploying the trained machine learning model to process input data and to output an output value that is indicative of a probability that the individual purchases a healthcare related product, wherein the output value enables privacy preserving targeted communication related to the healthcare related product.
Woytarowicz teaches training a machine learning model configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data (paragraph 54 discloses the training data comprises examples of values for the features. In some example embodiments, the training data comprises labeled data with examples of values for the features and labels indicating the outcome, such as a purchase, a visit to a website, adding an item to a cart, etc. The machine-learning algorithms utilize the training data to find correlations/relationships among identified features that affect the outcome. Paragraphs 63 further disclose training the ML algorithm includes labeled data which is known data for one or more identified features. Per paragraph 61, the features include features associated with the user profile (e.g., name, persona ID, email address, phone number, city of residence, etc.), and features associated with events (e.g., purchases, timestamp, website views, amount spent, time window for the events, etc.), wherein the training data comprises the combined sequences of embedding vectors that represents the healthcare dataset and the supplementary dataset in fewer dimensions than the healthcare dataset and the supplementary dataset (paragraph 70 discloses the embedding vector that represents user information (personal information such as user identifier. Paragraph 125 reveals the dataset can include age which is health data) and event information (supplementary data); deploying the trained machine learning model to process input data and to output an output value that is indicative of a probability that the individual purchases a healthcare related product (paragraph 21 discloses the propensity ML model being trained with training data comprising values for features that include event features, user information features, and audience labels; and generating, by the propensity ML model, a score for the user indicating a probability that the user will purchase the first product).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify Chakraborty’s method of generating synthetic trends in view of Lesh’s teachings of combining the embedded vectors to generate the synthetic dataset using machine learning models, Aptekar’s teachings of storing the synthetic dataset, and Pauly’s teachings of the encoded data set having less PII wherein the output value enables privacy preserving targeted communication related to the healthcare related product with Woytarowicz’s teachings of training a machine learning model to provide a system that monitor customers and consolidate customer information from multiple sources. The consolidation information may be used by machine-learning systems that predict possible purchases of customers, as well as prioritize those customers with high purchasing potential (paragraph 17 of Woytarowicz).
The combination of Chakraborty in view of Lesh, Aptekar, and Pauly does not teach wherein the output value enables privacy preserving targeted communication related to the healthcare related product.
Choudhury teaches training a machine learning model (paragraph 25 discloses training the federated ML model), deploying the trained machine learning model to output an output value that is indicative of a probability that the individual purchases a healthcare related product (paragraph 20 discloses the machine learning model may be trained to predict a user’s probability of purchasing an item); wherein the output value enables privacy preserving targeted communication related to the healthcare related product (paragraph 27 discloses applying syntactic anonymization to a federated learning system to privacy-protect data being shared among various components of the federated learning system).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify Chakraborty’s method of generating synthetic trends in view of Lesh’s teachings of combining the embedded vectors to generate the synthetic dataset using machine learning models, Aptekar’s teachings of storing the synthetic dataset, Pauly’s teachings of the encoded data set having less PII wherein the output value enables privacy preserving targeted communication related to the healthcare related product, and Woytarowicz’s teachings of training a machine learning model with Choudhury’s teaching of privacy preserving to facilitate compliance with privacy regulations that provide requirements around adequate data de-identification/data anonymization (paragraph 5 of Choudhury).
As to claim 2, the combination of Chakraborty in view of Lesh, Aptekar, Pauly, Woytarowicz, and Choudhury teaches wherein the transformation of the first sequence of embedding vectors comprises reducing a dimensionality of the first sequence of embedding vectors (Chakraborty: paragraph 56 discloses the transformation of the first embedded vector/sequence by adding noise to each tensor vector of the anonymized databases. Figure 4 and paragraphs 45-50 illustrate the noise propagation module which include input layer, a hidden layer, an anonymizer layer, and an output layer. The tensors are represented as a dimensional vector. An encoder maps each data point to a representation point h (e.g., h.sub.1, h.sub.2, or h.sub.3) (h=f(d)). The mapping of the entire data set includes extracting relevant features or weights associated with the data points and projecting the features from a first dimension (e.g., a higher dimension) to another dimension (e.g., a lower dimension). For example, a 300 dimensional vector represented as a table of token records (d.sub.1, −d.sub.4) can be mapped to a 100 dimensional vector (h.sub.1, −h.sub.3). This can be useful for principles such as data compression, which reduces dimensionality by removing static or non-important features of data. Each data point within the hidden layer is then mapped to another representation of the same or analogous dimension within the anonymizer layer. The processing for this mapping includes first using a Mondrian algorithm) .
As to claim 3, the combination of Chakraborty in view of Lesh, Aptekar, Pauly, Woytarowicz, and Choudhury teaches wherein the transformation of the first sequence of embedding vectors comprises a lossy compression of the first sequence of embedding vectors (Chakraborty: paragraph 56 discloses the transformation of the first embedded vector/sequence by adding noise to each tensor vector of the anonymized databases. Figure 4 and paragraphs 45-50 illustrate the noise propagation module which include input layer, a hidden layer, an anonymizer layer, and an output layer. The tensors are represented as a dimensional vector. An encoder maps each data point to a representation point h (e.g., h.sub.1, h.sub.2, or h.sub.3) (h=f(d)). The mapping of the entire data set includes extracting relevant features or weights associated with the data points and projecting the features from a first dimension (e.g., a higher dimension) to another dimension (e.g., a lower dimension). This can be useful for principles such as data compression, which reduces dimensionality by removing static or non-important features of data. Each data point within the hidden layer is then mapped to another representation of the same or analogous dimension within the anonymizer layer. The processing for this mapping includes first using a Mondrian algorithm) .
As to claim 4, the combination of Chakraborty in view of Lesh, Aptekar, Pauly, Woytarowicz, and Choudhury teaches wherein the transformation of the first sequence of embedding vectors comprises adding noise to the first sequence of embedding vectors, wherein the noise comprises one or more sources of noise (Chakraborty: paragraph 56 discloses nose is added by the noise propagation module (source of noise) to the tensors which are the first sequence of embedding vectors).
As to claim 5, the combination of Chakraborty in view of Lesh, Aptekar, Pauly, Woytarowicz, and Choudhury teaches wherein the generation of the first sequence of embedding vectors for the segregated data environments comprises a principal component analysis of the corresponding dataset (Chakraborty: paragraph 27 discloses analysis of the dataset via the embedding vector module to generate the tensor vectors).
As to claim 6, the combination of Chakraborty in view of Lesh, Aptekar, Pauly, Woytarowicz, and Choudhury teaches further comprising generating a token from the PII of the healthcare dataset and a token from the PII of the supplementary dataset, wherein the tokens are operative to link corresponding datasets to data stored outside of the federated data cleanroom (Chakraborty: paragraphs 30, 33, and 35 disclose the generated data tensors are representations of data in a different form, such as a token representing the dataset and are linked/corresponds to the real data values in the data records (health and supplementary as disclosed in claim 1above) ).
