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 .
DETAILED ACTION
Response to Amendment
The present Office Action is in response to the Request for Continued Examination dated 03/05/2026.
In the amendment dated 03/05/2026, the following occurred: Claims 1, 3, 7, 8, 11, 13, 16, 17, 20 and 21 have been amended. Claims 4, 14, 19 and 23 were canceled. Claim 24 is new.
Claims 1-3, 5-13, 15-18, 20-22 and 24 are currently pending.
Request for Continued Examination
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 03/05/2026 has been entered.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-3, 5-13, 15-18, 20-22 and 24 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more.
Claims 1, 11 and 20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1
The claim recites a method, a system and a non-transitory computer-readable medium for optimizing a sample size of an investigation, which are within a statutory category.
Step 2A1
Regarding claims 1, 11 and 20, the limitation of (claim 1 being representative) accessing a plurality of datasets representing a population stored, wherein the plurality of datasets include a plurality of attributes associated with a plurality of entities comprising the population; applying a […] to the plurality of datasets to identify a target population, wherein (i) the […] is trained to learn associations between the plurality of attributes and entity eligibility for the target population, and (ii) the target population is a subset of the population including (a) a first subset of the plurality of datasets wherein each dataset of the first subset comprises, as one or more of the plurality of attributes, an indicator explicitly representing one or more conditions based on one or more deterministic criteria associated with the entity eligibility, and (b) a second subset of the plurality of datasets, wherein each dataset of the second subset excludes the indicator and comprises, as one or more of the plurality of attributes, data implicitly representing the one or more conditions based on the one or more deterministic criteria; merging the first subset and the second subset of the plurality of datasets into a first data object representing the target population; receiving a user input indicative of a plurality of parameters that correspond to values contained in the first data object, wherein the plurality of parameters exhibit variation within the target population; generating a plurality of categories corresponding to a plurality of unique combinations of options identified across one or more of the plurality of parameters selected to represent the variation; determining a first distribution across the plurality of categories within the target population; generating a plurality of first similarity-based subsets by applying one or more sampling techniques to the first data object across the one or more of the plurality of parameters, wherein the plurality of first similarity-based subsets are associated with a plurality of deviation measures determined based on the first distribution and a second distribution determined across the plurality of categories within the plurality of first similarity-based subsets, the plurality of deviation measures indicative of a representativeness of the plurality of first similarity-based subsets relative to the target population across the plurality of categories; generating a second data object comprising one of the plurality of first similarity-based subsets associated with a lowest deviation measure of the plurality of deviation measures indicative of a most accurate representation of the target population among the plurality of first similarity-based subset; providing the second data object for display in response to the user input; generating based on feedback associated with the plurality of deviation measures determined for the plurality of first similarity- based subsets, a plurality of second similarity-based subsets by adjusting the one or more sampling techniques and applying the adjusted one or more sampling techniques to the first data object across the one or more plurality of parameters; receiving performance data indicative of one or more of (i) whether the target population identified is a true eligible population of the population or (ii) whether the plurality of categories generated are representative of the true eligible population: and updating, based on the performance data, one or more of the […] or the […], wherein the updating includes adjusting one or more layers or one or more weights of the one or more of the [..] or the […] to increase an accuracy of one or more of a next target population identified or a next plurality of categories generated, respectively, for future sampling as drafted, is a process that, under the broadest reasonable interpretation, covers certain methods of organizing human activity (i.e., managing personal behavior including following rules or instructions) but for recitation of generic computer components. The claims encompass a series of rules or instructions for a person or persons to follow, with or without the aid of a computer, to access a plurality of datasets, apply a first machine learning model, merge the first and second subset, receive a user input, generate a plurality of categories, determine a first distribution, generate a plurality of similarity-based subsets, generate a second data object, provide the second data object, generate a plurality of second similarity-based subsets, receive performance data and update one or more of the first or second machine learning model in the manner described in the identified abstract idea, supra. The rules or instructions are the claimed steps of “accessing…applying…merging… receiving… generating… determining…generating…generating…providing… generating…receiving…and updating” as indicated supra.
Other than reciting generic computer components (discussed infra), i.e., (in claim 1) one or more processors, (in claim 11) one or more processors and one or more non-transitory computer-readable media, and (in claim 20) one or more non-transitory computer-readable media and one or more processors, the claimed invention amounts to managing personal behavior or interaction between people. If a claim limitation, under its broadest reasonable interpretation, covers managing personal behavior or interactions between people but for the recitation of generic computer components, then it falls within the “certain methods of organizing human activity” grouping of abstract ideas. Accordingly, the claim recites an abstract idea.
The claim further recites “a first machine learning model that is trained and a second machine learning model” and “adjusting one or more layers or one or more weights of the one or more of the first machine learning model or the second machine learning model.” When given their broadest reasonable interpretation in light of the disclosure, the first machine model which is applied to identify a target population and trained to learn associations, the second machine learning model that is used to generate a plurality of categories and adjusting one or more layers or one or more weights of the one or more of the first machine learning model or the second machine learning model using decision trees, neural networks, or support vector machines, linear regression, random forest, gradient boosted machine (GBM), deep learning, and/or a deep neural network, K-means clustering or K-Nearest Neighbors, Fully Convolutional Networks (FCN) and Recurrent Neural Networks (RCN), probabilistic models such as Bayesian Networks, and/or discriminative models such as Decision Forests and maximum margin methods described in the Spec. at para. 0034, 0035, 0055 and 00110, represents the creation of mathematical interrelationships between data. As such, the first machine model which is applied to identify a target population and trained to learn associations, the second machine learning model that is used to generate a plurality of categories and adjusting one or more layers or one or more weights of the one or more of the first machine learning model or the second machine learning model represents a mathematical concept that is interpreted to be part of the identified abstract idea, supra. The types of identified abstract ideas are considered together as a single abstract idea for analysis purposes.