As to claim 7, the combination of Chakraborty in view of Lesh, Aptekar, Pauly, Woytarowicz, and Choudhury teaches further comprising:
determining a utility of the synthetic trends dataset, wherein the utility is indicative of a quality of the synthetic trends dataset with respect to a particular task (Lesh: paragraphs 80-83 discloses computing distance metrics that compare the real records to the synthetic records. After comparing, if any of the pairs have a distance value that is too low or within a threshold distance, the system may highlight that pair for manual review by a human to determine if any authentic patient medical data has leaked into the generated synthetic dataset. The synthetic dataset may be assumed validated if a certain threshold of confidence between a relevant metric of predictive performance is achieved. See also paragraphs 94-95 regarding determining first and second gradients to generate synthetic data records);
determining that the utility of the synthetic trends dataset is below a utility threshold that represents a minimum required quality of insights generated based on analytics of the synthetic trends dataset (Lesh: paragraphs 80-83 discloses computing distance metrics that compare the real records to the synthetic records. After comparing, if any of the pairs have a distance value that is too low or within a threshold distance, the system may highlight that pair for manual review by a human to determine if any authentic patient medical data has leaked into the generated synthetic dataset. The synthetic dataset may be assumed validated if a certain threshold of confidence between a relevant metric of predictive performance is achieved);
modifying, based on the utility of the synthetic trends dataset being below the utility threshold, one or more of the determined data features to increase the utility of the synthetic trends dataset (Lesh: paragraphs 80-83 discloses computing distance metrics that compare the real records to the synthetic records. After comparing, if any of the pairs have a distance value that is too low or within a threshold distance, the system may highlight that pair for manual review by a human to determine if any authentic patient medical data has leaked into the generated synthetic dataset. The synthetic dataset may be assumed validated if a certain threshold of confidence between a relevant metric of predictive performance is achieved. Paragraphs 97-95 disclose updating first and second gradients to meet threshold loss distribution that contributes to the generation of the synthetic data records. Paragraphs 5 and 46-47, 51 disclose adding noise to modify and control the generation of the synthetic dataset); and
after the modifying, outputting insights generated based on analytics of the synthetic trends dataset (Woytarowicz: paragraph 21 discloses the propensity ML model being trained with training data comprising values for features that include event features, user information features, and audience labels; and generating, by the propensity ML model, a score for the user indicating a probability that the user will purchase the first product). Motivation is similar to the motivation presented in claim 1.
As to claim 9, the combination of Chakraborty in view of Lesh, Aptekar, Pauly, Woytarowicz, and Choudhury teaches wherein generating the first sequence of embedding vectors comprises capturing a variance of the anonymized data in fewer dimensions than the dimensionality of the anonymized data (Chakraborty: paragraph 33 discloses in generating the tensors, the numerical data is normalized. Normalization captures the variance/difference of a anonymized data value with the mean. The mean is less dimensional than the anonymized dataset. For word data paragraphs 39-40 reveal equation that yields the normalized probabilistic model for language modeling).
As to claim 10, the combination of Chakraborty in view of Lesh, Aptekar, Pauly, Woytarowicz, and Choudhury teaches wherein the healthcare dataset comprises a plurality of alphanumeric codes, wherein the alphanumeric codes are mapped to embedding vectors (Chakraborty: paragraphs 32-34 disclose that the data records includes various fields or attributes with both real numbers and strings. These fields are linked to tensors/embedded vectors. Paragraph 48 disclose the output of the Mondrian algorithm is a set of equivalence classes. The equivalence class members are codes (derived from the data vectors using the encoder)).
As to claim 11, Chakraborty teaches a system (Figure 10 discloses a system computing environment) comprising:
one or more computers (Figure 13, reference number 12 “Computing Device”; paragraph 109 discloses the computing device shown in Figure 13 is a general purpose computing device. The computing device represents computing devices that includes the anonymized databases, and the user devices shown in Figure 13);
one or more computer-readable media (Figure 13, reference number 28 “Memory”) storing instructions that are operable, when executed by the one or more computers, to perform operations (paragraph 112 reveals memory may include at least one program product having a set (e.g., at least one) of program modules that are configured to carry out the functions of embodiments of the invention ) for generating a synthetic trends dataset (abstract discloses generating anonymized synthetic data by adding noise to a first set of data from one or more data stores. Paragraph 78 reveals the data can come from private databases. Claim 19 discloses aggregating/combining the plurality of databases (see Figure 10, reference number 1006). Paragraphs 77-78 reveals the data record are health records from hospital private database. Paragraphs 77-78, 32, 52, and 68 disclose additional data records from private databases that include secondary information such as education level, zip code, and/or salary data) for training a machine learning model (paragraphs 31, 60-62 and steps 608 and 612 reveal the output candidate for anonymization is retrained in the embedding vector model. The data record strings is sent through the word embedding vector model to obtain a candidate output for anonymization/synthetic data), the operations comprising:
receiving a healthcare dataset from a database in a first segregated data environment of a federated data cleanroom, the healthcare dataset comprising personally identifiable information (PII) pertaining to the individual (paragraphs 4 and 52 reveal receiving raw data/first data associated with a data stores. The raw data comes from database records, unstructured data, and/or any other set of data. Paragraphs 23 and 68 discloses the datastore record can include information such as disease. Paragraphs 77- 78 reveal the data can come from private databases such as health data from hospital; therefore, the databases are federated. See also Figure 9, reference number 910 of Figure 10, reference number 1008);
receiving a supplementary dataset from a database in a second segregated data environment of the federated data cleanroom, the supplementary data comprising PII pertaining to the individual (paragraphs 4 and 52 reveal receiving raw data/first data associated with a data stores. The raw data comes from database records, unstructured data, and/or any other set of data. Paragraphs 32, 52, and 68 disclose the datastore record can include information such as education level, zip code, and/or salary. The data can also include name of a person, SSN. Paragraphs 77- 78 reveal the data can come from private databases; therefore, the databases are federated. See also Figure 9, reference number 912 or Figure 10, reference number 1010);
anonymizing the data stored in the databases, wherein the anonymized data from the databases is stored in the corresponding segregated databases (paragraph 52 discloses anonymizing received data values by deleting, masking, changing, and/or encrypting the dataset such that the data are not discernable to a human reader. Paragraph 79 further discloses each of the private databases are anonymized to a corresponding anonymized database. See also Figure 9, reference numbers 904, 906, and 908);
generating, for the segregated data environments, a numerical representation of the data features of a plurality of data features of the anonymized data stored in the corresponding segregated data environment, wherein the numerical representation comprises a first sequence of embedding vectors (paragraphs 30 and 53-54 disclose for the plurality of real strings and real number values in the datastores of the anonymized datastore, identifying quasi-identifiers/data features and sensitive identifier values for both string data and numeral data; normalizing the data (if the data is a numerical dataset) and generating one or more tensors values for the strings data and the normalized numerical data. Tensors are numerical representation of the data record token that are fed through an embedding vector model (for the word ) or normalize vector. Vector embedding is converting words/numbers into number sequence), wherein the first sequence of embedding vectors for a segregated data environment is a compressed representation of the corresponding data feature (paragraphs 30-31 disclose the data tensor are representations of data in different form than the data itself, such as a token representing the data. Text strings that are fed through the word embedding vector model, the output is the data tensors which may be represented as integers or other real numbers. The data tensor are mapped from its high dimensionality space to a lower dimensionality space, the data is compressed);