Step 2A2
This judicial exception is not integrated into a practical application. In particular, claim 1 recites the additional element of one or more processors. Claims 11 and 20 recite the additional elements of one or more non-transitory computer-readable media and one or more processors. These additional elements are not exclusively defined by the applicant and are recited at a high-level of generality (i.e., a generic computer components for enabling access to medical information or for performing generic computer functions. See Spec at Para. [00112], [00116], [00117], [00124] and [00125]) such that they amounts to no more than mere instructions to apply the exception using a generic computer component. As set forth in MPEP 2106.04(d) “merely including instructions to implement an abstract idea on a computer” is an example of when an abstract idea has not been integrated into a practical application. Accordingly, even in combination, these additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. The claim is directed to an abstract idea.
Claims 1, 11 and 20 also recite the additional element of a database and an interactive interface. The additional element of a database is recited at a high level of generality (i.e. a general means to store data) and amount to extra solution activity. MPEP 2106.04(d)(I) indicates that extra-solution data gathering activity cannot provide a practical application. The additional element of an interactive interface merely generally links the abstract idea to a particular technological environment or field of use. MPEP 2106.04(d)(I) indicates that generally linking an abstract idea to a particular technological environment or field of use cannot provide a practical application. Accordingly, even in combination, these additional elements do not integrate the abstract idea into a practical application.
The claims further recite the additional elements of a first machine model which is applied to identify a target population and trained to learn associations, a second machine learning model that is used to generate a plurality of categories and adjusting one or more layers or one or more weights of the one or more of the first machine learning model or the second machine learning model. As noted above, there represent a mathematical concept as described in the Specification at Para. 0034, 0035, 0055 and 00110. This mathematical concept is applied to (“apply it’) the abstract idea. MPEP 2106.04(d)(I) indicates that merely saying “apply it” or equivalent to the abstract idea cannot provide a practical application.
Step 2B
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional element of the one or more processors and one or more non-transitory computer-readable media to perform the noted steps amounts to no more than mere instructions to apply the exception using a generic computer component. Mere instructions to apply an exception using a generic computer component cannot provide an inventive concept (“significantly more”). Moreover, using generic computer components to perform abstract ideas does not provide a necessary inventive concept. See Alice, 573 U.S. at 223 (“mere recitation of a generic computer cannot transform a patent-ineligible abstract idea into a patent-eligible invention”). Therefore, whether considered alone or in combination, the additional elements do not amount to significantly more than the abstract idea.
As discussed above with respect to integration of the abstract idea into a practical application, the additional element of the database was considered extra-solution activity. This has been re-evaluated under “significantly more” analysis and determined to be well-understood, routine and conventional activity in the field, see Knuff (US 2022/0188654) at [0113], [0116] and [01531] and see Baldauf-Lenschen (US 2023/0146840) at [0031] and [0078]. The additional element of the interactive interface was determined to generally link the abstract idea to a particular technological environment or field of use. This has also been re-evaluated under the “significantly more” analysis and has also been found insufficient to provide significantly more. MPEP 2106.05(A) indicates that generally linking an abstract idea to a particular technological environment or field of use cannot provide significantly more. Well-understood, routine and conventional activity cannot provide an inventive concept (“significantly more”). Therefore when considering the additional elements alone, and in combination, there is no inventive concept in the claim, and thus the claim is not patent eligible.
Also, as discussed above with respect to integration of the abstract idea into a practical application, the additional elements of a first machine model which is applied to identify a target population and trained to learn associations, a second machine learning model that is used to generate a plurality of categories and adjusting one or more layers or one or more weights of the one or more of the first machine learning model or the second machine learning model were determined to be the application of mathematical concepts to the identified abstract idea. This has been re-evaluated under the “significantly more” analysis and has also been found insufficient to provide significantly more. MPEP2106.05(1)(A) indicates that merely saying “apply it’ or equivalent to the abstract idea cannot provide an inventive concept (“significantly more’). As such the claim is not patent eligible.
The examiner notes that: A well-known, general-purpose computer has been determined by the courts to be a well-understood, routine and conventional element (see, e.g., Alice Corp. v. CLS Bank; see also MPEP 2106.05(d)); Receiving and/or transmitting data over a network (“a communications network”) has also been recognized by the courts as a well - understood, routine and conventional function (see, e.g., buySAFE v. Google; MPEP 2016(d)(II)); and Performing repetitive calculations is/are also well-understood, routine and conventional computer functions when they are claimed in a merely generic manner (see, e.g., Parker v. Flook; MPEP 2016.05(d)).
Claims 2-3, 5-10, 12-13, 15-18, 21-22 and 24 are similarly rejected because they either further define/narrow the abstract idea and/or do not further limit the claim to a practical application or provide as inventive concept such that the claims are subject matter eligible even when considered individually or as an ordered combination. Claim(s) 2 and 12 further merely describe(s) the second data object. Claim(s) 3 and 13 further merely describe(s) the first subset and the second subset. Claim(s) 5 further merely describe(s) determining the deviation measure. Claim(s) 6 and 15 further merely describe(s) what determining a deviation measure for each of the plurality of similarity-based subsets includes. Claim(s) 7 and 16 further merely describe(s) generating a plurality of member size categories and assigning each similarity-based subset of the plurality of similarity-based subsets to a member size category. Claim(s) 8 and 17 further merely describe(s) determining a similarity-based subset with the lowest deviation measure. Claim(s) 9 and 18 further merely describe(s) the second data object. Claim(s) 10 further merely describe(s) the plurality of parameters, what providing the second data object for display includes, determining satisfactory member size categories and providing a visual indicia. Claim(s) 21 further merely describe(s) adjusting the one or more sampling techniques and applying the adjusted one or more sampling techniques. Claim(s) 22 further merely describe(s) determining the first and second distribution. Claim(s) 24 further merely describe(s) the second machine learning model performs feature selection and automatically selects the one or more of the plurality of parameters. As can be seen, claims 2-3, 5-10, 12-13, 15-18, 21-22 and 24 further define the abstract idea and are rejected for the same reason presented above with respect to claims 1, 11 and 20.
Claim Rejections - 35 USC § 102
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claims 1-3, 5-13, 15-18, 20-22 and 24 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Knuff (US 2022/0188654).
REGARDING CLAIM 1
Knuff discloses a computer-implemented method comprising: accessing, by one or more processors, a plurality of datasets representing a population stored in a database, wherein the plurality of datasets include a plurality of attributes associated with a plurality of entities comprising a population ([0019] teaches the use of a processor. [0018] teaches receiving data packets and [0077] teaches access to all data including biomarkers and self-reported AEs, [0086] teaches the system may receive as input a list of biomarkers such as chronic heart failure, Caucasian (a subset of population), and cholesterol and [0116] also teaches searching a large database of clinical trials (interpreted by Examiner as means for accessing a plurality of datasets representing a population stored in a database, wherein the plurality of datasets include a plurality of attributes associated with a plurality of entities comprising a population)); applying, by the one or more processors, a first machine learning model to the plurality of datasets to identify a target population, wherein (i) the first machine learning model is trained to learn associations between the plurality of attributes and entity eligibility for the target population, and (ii) the target population is a subset of the population including (a) a first subset of the plurality of datasets, wherein each dataset of the first subset comprise, as one or more of the plurality of attributes, an indicator explicitly representing one or more conditions based on one or more deterministic criteria associated with the entity eligibility, and (b) a second subset of the plurality of datasets, wherein each dataset of the second subset excludes the indicator and comprises, as one or more of the plurality of attributes, data implicitly representing the one or more conditions based on the one or more deterministic criteria ([0019] teaches using machine learning and training machine learning. [0021] teaches trial data is used by machine learning to determine target patient groups for a clinical trial (interpreted by Examiner as applying a first machine learning model to the plurality of datasets to identify a target population); wherein the preclinical trial data is comparatively analyzed by machine learning against past and current clinical trials; wherein the analytical comparison is used to generate a report of the analytical comparison to the software application; wherein the generated analytical comparison report is sent to the software application, wherein the machine learning model training is supplemented by data from one or more of the modules (interpreted by Examiner as wherein the first machine learning model is trained to learn associations between the plurality of attributes and entity eligibility for the target population) [0086] teaches that by combining a standard data model with the linked and ranked biomarker-adverse event associations the system may facilitate demographic based queries to attain new insights derived from historical clinical trial data. For example, the system may receive as input a list of biomarkers such as chronic heart failure, Caucasian (a subset of population), and cholesterol (interpreted by Examiner as a first subset of the plurality of datasets, wherein each dataset of the first subset comprise, as one or more of the plurality of attributes, an indicator explicitly representing one or more conditions based on one or more deterministic criteria associated with the entity eligibility) and return a list of papers that provide relevant information for that subset of the population in the context of the other biomarkers (interpreted by Examiner as a second subset of the plurality of datasets, wherein each dataset of the second subset excludes the indicator and comprises, as one or more of the plurality of attributes, data implicitly representing the one or more conditions based on the one or more deterministic criteria). The system may identify at-risk populations (interpreted by Examiner as the target population) based on biomarkers and trial drug characteristics and [0241] and Fig. 43 teach inputs and outputs to machine learning model for clinical trial analysis. The inputs such as preclinical trial data 4310 is used for various operations, some of which comprise providing preclinical comparative analytical comparison of a sponsors target and compound with data for former and current clinical trial sites by variables such as: target, drug, endpoints, SAEs and AEs 4340 and assist in the defining the patient populations that would be best suited for a specific clinical trial (disease target and drug) [0245] teaches this at-risk cohort, generated by machine learning, provides a most expeditious method to intervene at scale to a pending medical emergency. Service comprises using preclinical trial data to determine the patient populations that would be best suited for a specific clinical trial.); merging, by the one or more processors, the first subset and the second subset of the plurality of datasets into a first data object representing the target population ([0080] teaches linking biomarkers with molecules, proteins, and genetic data to provide insight into the relationship between biomarkers, outcomes, and adverse events. The system uses natural language processing techniques on a large corpus of medical literature to perform advanced text mining to identify biomarkers associated with adverse events and to curate a comprehensive profile of biomarker-outcome associations. Having a comprehensive profile of ranked biomarker-outcome data allows the system to predict biomarkers associated with a given disease and serious adverse events linked to biomarker data (interpreted by Examiner as merging into a first data object representing a target population)); receiving, by the one or more processors and as input via an interactive interface, a user input indicative of a plurality of parameters that correspond to values contained in the first data object, wherein the plurality of parameters exhibit variation within the target population ([0084] teaches allowing a client to input a list of biomarkers that will be measured during screening or continuously throughout a clinical trial for each patient. [0113] teaches the use of a user interface and [0195] teaches sponsors and sometimes clinicians who will input a series of parameters such as trial endpoints and biomarkers of interest (interpreted by examiner as receiving a user input indicative of a plurality of parameters that correspond to values contained in the first data object)); generating, by the one or more processors and using a second machine learning model, a plurality of categories corresponding to a plurality of unique combinations of options identified across one or more of the plurality of parameters selected to represent the variation ([0126] teaches data is received from a data extraction engine (interpreted by Examiner as the second machine learning model) in each of several categories of data, nodes are assigned to each entity identified in each category and attributes of the entity are assigned to the node. The relationships between entities are assigned, both within the category of knowledge and between all other categories of knowledge (interpreted by Examiner as generating a plurality of categories corresponding to a plurality of unique combinations of the plurality of options identifies across one or more of the plurality of parameters selected to represent the variation) [0113] teaches the data extraction engine uses natural language processing techniques to extract and classify information and [0242] teaches machine learning tasks and the methods thereof may be implemented by the disclosed information regarding the various modules contained within this specification (e.g., data extraction engine.); determining, by the one or more processors, a first distribution across the plurality of categories within the target population ([0117] teaches predicting distribution. [0153] teaches printing out the summary of the distributions and generating plots comparing the molecular properties as defined in the computer properties function of the active and generated distributions (interpreted by examiner as determining a first distribution)); generating, by the one or more processors, a plurality of first similarity-based subsets by applying one or more sampling techniques to the first data object across the one or more of the plurality of parameters, wherein the plurality of first similarity-based subsets are associated with a plurality of deviation measures determined based on the first distribution and a second distribution determined across the plurality of categories within the plurality of first similarity-based subsets, the plurality of deviation measures indicative of a representativeness of the plurality of first similarity-based subsets relative to