determining, for the segregated data environments, a second sequence of embedding vectors, wherein the second sequence of embedding vectors is a transformation of the corresponding first sequence of embedding vectors, the transformation comprising a reduction of information from the first sequence of embedding vectors (paragraph 56 discloses adding noise to each tensor vector of the anonymized databases. Figure 4 and paragraphs 45-50 illustrate the noise propagation module which include input layer, a hidden layer, an anonymizer layer, and an output layer. The tensors are represented as a dimensional vector. An encoder maps each data point to a representation point h (e.g., h.sub.1, h.sub.2, or h.sub.3) (h=f(d)). The mapping of the entire data set includes extracting relevant features or weights associated with the data points and projecting the features from a first dimension (e.g., a higher dimension) to another dimension (e.g., a lower dimension). For example, a 300 dimensional vector represented as a table of token records (d.sub.1, −d.sub.4) can be mapped to a 100 dimensional vector (h.sub.1, −h.sub.3). This can be useful for principles such as data compression, which reduces dimensionality by removing static or non-important features of data. Each data point within the hidden layer is then mapped to another representation of the same or analogous dimension within the anonymizer layer. The processing for this mapping includes first using a Mondrian algorithm);
generating the synthetic trends (paragraph 48 discloses the output of the Mondrian algorithm is a set of equivalence classes, where each equivalence class is of size at least k to meet k-anonymity. The equivalence class members at this point can be codes (derived from the data vectors using the encoder). Accordingly, at the output of the anonymizer layer (e.g., the anonymizer of FIG. 1), the c1, c2, c3, and c4 are indistinguishable. They are all represented by noisy code C. In some embodiments, different equivalence classes have different centroids and thus different outputs. The output layer is the synthetic trends dataset that comprises the second sequence of embedding vectors/noise associated with the anonymized databases);
training, by a processor configured to access training data comprising the ….synthetic dataset, a machine learning model …using the training data (Figure 1 and paragraphs 31 and 62 disclose the word embedding vector model(s) are retrained using the anonymized data. After the decoder turns the data into the anonymized data, the word embedding vector model(s) is retrained, which is described in more detail below. This is indicated by the arrow from the decoder to the word embedding vector model(s)), wherein the training data comprises the combined sequences of embedding vectors that represents [the datasets] in fewer dimensions than the healthcare dataset and the supplementary dataset (paragraph 37 discloses the training data includes the word embedding vector model. Paragraph 56 discloses adding noise to each tensor vector of the anonymized databases. Figure 4 and paragraphs 45-50 illustrate the noise propagation module which include input layer, a hidden layer, an anonymizer layer, and an output layer. The tensors are represented as a dimensional vector. An encoder maps each data point to a representation point h (e.g., h.sub.1, h.sub.2, or h.sub.3) (h=f(d)). The mapping of the entire data set includes extracting relevant features or weights associated with the data points and projecting the features from a first dimension (e.g., a higher dimension) to another dimension (e.g., a lower dimension). For example, a 300 dimensional vector represented as a table of token records (d.sub.1, −d.sub.4) can be mapped to a 100 dimensional vector (h.sub.1, −h.sub.3). This can be useful for principles such as data compression, which reduces dimensionality by removing static or non-important features of data. Each data point within the hidden layer is then mapped to another representation of the same or analogous dimension within the anonymizer layer. The processing for this mapping includes first using a Mondrian algorithm; and
…the trained machine learning model to process input data indicative of healthcare information and consumer information related to the individual (Figures 1 and 6 and paragraphs 31, 58, and 62 reveal the retrain embedded vector model processes data record strings/input data. Paragraph 78 reveals the data can come from private databases. Claim 19 discloses aggregating/combining the plurality of databases (see Figure 10, reference number 1006). Paragraphs 77-78 reveals the data record are health records from hospital private database. Paragraphs 77-78, 32, 52, and 68 disclose additional data records from private databases that include secondary information such as education level, zip code, and/or salary data).
Chakraborty does not teach generating, in a shared data environment, the synthetic trends dataset by combining the second sequence of embedding vectors associated with the segregated data environments; storing the synthetic trends dataset in a storage device, wherein the stored synthetic trends dataset comprising the combined sequences of embedding vectors requires less storage and has less PII pertaining to the individual than a combination of the received healthcare dataset and the received supplementary dataset; training a machine learning model configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data, wherein the training data comprises the combined sequences of embedding vectors that represents the healthcare dataset and the supplementary dataset; and deploying the trained machine learning model to process input data and to output an output value that is indicative of a probability that the individual purchases a healthcare related product, wherein the output value enables privacy preserving targeted communication related to the healthcare related product.
Lesh teaches generating, in a shared data environment, the synthetic trends dataset by combining the second sequence of embedding vectors associated with the segregated data environments (paragraphs 64-65 and Figure 5B disclose the medical records (from databases (see paragraph 48) for a patient is transformed and for each record, a numerical vector/embedded vector is generated. The medical records are transformed into concatenated vector using the embedding vectors and are inputted into a generative model (shared data environment, per paragraph 48 the synthetic data is generated from a single or multiple source database ) to generate synthetic dataset), wherein the synthetic trends dataset has [threshold metric difference] with the healthcare dataset and the supplementary dataset (paragraph 66 further discloses a discriminator can determine a distribution over the state of the same based on the received input of the synthetic dataset and the authentic medical datasets (which can be the combined healthcare dataset and the dosage dataset), detecting whether the sample is synthetic. Paragraphs 80-83 also discloses computing distance metrics that compare the real records to the synthetic records. After comparing, if any of the pairs have a distance value that is too low or within a threshold distance, the system may highlight that pair for manual review by a human to determine if any authentic patient medical data has leaked into the generated synthetic dataset. The synthetic dataset may be assumed validated if a certain threshold of confidence between a relevant metric of predictive performance is achieved. For statistical validation, the generative system may utilize statistical measures over a population to demonstrate that the generated data has similar characteristics to the original/authentic data. For example, the system may utilize a distribution visualization called a violin plot to determine whether the synthetic data is comparable to the authentic data within an acceptable threshold. The system may also utilize other statistical validation such as a plot of a univariate correlation between each pair of variables in the original authentic electronic health record dataset and compare this against the same correlation plot in the synthetic electronic health record dataset).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify Chakraborty’s method of generating synthetic trends with Lesh’s teachings of combining the embedded vectors to generate the synthetic dataset using machine learning models such that the system can provide quality synthetic datasets that provide a threshold distance with the authentic datasets, and quickly update the synthetic dataset using the machine learning model, such that the gradient/distance in distribution between the synthetic dataset and the authentic dataset meets a threshold value (paragraph 95 of Lesh).
The combination of Chakraborty in view of Lesh does not teach storing the synthetic trends dataset in a storage device, wherein the stored synthetic trends dataset comprising the combined sequences of embedding vectors requires less storage and has less PII pertaining to the individual than a combination of the received healthcare dataset and the received supplementary dataset; training a machine learning model configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data, wherein the training data comprises the combined sequences of embedding vectors that represents the healthcare dataset and the supplementary dataset; and deploying the trained machine learning model to process input data and to output an output value that is indicative of a probability that the individual purchases a healthcare related product, wherein the output value enables privacy preserving targeted communication related to the healthcare related product.