the target population across the plurality of categories ([0078] teaches identifying those patients and potentially other patients who share similar traits and biomarkers and [0116] teaches that the module utilizes the knowledge graph and data analysis engine capabilities of the data platform, and in one embodiment is configured to return clinical trials similar to a specified clinical trial in one or more aspects (e.g., proteins and ligands studied, methodology, results, etc.) based on semantic clustering within the knowledge graph (interpreted by Examiner as means for generating a plurality of first similarity-based subsets by applying one or more sampling techniques to the first data object across the one or more the plurality of parameters) [0143] teaches passing examples through a variational autoencoder (VAE) together with a reinforcement learning. The readout vector is subsequently split into the mean and variance vectors which serve and as the parameters of the posterior distribution from the sampling, [0144] teaches a prior distribution as well as posterior distribution (interpreted by Examiner as the second distribution) [0162] teaches a model architecture of a Sampling module for de novo drug discovery according to one embodiment. The sampling module comprises a split readout function that produces the mean of the batch (interpreted by Examiner as wherein the plurality of first similarity-based subsets are associated with a plurality of deviation measures, the plurality of deviation measures indicative of a representativeness of the plurality of first similarity-based subsets relative to the target population across the plurality of categories)); generating, by the one or more processors, a second data object comprising one of the plurality of first similarity-based subsets associated with a lowest deviation measure of the plurality of deviation measures indicative of a most accurate representation of the target population among the plurality of first similarity-based subsets ([0113] teaches the user may submit a query for identification of molecules likely to have similar bioactivity to a molecule with known bioactivity. The data analysis engine may process the knowledge graph through a GNN to identify such molecules based on the information and relationships in the knowledge graph. [0125] teaches that the system determines a number of similarities and relationships between the first molecule (interpreted by Examiner as the first subset) and the second molecule (interpreted by Examiner as the second subset). Thus, the system determines that the second molecule is likely to have a similar effect on the disease as the first molecule. Further, the system identifies a second clinical trial that suggests that the second molecule has lesser side effects than the first molecule. As the second molecule meets the query criteria, it is returned as a response to the query (interpreted by Examiner as generating a second data object comprising one of the plurality of first similarity-based subsets associated with a lowest deviation measure of the plurality of deviation measures indicative of a most accurate representation of the target population among the plurality of first similarity-based subsets)); providing, by the one or more processors, the second data object for display via the interactive interface in response to the user input ([0113] teaches a user interface and [0263] teaches a display and an input output I/O unit. [0210] teaches display of outputs (interpreted by Examiner as means to providing the second data object for display via the interactive interface in response to the user input)); generating, by the one or more processors and based on feedback associated with the plurality of deviation measures determined for the plurality of first similarity- based subsets, a plurality of second similarity-based subsets by adjusting the one or more sampling techniques and applying the adjusted one or more sampling techniques to the first data object across the one or more of the plurality of parameters ([0162] and Fig. 22 teach a sampling module. [0178] teaches another type of query that may be received and processed by the system is a sample size query. A sample size query may consist of input terms which describe the goal, methodology, and design of a potential clinical trial in order to provide context for the sample size estimator 3503 to calculate a sample size. Sample size estimator 3503 may use prior information sampled from similar historical clinical trial data based on input terms. [0074] teaches adjusting a required sample size based on the treatment effect at any point in time, [0076] teaches the statistical significance of the difference in treatment effect is calculated by getting the changes of biomarkers from the start of the clinical trial up to a time T. Therefore, having access to the primary endpoint value at any time T would allow to compute the statistical significance at this time T. Then, what would be left to do would be to use the actual difference in treatment effect at this time T to estimate the sample size needed to achieve a 5% or 1% statistical significance (interpreted by Examiner as generating a plurality of second similarity-based subsets by adjusting the one or more sampling techniques and applying the adjusted one or more sampling techniques to the first data object across the one or more of the the plurality of parameters)) receiving, by the one or more processors, performance data indicative of one or more of (i) whether the target population identified using the first machine learning model is a true eligible population of the population or (ii) whether the plurality of categories generated using the second machine learning model are representative of the true eligible population ([0086] teaches to help with clinical trial design the system may identify at-risk populations based on biomarkers and trial drug characteristics. For instance, a biomarker may be associated with some biological process and the biological process may be regulated by certain proteins and furthermore, the protein function may be impacted by some molecule which may be present in a drug. Using the data platform knowledge graph the system may be able quickly identify the connection, via biological pathways, between a biomarker and a trial drug. At-risk populations may be selected based upon identified biological pathways that may be compromised due to underlying conditions, genetics, physical disposition, etc. For example, a population with low blood pressure biomarkers could be considered an at-risk population for a drug that purports to lower blood pressure to cause some effect and [0189] teaches data from preclinical analytics could assist in the defining the patient populations that would be best suited for a specific clinical trial (disease target and drug) (interpreted by Examiner as performance data indicative of whether the plurality of categories generated using the second machine learning model are representative of the true eligible population)); and updating, by the one or more processors and based on the performance data, one or more of the first machine learning model or the second machine learning model, wherein the updating includes adjusting one or more layers or one or more weights of the one or more of the first machine learning model or the second machine learning model to increase an accuracy of one or more of a next target population identified or a next plurality of categories generated, respectively, for future sampling ([0078] teaches the machine learning aspect of the system arriving at a higher confidence decision to remove those patients and potentially other patients who share similar traits and biomarkers faster than it took the sites in the first example to share information, thus better protecting human life and producing successful trials. [0084] teaches input a list of biomarkers that will be measured during screening or continuously throughout a clinical trial for each patient. [0155] teaches during training of the neural network machine learning model with inputs of a 3D grid 1802 of Gaussian-like atom type densities, the weights are iteratively modified in order to minimize the losses 1804, which is some measure of the goodness of fit of the model outputs to the training data. In an embodiment, the procedure is performed using some variation of gradient descent, where the changes applied to each weight during the update step are proportional in some way to the gradient of the loss with respect to the weight in question. [0193] teaches sponsor and trial site edge machine learning apps may be updated with relevant longitudinal patient data results from Phase 1 trials to the Phase 2 trials, then to the Phase 3 trials. (interpreted by Examiner as adjusting one or more layers or one or more weights of the one or more of the first machine learning model or the second machine learning model to increase an accuracy of one or more of a next target population identified or a next plurality of categories generated, respectively, for future sampling)).