Aptekar teaches storing the synthetic trends dataset in a storage device (column 3, lines 62-65 disclose outputting the synthetic dataset to a storage system), wherein the stored synthetic trends dataset comprising the combined sequences of embedding vectors requires less storage (column 11, lines 11-40 disclose the synthetic dataset is generated based on combined sequences/subsequence patterns of feature vector vectors. The feature vectors are embedded in a lower dimension space).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify Chakraborty’s method of generating synthetic trends in view of Lesh’s teachings of combining the embedded vectors to generate the synthetic dataset using machine learning models with Aptekar’s teachings of storing the synthetic dataset to provide improve data fidelity and data privacy by obfuscating the data in a way that preserves its statistical properties while rendering it unidentifiable to unauthorized users. This may allow researchers to more widely access and analyze such data at an earlier stage of clinical trials, which can accelerate the development of new drugs and medical treatments (column 2, lines 2-9 of Aptekar).
The combination of Chakraborty in view of Kwatra, Lesh, and Aptekar does not teach but wherein the synthetic trends dataset has less PII pertaining to the individual than a combination of the received healthcare dataset and the received supplementary dataset; training a machine learning model configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data, wherein the training data comprises the combined sequences of embedding vectors that represents the healthcare dataset and the supplementary dataset; and deploying the trained machine learning model to process input data and to output an output value that is indicative of a probability that the individual purchases a healthcare related product, wherein the output value enables privacy preserving targeted communication related to the healthcare related product.
Pauly teaches wherein the synthetic trends dataset has less PII pertaining to the individual than a combination of the received healthcare dataset and the received supplementary dataset (paragraphs 102 and 147 disclose the encoded output data/synthetic data has a dimension that is lower than the medical data. The length of a vector representing medical data of the patient may be longer than the length of a vector that is an encoded output matrix encoded from the vector of medical data by an encoding module. Paragraph 4 reveals the data can include imaging, laboratory test measurements, clinical and family history data. Family history can be supplementary dataset. Paragraph 207 reveals the encoded data set has a high degree of privacy than the raw dataset).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify Chakraborty’s method of generating synthetic trends in view of Lesh’s teachings of combining the embedded vectors to generate the synthetic dataset using machine learning models and Aptekar’s teachings of storing the synthetic dataset with Pauly’s teachings of the encoded data set having less PII such that the data can be sent to remote computing environments that have lower privacy standard without compromising the data (paragraph 207 of Pauly).
The combination of Chakraborty in view of Lesh, Aptekar, and Pauly does not teach training a machine learning model configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data, wherein the training data comprises the combined sequences of embedding vectors that represents the healthcare dataset and the supplementary dataset; and deploying the trained machine learning model to process input data and to output an output value that is indicative of a probability that the individual purchases a healthcare related product, wherein the output value enables privacy preserving targeted communication related to the healthcare related product.
Woytarowicz teaches training a machine learning model configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data (paragraph 54 discloses the training data comprises examples of values for the features. In some example embodiments, the training data comprises labeled data with examples of values for the features and labels indicating the outcome, such as a purchase, a visit to a website, adding an item to a cart, etc. The machine-learning algorithms utilize the training data to find correlations/relationships among identified features that affect the outcome. Paragraphs 63 further disclose training the ML algorithm includes labeled data which is known data for one or more identified features. Per paragraph 61, the features include features associated with the user profile (e.g., name, persona ID, email address, phone number, city of residence, etc.), and features associated with events (e.g., purchases, timestamp, website views, amount spent, time window for the events, etc.), wherein the training data comprises the combined sequences of embedding vectors that represents the healthcare dataset and the supplementary dataset in fewer dimensions than the healthcare dataset and the supplementary dataset (paragraph 70 discloses the embedding vector that represents user information (personal information such as user identifier. Paragraph 125 reveals the dataset can include age which is health data) and event information (supplementary data); deploying the trained machine learning model to process input data and to output an output value that is indicative of a probability that the individual purchases a healthcare related product (paragraph 21 discloses the propensity ML model being trained with training data comprising values for features that include event features, user information features, and audience labels; and generating, by the propensity ML model, a score for the user indicating a probability that the user will purchase the first product).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify Chakraborty’s method of generating synthetic trends in view of Lesh’s teachings of combining the embedded vectors to generate the synthetic dataset using machine learning models, Aptekar’s teachings of storing the synthetic dataset, and Pauly’s teachings of the encoded data set having less PII wherein the output value enables privacy preserving targeted communication related to the healthcare related product with Woytarowicz’s teachings of training a machine learning model to provide a system that monitor customers and consolidate customer information from multiple sources. The consolidation information may be used by machine-learning systems that predict possible purchases of customers, as well as prioritize those customers with high purchasing potential (paragraph 17 of Woytarowicz).
The combination of Chakraborty in view of Lesh, Aptekar, and Pauly does not teach wherein the output value enables privacy preserving targeted communication related to the healthcare related product.
Choudhury teaches training a machine learning model (paragraph 25 discloses training the federated ML model), deploying the trained machine learning model to output an output value that is indicative of a probability that the individual purchases a healthcare related product (paragraph 20 discloses the machine learning model may be trained to predict a user’s probability of purchasing an item); wherein the output value enables privacy preserving targeted communication related to the healthcare related product (paragraph 27 discloses applying syntactic anonymization to a federated learning system to privacy-protect data being shared among various components of the federated learning system).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify Chakraborty’s method of generating synthetic trends in view of Lesh’s teachings of combining the embedded vectors to generate the synthetic dataset using machine learning models, Aptekar’s teachings of storing the synthetic dataset, Pauly’s teachings of the encoded data set having less PII wherein the output value enables privacy preserving targeted communication related to the healthcare related product, and Woytarowicz’s teachings of training a machine learning model with Choudhury’s teaching of privacy preserving to facilitate compliance with privacy regulations that provide requirements around adequate data de-identification/data anonymization (paragraph 5 of Choudhury).