REGARDING CLAIM 2
Knuff disclose the limitation of claim 1.
Knuff further discloses:
The method of claim 1, wherein the second data object further comprises one or more additional first similarity-based subsets of the plurality of first similarity-based subsets, wherein the one or more first additional similarity-based subsets of the second data object are sorted by deviation measure (Knuff at [0117] teaches similar absorption or metabolism characteristics and that users can research a large ADMET database based on aspects of interest by entering queries through the EDA interface (interpreted by Examiner as means for the second data object to comprise one or more additional first similarity-based subsets) [0080] teaches identifying biomarkers associated with adverse events and to curate a comprehensive profile of biomarker-outcome associations. These associations may then be ranked to identify the most-common biomarker-outcome association pairs. Having a comprehensive profile of ranked biomarker-outcome data allows the system to predict biomarkers associated with a given disease and serious adverse events linked to biomarker data.).
REGARDING CLAIM 3
Knuff disclose the limitation of claim 3.
Knuff further discloses:
The method of claim 1, wherein the first subset of the plurality of datasets are associated with a first subset of the plurality of entities and the second subset of the plurality of datasets are associated with a second subset of the plurality of entities unique from the first subset of the plurality of entities (Knuff at [0086] teaches the system may receive as input a list of biomarkers such as chronic heart failure, Caucasian (a subset of population), and cholesterol (interpreted by Examiner as first subset of the plurality of datasets are associated with a first subset of the plurality of entities) and return a list of papers that provide relevant information for that subset of the population in the context of the other biomarkers (interpreted by Examiner as the second subset of the plurality of datasets are associated with a second subset of the plurality of entities unique from the first subset of the plurality of entities)).
REGARDING CLAIM 5
Knuff disclose the limitation of claim 4.
Knuff further discloses:
The method of claim 4, further comprising determining the plurality of deviation measures for the plurality of first similarity-based subsets (Knuff at [0220] teaches a root mean square deviation and [0106] an outcome which is used herein as a measure within a clinical trial which is used to assess the effect, both positive and negative, of an intervention or treatment. In clinical trials such measures of direct importance of for an individual may include, but are not limited to, survival, quality of life, morbidity, suffering, functional impairment, and changes in symptoms.).
REGARDING CLAIM 6
Knuff disclose the limitation of claim 5.
Knuff further discloses:
The method of claim 5, wherein the determining the plurality of deviation measures for the plurality of first similarity-based subsets includes: for a category of plurality of categories: determining a probability of the representative category being selected from the first data object, determining a probability of the category being selected from the plurality of first similarity-based subsets, and generating a plurality of divergence values for the category within the plurality of first similarity-based subsets based at least in part on the probability of the category being selected from the first data object and/or the plurality of probabilities of the category being selected from the plurality of first similarity-based subsets; and generating the plurality of deviation measures for the plurality of first similarity-based subsets based on the plurality of divergence values generated across the plurality of categories within the plurality of first similarity-based subsets (Knuff at [0099] teaches edges may also comprise value, conditions, or other information, such as edge weights or probabilities. [0105] teaches measure of how much the actual probability of a particular co-occurrence of events (word-pairs) differs from its expected probability on the basis of the probabilities of the individual events and the assumption of independence. The calculated NPMI value is bounded between the values of negative one and one (−1, 1), inclusive. A value of negative one indicates the word-pair occur separately, but never occur together. A value of zero indicates independence of the word-pair in which co-occurrences happen at random. A value of one indicates complete co-occurrence, or that the word-pair only exist together (interpreted by Examiner as determining a probability of the category being selected from the plurality of first similarity-based subsets, and generating a plurality of divergence values for the category within the plurality of first similarity-based subsets and generating the plurality of deviation measures)).
REGARDING CLAIM 7
Knuff disclose the limitation of claim 1.