As to claim 12, the combination of Chakraborty in view of Lesh, Aptekar, Pauly, Woytarowicz, and Choudhury teaches wherein the transformation of the first sequence of embedding vectors comprises one or more of reducing a dimensionality of the first sequence of embedding vectors (Chakraborty: paragraph 56 discloses the transformation of the first embedded vector/sequence by adding noise to each tensor vector of the anonymized databases. Figure 4 and paragraphs 45-50 illustrate the noise propagation module which include input layer, a hidden layer, an anonymizer layer, and an output layer. The tensors are represented as a dimensional vector. An encoder maps each data point to a representation point h (e.g., h.sub.1, h.sub.2, or h.sub.3) (h=f(d)). The mapping of the entire data set includes extracting relevant features or weights associated with the data points and projecting the features from a first dimension (e.g., a higher dimension) to another dimension (e.g., a lower dimension). For example, a 300 dimensional vector represented as a table of token records (d.sub.1, −d.sub.4) can be mapped to a 100 dimensional vector (h.sub.1, −h.sub.3). This can be useful for principles such as data compression, which reduces dimensionality by removing static or non-important features of data. Each data point within the hidden layer is then mapped to another representation of the same or analogous dimension within the anonymizer layer. The processing for this mapping includes first using a Mondrian algorithm), a lossy compression of the first sequence of embedding vectors (Chakraborty: paragraph 56 discloses the transformation of the first embedded vector/sequence by adding noise to each tensor vector of the anonymized databases. Figure 4 and paragraphs 45-50 illustrate the noise propagation module which include input layer, a hidden layer, an anonymizer layer, and an output layer. The tensors are represented as a dimensional vector. An encoder maps each data point to a representation point h (e.g., h.sub.1, h.sub.2, or h.sub.3) (h=f(d)). The mapping of the entire data set includes extracting relevant features or weights associated with the data points and projecting the features from a first dimension (e.g., a higher dimension) to another dimension (e.g., a lower dimension). This can be useful for principles such as data compression, which reduces dimensionality by removing static or non-important features of data. Each data point within the hidden layer is then mapped to another representation of the same or analogous dimension within the anonymizer layer. The processing for this mapping includes first using a Mondrian algorithm), and adding noise to the first sequence of embedding vectors, wherein the noise comprises one or more sources of noise (Chakraborty: paragraph 56 discloses nose is added by the noise propagation module (source of noise) to the tensors which are the first sequence of embedding vectors).
As to claim 13, the combination of Chakraborty in view of Lesh, Aptekar, Pauly, Woytarowicz, and Choudhury teaches wherein the generation of the first sequence of embedding vectors for the segregated data environments comprises a principal component analysis of the corresponding dataset (Chakraborty: paragraph 27 discloses analysis of the dataset via the embedding vector module to generate the tensor vectors).
As to claim 14, the combination of Chakraborty in view of Lesh, Aptekar, Pauly, Woytarowicz, and Choudhury teaches wherein the operations further comprise generating a token from the PII of the healthcare dataset and a token from the PII of the supplementary dataset, wherein the token are operative to link corresponding datasets to data stored outside of the federated data cleanroom (Chakraborty: paragraphs 30, 33, and 35 disclose the generated data tensors are representations of data in a different form, such as a token representing the dataset and are linked/corresponds to the real data values in the data records (health and supplementary as disclosed in claim 1above) ).
As to claim 15, the combination of Chakraborty in view of Lesh, Aptekar, Pauly, Woytarowicz, and Choudhury teaches the operation further comprising:
determining a utility of the synthetic trends dataset, wherein the utility is indicative of a quality of the synthetic trends dataset with respect to a particular task (Lesh: paragraphs 80-83 discloses computing distance metrics that compare the real records to the synthetic records. After comparing, if any of the pairs have a distance value that is too low or within a threshold distance, the system may highlight that pair for manual review by a human to determine if any authentic patient medical data has leaked into the generated synthetic dataset. The synthetic dataset may be assumed validated if a certain threshold of confidence between a relevant metric of predictive performance is achieved. See also paragraphs 94-95 regarding determining first and second gradients to generate synthetic data records);
determining that the utility of the synthetic trends dataset is below a utility threshold that represents a minimum required quality of insights generated based on analytics of the synthetic trends dataset (Lesh: paragraphs 80-83 discloses computing distance metrics that compare the real records to the synthetic records. After comparing, if any of the pairs have a distance value that is too low or within a threshold distance, the system may highlight that pair for manual review by a human to determine if any authentic patient medical data has leaked into the generated synthetic dataset. The synthetic dataset may be assumed validated if a certain threshold of confidence between a relevant metric of predictive performance is achieved );
modifying, based on the utility of the synthetic trends dataset being below the utility threshold, one or more of the determined data features to increase the utility of the synthetic trends dataset (Lesh: paragraphs 80-83 discloses computing distance metrics that compare the real records to the synthetic records. After comparing, if any of the pairs have a distance value that is too low or within a threshold distance, the system may highlight that pair for manual review by a human to determine if any authentic patient medical data has leaked into the generated synthetic dataset. The synthetic dataset may be assumed validated if a certain threshold of confidence between a relevant metric of predictive performance is achieved. Paragraphs 97-95 disclose updating first and second gradients to meet threshold loss distribution that contributes to the generation of the synthetic data records. Paragraphs 5 and 46-47, 51 disclose adding noise to modify and control the generation of the synthetic dataset); and
after the modifying, outputting insights generated based on analytics of the synthetic trends dataset (Woytarowicz: paragraph 21 discloses generating, by the propensity ML model, a score for the user indicating a probability that the user will purchase the first product). Motivation is similar to the motivation presented in claim 11.
As to claim 17, the combination of Chakraborty in view of Lesh, Aptekar, Pauly, Woytarowicz, and Choudhury teaches wherein generating the first sequence of embedding vectors comprises capturing a variance of the anonymized data in fewer dimensions than the dimensionality of the anonymized data (Chakraborty: paragraph 33 discloses in generating the tensors, the numerical data is normalized. Normalization captures the variance/difference of a anonymized data value with the mean. The mean being fewer dimensionality than the anonymized dataset). For word data paragraphs 39-40 reveals equation that yields the normalized probabilistic model for language modeling).
As to claim 18, the combination of Chakraborty in view of Lesh, Aptekar, Pauly, Woytarowicz, and Choudhury wherein the healthcare dataset comprises a plurality of alphanumeric codes, wherein the alphanumeric codes are mapped to embedding vectors (Chakraborty: paragraphs 32-34 discloses that the data records includes various fields or attributes with both real numbers and strings. These fields are linked to tensors/embedded vectors. Paragraph 48 disclose the output of the Mondrian algorithm is a set of equivalence classes. The equivalence class members are codes (derived from the data vectors using the encoder)).