Knuff further discloses:
The method of claim 1, further comprising: generating, by the one or more processors, a plurality of member size categories, each member size category of the plurality of member size categories associated with a unique number of group members; and assigning, by the one or more processors, one or more first similarity-based subsets of the plurality of first similarity-based subsets to a member size category of the plurality of member size categories based on the number of group members in the one or more first similarity-based subsets (Knuff at [0074] teaches allow the Sponsor/CRO to adjust their required sample size (interpreted by examiner as member size category) based on the treatment effect at any point in time, or to withdraw patients (interpreted by Examiner as the members) faster in case of emergencies. [0076] teaches estimate the sample size needed to achieve a 5% or 1% statistical significance and [0178] teaches the sample size estimator may use prior information sampled from similar historical clinical trial data based on input terms (interpreted by Examiner as generating, by the one or more processors, a plurality of member size categories, each member size category of the plurality of member size categories associated with a unique number of group members; and assigning, by the one or more processors, one or more first similarity-based subsets of the plurality of first similarity-based subsets to a member size category of the plurality of member size categories based on the number of group members in the one or more first similarity-based subsets)).
REGARDING CLAIM 8
Knuff disclose the limitation of claim 1.
Knuff further discloses:
The method of claim 7, further comprising: determining, for the member size category of the plurality of member size categories, a first similarity-based subset of the one or more first similarity-based subsets assigned to the member size category with a lowest deviation measure (Knuff at [0113] teaches the user may submit a query for identification of molecules likely to have similar bioactivity to a molecule with known bioactivity. The data analysis engine may process the knowledge graph through a GNN to identify such molecules based on the information and relationships in the knowledge graph. [0125] teaches that the system determines a number of similarities and relationships between the first molecule and the second molecule. Thus, the system determines that the second molecule is likely to have a similar effect on the disease as the first molecule. Further, the system identifies a second clinical trial that suggests that the second molecule has lesser side effects than the first molecule. As the second molecule meets the query criteria, it is returned as a response to the query (interpreted by Examiner as a first similarity-based subset with a lowest deviation measure)).
REGARDING CLAIM 9
Claim 9 is analogous to Claim 8 thus Claim 9 is similarly analyzed and rejected in a manner consistent with the rejection of Claim 8.
REGARDING CLAIM 10
Knuff disclose the limitation of claim 1.
Knuff further discloses:
The method of claim 9, wherein the plurality of parameters includes a threshold deviation measure, and wherein providing the second data object for display via the interactive interface includes: comparing, for the member size category, the lowest deviation measure against the threshold deviation measure; determining, based on a result of the comparing, a satisfactory member size category; and providing, along with the second data object, a visual indicia of the satisfactory member size category (Knuff at [0170] teaches threshold used to determine similarities and [0244] teaches reports of those biomarkers may be generated either by default, or by the biomarker surpassing some threshold—which may be arbitrarily decided by sponsors and clinicians (interpreted by Examiner as wherein the plurality of parameters includes a threshold deviation measure) [0078] teaches the claimed invention is receiving those blood analyses in real-time, comparing those patients and the patient's other biometrics (i.e., vital signs, etc.) data with the rest of the patients for differences and commonalities. The system also compares the current trial with past and other ongoing clinical trials. The machine learning aspect of the system arriving at a higher confidence decision to remove those patients and potentially other patients who share similar traits and biomarkers faster than it took the sites in the first example to share information, thus better protecting human life and producing successful trials (interpreted by Examiner as comparing, for the member size category, the lowest deviation measure against the threshold deviation measure and determining, based on a result of the comparing, a satisfactory member size category) [0153] teaches results may then be summarized in two ways: by printing out the summary of the distributions and generating plots comparing the molecular properties as defined in the computer properties function of the active and generated distributions (interpreted by Examiner as providing, along with the second data object, a visual indicia of the satisfactory member size category)).
REGARDING CLAIMS 11-13, 15-18 and 20
Claims 11-13, 15-18 and 20 are analogous to Claims 1-3 and 5-10 thus Claims 11-13, 15-18 and 20 are similarly analyzed and rejected in a manner consistent with the rejection of Claims 1-3 and 5-10.
REGARDING CLAIM 21
Knuff disclose the limitation of claim 1.
Knuff further discloses:
The computer-implemented method of claim 1, wherein adjusting the one or more sampling techniques and applying the adjusted one or more sampling techniques to the first data object across the one or more of the plurality of parameters comprises adjusting, based on the feedback, the one or more sampling techniques to minimize associated deviation measures and increase representativeness of the plurality of second similarity-based subsets relative to the target population across at least one of the plurality of categories ([0162] and Fig. 22 teach a sampling module. [0178] teaches another type of query that may be received and processed by the system is a sample size query. A sample size query may consist of input terms which describe the goal, methodology, and design of a potential clinical trial in order to provide context for the sample size estimator 3503 to calculate a sample size. Sample size estimator 3503 may use prior information sampled from similar historical clinical trial data based on input terms. [0074] teaches adjusting a required sample size based on the treatment effect at any point in time, [0076] teaches the statistical significance of the difference in treatment effect is calculated by getting the changes of biomarkers from the start of the clinical trial up to a time T. Therefore, having access to the primary endpoint value at any time T would allow to compute the statistical significance at this time T. Then, what would be left to do would be to use the actual difference in treatment effect at this time T to estimate the sample size needed to achieve a 5% or 1% statistical significance (interpreted by Examiner as adjusting, based on the feedback, the one or more sampling techniques)).
REGARDING CLAIM 22
Knuff disclose the limitation of claim 1.
Knuff further discloses:
The computer-implemented method of claim 1, further comprising: determining the first distribution by determining a first plurality of probabilities of the plurality of categories being selected from the first data object; and determining the second distribution by determining a second plurality of probabilities of the plurality of categories being selected from the plurality of first similarity-based subsets ([0117] teaches predicting distribution and [0153] teaches printing out the summary of the distributions and generating plots comparing the molecular properties as defined in the computer properties function of the active and generated distributions, [0177] teaches accounting for probability to decide whether to take a bioactivity rating or disregard it as inaccurate and [0205] teaches assigning probability to each node (interpreted by Examiner as means to determine the first/second distribution by determining a first/second plurality of probabilities)).
REGARDING CLAIM 24
Knuff disclose the limitation of claim 1.
Knuff further discloses:
The computer-implemented method of claim 1, wherein the second machine learning model performs feature selection to identify a feature importance associated with the plurality of parameters in stratifying the target population, and automatically selects the one or more of the plurality of parameters based on the feature importance ([0119] teaches as the medical information is downloaded, it is fed to a data extraction engine which may perform a series of operations to extract data from the medical information materials. Once the text has been extracted from the materials, natural language processing (NLP) techniques may be used to extract useful information from the materials for use in analysis by machine learning algorithms. Of particular importance is recognition of standardized biochemistry naming conventions. The data extraction engine feeds the extracted data to a knowledge graph constructor, which constructs a knowledge graph based on the information in the data, representing informational entities (e.g., proteins, molecules, diseases, study results, people) as vertices of a graph and relationships between the entities as edges of the graph (interpreted by Examiner as identify a feature importance associated with the plurality of parameters in stratifying the target population, and automatically selects the one or more of the plurality of parameters based on the feature importance)).
Response to Arguments
Rejection under 35 U.S.C. § 101
Regarding the rejection of claims 1-3, 5-13, 15-18, 20-22 and 24, the Examiner has considered the Applicant’s arguments, but does not find them persuasive. Applicant argues:
The specification identifies several problems in the technical field of data science with relation to data sampling for clinical trial cohort selection. Specifically, "the current landscape of clinical trials often grapples with the challenge of ensuring that the selected participant cohorts are truly representative of the broader target population." (Specification at para. [0002].) "A key issue [suffered by traditional approaches to these challenges] lies in the initial estimation of the true prevalence of a condition, especially when a significant proportion of individuals may be undiagnosed or misdiagnosed," which "results in an inaccurate picture of the true target population, thereby hindering the effectiveness of devising representative trial quotas." (Id. at para. [0003].) Additionally, the traditional approaches do not adequately "account for the complexity and diversity of the target population." (Id. at para. [0003].) Such inadequate representation of the broader target population "could lead to skewed results and an over- or underestimation of the effectiveness of a new treatment or drug on certain demographic or clinical groups," which may "impede the advancement of medical research and potentially limit the applicability of new treatments or drugs to all those who may benefit." (Id. at para. [0005].). The specification then proceeds to describe a particular solution to these problems, as reflected by the independent claims as amended, by applying machine learning models to identify a target population (e.g., to detect eligibility) and generate representative categories of the target population. (See Id. at paras. [0026], [0051].)… This solves one of the challenges discussed above related to the initial estimation of the true prevalence of a condition as a result of many individuals being undiagnosed or misdiagnosed. Thus, the claimed first machine learning model is specifically applied to the technical field of data science to provide a particular solution for identifying both entities having explicit indicators of eligibility as well as those that do not, but otherwise are associated with data implicitly indicating eligibility (e.g., to identify those that have gone un- or misdiagnosed), for inclusion in the target population. Additionally, and as reflected in the claims, a second machine learning model may be employed to generate representative categories for the true eligible population (e.g., the target population). (Id. at paras. [0051], [0067]-[0068].)… Thus, the claimed second machine learning-based generation of representative categories (e.g., "categories corresponding to a plurality of unique combinations of options identified across one or more of the plurality of parameters selected to represent the variation [of the target population]"), that are then leveraged for subsequent deviation measure determinations and comparisons among the similarity-based subsets, is specifically applied to the technical field of data science to provide a particular solution to account for the complexity and diversity of the target population to help ensure highly representative similarity-based subsets are generated and selectable as a sample cohort for the clinical trial. "Furthermore, the cohort generation algorithm 127 also takes into account the performance of past generated populations [such as eligible populations, representative categories, sample cohorts, cohort rankings, or the like determined in the past] in similar contexts." (Id. at para. [0053].) For example, the machine learning models may be updated or retrained based on "comparison results 416 that compare a previous output of the corresponding machine- learning model to apply the previous result [a known outcome indicated by the performance] to re-train the machine-learning model." (Id. at para. [0110].) Specifically, the machine learning modeled may be updated or re-trained "by adjusting one or more weights and/or one or more layers of the machine-learning model." (Id. at para. [0111].) Therefore, similar to the claim at issue in Ex Parte Desjardins that was related to training a machine learning model, when evaluating the present claim as a whole, at least the "updating, ... based on the performance data, one or more of the first machine learning model or the second machine learning model" element reflects a disclosed improvement to how the machine learning model itself operates, and thus integrates the alleged abstract idea into a practical application.
Regarding 1, The Examiner respectfully disagrees and notes that the problems/solutions the Applicant refers to are all non-technical problem/solution and none of them are not rooted in computer technology, nor relate to improvements in machine learning technology. Moreover, the specification discloses improvements to the accuracy and efficiency of cohort generation and the selection of representative cohorts ([0026]), improvement to ability to generate true eligible populations and representative categories ([0055]), improvements to the representation ([0075]) and improvements to the selection of representative similarity-based subsets ([0076]), all of which are non-technical and provide improvements in the medical field for clinical trial cohort generation. None of the above mentioned improvements, nor the improvements mentioned in Applicants specification support improvements to machine learning technology. Rather, the claims apply machine learning technology to identify a target population and generate representative categories of the target population to increase the accuracy of one or more of a next target population identified or a next plurality of categories generated for future sampling. As can be seen, the claim apply conventional machine learning technology and so the first and second machine learning models are interpreted as additional elements that are applied to the abstract idea. MPEP 2106.04(d)(I) indicates that merely saying “apply it” or equivalent to the abstract idea cannot provide a practical application. Accordingly, even in combination, the additional elements of a first and second machine learning models do not integrate the abstract idea into a practical application.