As to claim 19, Chakraborty teaches a non-transitory computer-readable medium (Figure 13, reference number 28 “Memory”) storing one or more instructions executable by a computer system to perform operations (paragraph 112 reveals memory may include at least one program product having a set (e.g., at least one) of program modules that are configured to carry out the functions of embodiments of the invention ) for generating a synthetic trends dataset (abstract discloses generating anonymized synthetic data by adding noise to a first set of data from one or more data stores. Paragraph 78 reveals the data can come from private databases. Claim 19 discloses aggregating/combining the plurality of databases (see Figure 10, reference number 1006). Paragraphs 77-78 reveals the data record are health records from hospital private database. Paragraphs 77-78, 32, 52, and 68 disclose additional data records from private databases that include secondary information such as education level, zip code, and/or salary data) for training a machine learning model (paragraphs 31, 60-62 and steps 608 and 612 reveal the output candidate for anonymization is retrained in the embedding vector model. The data record strings is sent through the word embedding vector model to obtain a candidate output for anonymization/synthetic data), the operations comprising:
receiving a healthcare dataset from a database in a first segregated data environment of a federated data cleanroom, the healthcare dataset comprising personally identifiable information (PII) pertaining to the individual (paragraphs 4 and 52 reveal receiving raw data/first data associated with a data stores. The raw data comes from database records, unstructured data, and/or any other set of data. Paragraphs 23 and 68 discloses the datastore record can include information such as disease. Paragraphs 77- 78 reveal the data can come from private databases such as health data from hospital; therefore, the databases are federated. See also Figure 9, reference number 910 of Figure 10, reference number 1008);
receiving a supplementary dataset from a database in a second segregated data environment of the federated data cleanroom, the supplementary data comprising PII pertaining to the individual (paragraphs 4 and 52 reveal receiving raw data/first data associated with a data stores. The raw data comes from database records, unstructured data, and/or any other set of data. Paragraphs 32, 52, and 68 disclose the datastore record can include information such as education level, zip code, and/or salary. The data can also include name of a person, SSN. Paragraphs 77- 78 reveal the data can come from private databases; therefore, the databases are federated. See also Figure 9, reference number 912 or Figure 10, reference number 1010);
anonymizing the data stored in the databases, wherein the anonymized data from the databases is stored in the corresponding segregated databases (paragraph 52 discloses anonymizing received data values by deleting, masking, changing, and/or encrypting the dataset such that the data are not discernable to a human reader. Paragraph 79 further discloses each of the private databases are anonymized to a corresponding anonymized database. See also Figure 9, reference numbers 904, 906, and 908);
generating, for the segregated data environments, a numerical representation of the data features of a plurality of data features of the anonymized data stored in the corresponding segregated data environment, wherein the numerical representation comprises a first sequence of embedding vectors (paragraphs 30 and 53-54 disclose for the plurality of real strings and real number values in the datastores of the anonymized datastore, identifying quasi-identifiers/data features and sensitive identifier values for both string data and numeral data; normalizing the data (if the data is a numerical dataset) and generating one or more tensors values for the strings data and the normalized numerical data. Tensors are numerical representation of the data record token that are fed through an embedding vector model (for the word ) or normalize vector. Vector embedding is converting words/numbers into number sequence), wherein the first sequence of embedding vectors for a segregated environment is a compressed representation of the corresponding data feature (paragraphs 30-31 disclose the data tensor are representations of data in different form than the data itself, such as a token representing the data. Text strings that are fed through the word embedding vector model, the output is the data tensors which may be represented as integers or other real numbers. The data tensor are mapped from its high dimensionality space to a lower dimensionality space, the data is compressed);
determining, for the segregated data environments, a second sequence of embedding vectors, wherein the second sequence of embedding vectors is a transformation of the corresponding first sequence of embedding vectors, the transformation comprising a reduction of information from the first sequence of embedding vectors (paragraph 56 discloses adding noise to each tensor vector of the anonymized databases. Figure 4 and paragraphs 45-50 illustrate the noise propagation module which include input layer, a hidden layer, an anonymizer layer, and an output layer. The tensors are represented as a dimensional vector. An encoder maps each data point to a representation point h (e.g., h.sub.1, h.sub.2, or h.sub.3) (h=f(d)). The mapping of the entire data set includes extracting relevant features or weights associated with the data points and projecting the features from a first dimension (e.g., a higher dimension) to another dimension (e.g., a lower dimension). For example, a 300 dimensional vector represented as a table of token records (d.sub.1, −d.sub.4) can be mapped to a 100 dimensional vector (h.sub.1, −h.sub.3). This can be useful for principles such as data compression, which reduces dimensionality by removing static or non-important features of data. Each data point within the hidden layer is then mapped to another representation of the same or analogous dimension within the anonymizer layer. The processing for this mapping includes first using a Mondrian algorithm);
generating the synthetic trends dataset (paragraph 48 discloses the output of the Mondrian algorithm is a set of equivalence classes, where each equivalence class is of size at least k to meet k-anonymity. The equivalence class members at this point can be codes (derived from the data vectors using the encoder). Accordingly, at the output of the anonymizer layer (e.g., the anonymizer of FIG. 1), the c1, c2, c3, and c4 are indistinguishable. They are all represented by noisy code C. In some embodiments, different equivalence classes have different centroids and thus different outputs. The output layer is the synthetic trends dataset that comprises the second sequence of embedding vectors/noise associated with the anonymized databases);
training, by a processor configured to access training data comprising the ….synthetic dataset, a machine learning model …using the training data (Figure 1 and paragraphs 31 and 62 disclose the word embedding vector model(s) are retrained using the anonymized data. After the decoder turns the data into the anonymized data, the word embedding vector model(s) is retrained, which is described in more detail below. This is indicated by the arrow from the decoder to the word embedding vector model(s)), wherein the training data comprises the combined sequences of embedding vectors that represents [the datasets] in fewer dimensions than the healthcare dataset and the supplementary dataset (paragraph 37 discloses the training data includes the word embedding vector model. Paragraph 56 discloses adding noise to each tensor vector of the anonymized databases. Figure 4 and paragraphs 45-50 illustrate the noise propagation module which include input layer, a hidden layer, an anonymizer layer, and an output layer. The tensors are represented as a dimensional vector. An encoder maps each data point to a representation point h (e.g., h.sub.1, h.sub.2, or h.sub.3) (h=f(d)). The mapping of the entire data set includes extracting relevant features or weights associated with the data points and projecting the features from a first dimension (e.g., a higher dimension) to another dimension (e.g., a lower dimension). For example, a 300 dimensional vector represented as a table of token records (d.sub.1, −d.sub.4) can be mapped to a 100 dimensional vector (h.sub.1, −h.sub.3). This can be useful for principles such as data compression, which reduces dimensionality by removing static or non-important features of data. Each data point within the hidden layer is then mapped to another representation of the same or analogous dimension within the anonymizer layer. The processing for this mapping includes first using a Mondrian algorithm; and
…the trained machine learning model to process input data indicative of healthcare information and consumer information related to the individual (Figures 1 and 6 and paragraphs 31, 58, and 62 reveal the retrain embedded vector model processes data record strings/input data. Paragraph 78 reveals the data can come from private databases. Claim 19 discloses aggregating/combining the plurality of databases (see Figure 10, reference number 1006). Paragraphs 77-78 reveals the data record are health records from hospital private database. Paragraphs 77-78, 32, 52, and 68 disclose additional data records from private databases that include secondary information such as education level, zip code, and/or salary data).
Chakraborty does not teach generating, in a shared data environment, the synthetic trends dataset by combining the second sequence of embedding vectors associated with the segregated data environments; storing the synthetic trends dataset in a storage device, wherein the stored synthetic trends dataset comprising the combined sequences of embedding vectors requires less storage and has less PII pertaining to the individual than a combination of the received healthcare dataset and the received supplementary dataset; training a machine learning model configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data, wherein the training data comprises the combined sequences of embedding vectors that represents the healthcare dataset and the supplementary dataset; and deploying the trained machine learning model to process input data and to output an output value that is indicative of a probability that the individual purchases a healthcare related product, wherein the output value enables privacy preserving targeted communication related to the healthcare related product.
Lesh teaches generating, in a shared data environment, the synthetic trends dataset by combining the second sequence of embedding vectors associated with the segregated data environments (paragraphs 64-65 and Figure 5B disclose the medical records (from databases (see paragraph 48) for a patient is transformed and for each record, a numerical vector/embedded vector is generated. The medical records are transformed into concatenated vector using the embedding vectors and are inputted into a generative model (shared data environment, per paragraph 48 the synthetic data is generated from a single or multiple source database ) to generate synthetic dataset), wherein the synthetic trends dataset has [threshold metric difference] with the healthcare dataset and the supplementary dataset (paragraph 66 further discloses a discriminator can determine a distribution over the state of the same based on the received input of the synthetic dataset and the authentic medical datasets (which can be the combined healthcare dataset and the dosage dataset), detecting whether the sample is synthetic. Paragraphs 80-83 also discloses computing distance metrics that compare the real records to the synthetic records. After comparing, if any of the pairs have a distance value that is too low or within a threshold distance, the system may highlight that pair for manual review by a human to determine if any authentic patient medical data has leaked into the generated synthetic dataset. The synthetic dataset may be assumed validated if a certain threshold of confidence between a relevant metric of predictive performance is achieved. For statistical validation, the generative system may utilize statistical measures over a population to demonstrate that the generated data has similar characteristics to the original/authentic data. For example, the system may utilize a distribution visualization called a violin plot to determine whether the synthetic data is comparable to the authentic data within an acceptable threshold. The system may also utilize other statistical validation such as a plot of a univariate correlation between each pair of variables in the original authentic electronic health record dataset and compare this against the same correlation plot in the synthetic electronic health record dataset).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify Chakraborty’s method of generating synthetic trends with Lesh’s teachings of combining the embedded vectors to generate the synthetic dataset using machine learning models such that the system can provide quality synthetic datasets that provide a threshold distance with the authentic datasets, and quickly update the synthetic dataset using the machine learning model, such that the gradient/distance in distribution between the synthetic dataset and the authentic dataset meets a threshold value (paragraph 95 of Lesh).
The combination of Chakraborty in view of Lesh does not teach storing the synthetic trends dataset in a storage device, wherein the stored synthetic trends dataset comprising the combined sequences of embedding vectors requires less storage and has less PII pertaining to the individual than a combination of the received healthcare dataset and the received supplementary dataset; training a machine learning model configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data, wherein the training data comprises the combined sequences of embedding vectors that represents the healthcare dataset and the supplementary dataset; and deploying the trained machine learning model to process input data and to output an output value that is indicative of a probability that the individual purchases a healthcare related product, wherein the output value enables privacy preserving targeted communication related to the healthcare related product.
Aptekar teaches storing the synthetic trends dataset in a storage device (column 3, lines 62-65 disclose outputting the synthetic dataset to a storage system), wherein the stored synthetic trends dataset comprising the combined sequences of embedding vectors requires less storage (column 11, lines 11-40 disclose the synthetic dataset is generated based on combined sequences/subsequence patterns of feature vector vectors. The feature vectors are embedded in a lower dimension space).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify Chakraborty’s method of generating synthetic trends in view of Lesh’s teachings of combining the embedded vectors to generate the synthetic dataset using machine learning models with Aptekar’s teachings of storing the synthetic dataset to provide improve data fidelity and data privacy by obfuscating the data in a way that preserves its statistical properties while rendering it unidentifiable to unauthorized users. This may allow researchers to more widely access and analyze such data at an earlier stage of clinical trials, which can accelerate the development of new drugs and medical treatments (column 2, lines 2-9 of Aptekar).
The combination of Chakraborty in view of Kwatra, Lesh, and Aptekar does not teach but wherein the synthetic trends dataset has less PII pertaining to the individual than a combination of the received healthcare dataset and the received supplementary dataset; training a machine learning model configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data, wherein the training data comprises the combined sequences of embedding vectors that represents the healthcare dataset and the supplementary dataset; and deploying the trained machine learning model to process input data and to output an output value that is indicative of a probability that the individual purchases a healthcare related product, wherein the output value enables privacy preserving targeted communication related to the healthcare related product.
Pauly teaches wherein the synthetic trends dataset has less PII pertaining to the individual than a combination of the received healthcare dataset and the received supplementary dataset (paragraphs 102 and 147 disclose the encoded output data/synthetic data has a dimension that is lower than the medical data. The length of a vector representing medical data of the patient may be longer than the length of a vector that is an encoded output matrix encoded from the vector of medical data by an encoding module. Paragraph 4 reveals the data can include imaging, laboratory test measurements, clinical and family history data. Family history can be supplementary dataset. Paragraph 207 reveals the encoded data set has a high degree of privacy than the raw dataset).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify Chakraborty’s method of generating synthetic trends in view of Lesh’s teachings of combining the embedded vectors to generate the synthetic dataset using machine learning models and Aptekar’s teachings of storing the synthetic dataset with Pauly’s teachings of the encoded data set having less PII such that the data can be sent to remote computing environments that have lower privacy standard without compromising the data (paragraph 207 of Pauly).
The combination of Chakraborty in view of Lesh, Aptekar, and Pauly does not teach training a machine learning model configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data, wherein the training data comprises the combined sequences of embedding vectors that represents the healthcare dataset and the supplementary dataset; and deploying the trained machine learning model to process input data and to output an output value that is indicative of a probability that the individual purchases a healthcare related product, wherein the output value enables privacy preserving targeted communication related to the healthcare related product.
Woytarowicz teaches training a machine learning model configured to learn relationships between features and target variables of interest as defined by an audience definition using the training data (paragraph 54 discloses the training data comprises examples of values for the features. In some example embodiments, the training data comprises labeled data with examples of values for the features and labels indicating the outcome, such as a purchase, a visit to a website, adding an item to a cart, etc. The machine-learning algorithms utilize the training data to find correlations/relationships among identified features that affect the outcome. Paragraphs 63 further disclose training the ML algorithm includes labeled data which is known data for one or more identified features. Per paragraph 61, the features include features associated with the user profile (e.g., name, persona ID, email address, phone number, city of residence, etc.), and features associated with events (e.g., purchases, timestamp, website views, amount spent, time window for the events, etc.), wherein the training data comprises the combined sequences of embedding vectors that represents the healthcare dataset and the supplementary dataset in fewer dimensions than the healthcare dataset and the supplementary dataset (paragraph 70 discloses the embedding vector that represents user information (personal information such as user identifier. Paragraph 125 reveals the dataset can include age which is health data) and event information (supplementary data); deploying the trained machine learning model to process input data and to output an output value that is indicative of a probability that the individual purchases a healthcare related product (paragraph 21 discloses the propensity ML model being trained with training data comprising values for features that include event features, user information features, and audience labels; and generating, by the propensity ML model, a score for the user indicating a probability that the user will purchase the first product).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify Chakraborty’s method of generating synthetic trends in view of Lesh’s teachings of combining the embedded vectors to generate the synthetic dataset using machine learning models, Aptekar’s teachings of storing the synthetic dataset, and Pauly’s teachings of the encoded data set having less PII wherein the output value enables privacy preserving targeted communication related to the healthcare related product with Woytarowicz’s teachings of training a machine learning model to provide a system that monitor customers and consolidate customer information from multiple sources. The consolidation information may be used by machine-learning systems that predict possible purchases of customers, as well as prioritize those customers with high purchasing potential (paragraph 17 of Woytarowicz).
The combination of Chakraborty in view of Lesh, Aptekar, and Pauly does not teach wherein the output value enables privacy preserving targeted communication related to the healthcare related product.
Choudhury teaches training a machine learning model (paragraph 25 discloses training the federated ML model), deploying the trained machine learning model to output an output value that is indicative of a probability that the individual purchases a healthcare related product (paragraph 20 discloses the machine learning model may be trained to predict a user’s probability of purchasing an item); wherein the output value enables privacy preserving targeted communication related to the healthcare related product (paragraph 27 discloses applying syntactic anonymization to a federated learning system to privacy-protect data being shared among various components of the federated learning system).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify Chakraborty’s method of generating synthetic trends in view of Lesh’s teachings of combining the embedded vectors to generate the synthetic dataset using machine learning models, Aptekar’s teachings of storing the synthetic dataset, Pauly’s teachings of the encoded data set having less PII wherein the output value enables privacy preserving targeted communication related to the healthcare related product, and Woytarowicz’s teachings of training a machine learning model with Choudhury’s teaching of privacy preserving to facilitate compliance with privacy regulations that provide requirements around adequate data de-identification/data anonymization (paragraph 5 of Choudhury).
As to claim 20, the combination of Chakraborty in view of Lesh, Aptekar, Pauly, Woytarowicz, and Choudhury teaches wherein the transformation of the first sequence of embedding vectors comprises one or more of reducing a dimensionality of the first sequence of embedding vectors (Chakraborty: paragraph 56 discloses the transformation of the first embedded vector/sequence by adding noise to each tensor vector of the anonymized databases. Figure 4 and paragraphs 45-50 illustrate the noise propagation module which include input layer, a hidden layer, an anonymizer layer, and an output layer. The tensors are represented as a dimensional vector. An encoder maps each data point to a representation point h (e.g., h.sub.1, h.sub.2, or h.sub.3) (h=f(d)). The mapping of the entire data set includes extracting relevant features or weights associated with the data points and projecting the features from a first dimension (e.g., a higher dimension) to another dimension (e.g., a lower dimension). For example, a 300 dimensional vector represented as a table of token records (d.sub.1, −d.sub.4) can be mapped to a 100 dimensional vector (h.sub.1, −h.sub.3). This can be useful for principles such as data compression, which reduces dimensionality by removing static or non-important features of data. Each data point within the hidden layer is then mapped to another representation of the same or analogous dimension within the anonymizer layer. The processing for this mapping includes first using a Mondrian algorithm), a lossy compression of the first sequence of embedding vectors (Chakraborty: paragraph 56 discloses the transformation of the first embedded vector/sequence by adding noise to each tensor vector of the anonymized databases. Figure 4 and paragraphs 45-50 illustrate the noise propagation module which include input layer, a hidden layer, an anonymizer layer, and an output layer. The tensors are represented as a dimensional vector. An encoder maps each data point to a representation point h (e.g., h.sub.1, h.sub.2, or h.sub.3) (h=f(d)). The mapping of the entire data set includes extracting relevant features or weights associated with the data points and projecting the features from a first dimension (e.g., a higher dimension) to another dimension (e.g., a lower dimension). This can be useful for principles such as data compression, which reduces dimensionality by removing static or non-important features of data. Each data point within the hidden layer is then mapped to another representation of the same or analogous dimension within the anonymizer layer. The processing for this mapping includes first using a Mondrian algorithm), and adding noise to the first sequence of embedding vectors, wherein the noise comprises one or more sources of noise (Chakraborty: paragraph 56 discloses nose is added by the noise propagation module (source of noise) to the tensors which are the first sequence of embedding vectors).
As to claim 21, the combination of Chakraborty in view of Lesh, Aptekar, Pauly, Woytarowicz, and Choudhury teaches wherein deploying the trained machine learning model comprises deploying to a production environment configured to receive the input data and generate the output value in real-time (Woytarowicz: paragraph 21 discloses the propensity ML model being trained with training data comprising values for features that include event features, user information features, and audience labels; and generating, by the propensity ML model, a score for the user indicating a probability that the user will purchase the first product within a predetermined time window/real time). The motivation is the same motivation as in claim 1.
Claim(s) 22 is/are rejected under 35 U.S.C. 103 as being unpatentable over Chakraborty et al US 20190156061 (hereinafter Chakraborty), in further view of Lesh et al US 20230010686 (hereinafter Lesh), in further view of Aptekar et al US 11977550 (hereinafter Aptekar), in further view of Pauly US 20200035365 (hereinafter Pauly), in further view of Woytarowicz et al US 20230298058 (hereinafter Woytarowicz), in further view of Choudhury et al US 20210150269 (hereinafter Choudhury), and in further view of Beyda et al US 9928522 (hereinafter Beyda).
As to claim 22, the combination of Chakraborty in view of Lesh, Aptekar, Pauly, Woytarowicz, and Choudhury teaches the limitations recited in claim 1 above. The combination of Chakraborty in view of Lesh, Aptekar, Pauly, Woytarowicz, and Choudhury does not teach, but Beyda teaches wherein the output value is associated with an online identifier linked to the individual, and wherein the online identifier enables delivery of the targeted communications (claim 17 disclose the concept of accessing the profile data (column 3, lines 59-61 reveal the profile data is obtain from a model) for the client computer in response to identification of the unique identifier for the client computer determining that the client computer is a member of a network segment that includes client computers having a set of attributes corresponding to profile data collected from the plurality of audience member information collection domains, wherein the network segment is affiliated with the network domain; transmitting a segment-targeting cookie, which includes a segment identifier for the network segment of client computers, to the client computer; and accommodating the delivery of an advertisement to the client computer based upon the segment identifier in the segment-targeting cookie and using performance criteria configured to benefit a publisher of content in which the advertisement is placed).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify Chakraborty’s method of generating synthetic trends in view of Lesh’s teachings of combining the embedded vectors to generate the synthetic dataset using machine learning models, Aptekar’s teachings of storing the synthetic dataset, Pauly’s teachings of the encoded data set having less PII wherein the output value enables privacy preserving targeted communication related to the healthcare related product, and Woytarowicz’s teachings of training a machine learning model and Choudhury’s teaching of privacy preserving with Beyda’s teachings of online identifier enables delivery of the targeted communications to lead targeted marketing that is more focused and provide robust interaction with the consumer (column 1, lines 33-36 of Beyda).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure: Shmunis et al US 20130057925 (hereinafter Shmunis).
Shmunis teaches wherein the output value is associated with an online identifier linked to the individual, and wherein the online identifier enables delivery of the targeted communications (paragraph 25 discloses the concept of providing delivery over communication in response to online identifier. Paragraph 18 reveals the identifier is associated with an endpoint/user).
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 FELICIA FARROW whose telephone number is (571)272-1856. The examiner can normally be reached M - F 7:30am-4:00pm (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, Alexander Lagor can be reached at (571)270-5143. 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.
/F.F/ Examiner, Art Unit 2437
/BENJAMIN E LANIER/ Primary Examiner, Art Unit 2437