Applicant respectfully asserts that the independent claims, when evaluated as whole, use the first and second machine learning models (e.g., additional elements) in a specific and particular manner, and in an ordered combination with one another, to provide for increased representativeness in data sampling (e.g., by including datasets with mis- or undiagnosed conditions in the target population and accounting for variation in the target population when generating categories to facilitate diverse representation) that is not well-understood, routine, or conventional activity. (See Specification, paras. [0002]-[0005], [0025]-[0026], [0051], [0055].) Therefore, even if the claims are found to be directed to an abstract idea under Step 2A, the claims nonetheless transform the nature of the claim into a patent-eligible application under Step 2B, and thus qualify as eligible subject matter. If the Examiner disagrees, Applicant respectfully requests that the Examiner expressly support the rejection in writing in accordance with MPEP 2106.07(a). For at least these reasons, Applicant respectfully requests reconsideration and withdrawal of the rejection under 35 U.S.C. § 101.
Regarding 2, The Examiner respectfully disagrees. The additional elements of the first and second machine learning models were determined to be (“apply it”) to the abstract idea. MPEP2106.05(1)(A) indicates that merely saying “apply it’ or equivalent to the abstract idea cannot provide an inventive concept (“significantly more’). Therefore when considering the additional elements alone, and in combination, there is no inventive concept in the claim, and thus the claim is not patent eligible.
Rejection under 35 U.S.C. § 102
Regarding the rejection of claims 1-3, 5-13, 15-18, 20-22 and 24, the Examiner has considered the Applicant’s arguments, but does not find them persuasive. Applicant argues:
Knuff does not disclose application of a machine learning model to a plurality of datasets to identify, as a target population, two specific types of datasets, including " (a) a first subset of the plurality of datasets, wherein each dataset of the first subset comprises, as one or more of the plurality of attributes, an indicator explicitly representing one or more conditions based on one or more deterministic criteria associated with the entity eligibility, and (b) a second subset of the plurality of datasets, wherein each dataset of the second subset excludes the indicator and comprises, as one or more of the plurality of attributes, data implicitly representing the one or more conditions based on the one or more deterministic criteria," as recited in independent claim 1.
Regarding 1, The Examiner respectfully disagrees. Knuff at [0019] teaches using machine learning and training machine learning and at [0021] teaches trial data is used by machine learning to determine target patient groups for a clinical trial (interpreted by Examiner as applying a first machine learning model to the plurality of datasets to identify a target population) [0086] teaches the system may receive as input a list of biomarkers such as chronic heart failure, Caucasian (a subset of population), and cholesterol (interpreted by Examiner as a first subset of the plurality of datasets, wherein each dataset of the first subset comprise, as one or more of the plurality of attributes, an indicator explicitly representing one or more conditions based on one or more deterministic criteria associated with the entity eligibility) and return a list of papers that provide relevant information for that subset of the population in the context of the other biomarkers (interpreted by Examiner as a second subset of the plurality of datasets, wherein each dataset of the second subset excludes the indicator and comprises, as one or more of the plurality of attributes, data implicitly representing the one or more conditions based on the one or more deterministic criteria). The system may identify at-risk populations (interpreted by Examiner as the target population) based on biomarkers and trial drug characteristics.
Given the broadest reasonable interpretation, the cited reference teaches the claimed features.
First, Knuff fails to disclose using a machine learning model to generate the categories of data. Second, Knuff fails to disclose categories generated to represent variation of a target population. Therefore, Knuff also does not disclose at least, "generating, by the one or more processors and using a second machine learning model, a plurality of categories corresponding to a plurality of unique combinations of options identified across one or more of the plurality of parameters selected to represent the variation [of the target population]."
Regarding 2, The Examiner respectfully disagrees. Knuff at [0126] teaches data is received from a data extraction engine (interpreted by Examiner as the second machine learning model) in each of several categories of data, nodes are assigned to each entity identified in each category and attributes of the entity are assigned to the node. The relationships between entities are assigned, both within the category of knowledge and between all other categories of knowledge, which is interpreted by Examiner as the plurality of categories corresponding to a plurality of unique combinations of the plurality of options identifies across one or more of the plurality of parameters selected to represent the variation.
Given the broadest reasonable interpretation, the cited reference teaches the claimed features.
Further, Knuff fails to disclose the new elements of independent claim 1 reciting: receiving, by the one or more processors, performance data indicative of one or more of (i) whether the target population identified using the first machine learning model is a true eligible population of the population or (ii) whether the plurality of categories generated using the second machine learning model are representative of the true eligible population; and updating, by the one or more processors and based on the performance data, one or more of the first machine learning model or the second machine learning model, wherein the updating includes adjusting one or more layers or one or more weights of the one or more of the first machine learning model or the second machine learning model to increase an accuracy of one or more of a next target population identified or a next plurality of categories generated, respectively, for future sampling.
Regarding 3, The Examiner respectfully disagrees. Knuff discloses the following limitations. Knuff at [0086] teaches to help with clinical trial design the system may identify at-risk populations based on biomarkers and trial drug characteristics. For instance, a biomarker may be associated with some biological process and the biological process may be regulated by certain proteins and furthermore, the protein function may be impacted by some molecule which may be present in a drug. Using the data platform knowledge graph the system may be able quickly identify the connection, via biological pathways, between a biomarker and a trial drug. At-risk populations may be selected based upon identified biological pathways that may be compromised due to underlying conditions, genetics, physical disposition, etc. For example, a population with low blood pressure biomarkers could be considered an at-risk population for a drug that purports to lower blood pressure to cause some effect and [0189] teaches data from preclinical analytics could assist in the defining the patient populations that would be best suited for a specific clinical trial (disease target and drug), which is interpreted by Examiner as performance data indicative of whether the plurality of categories generated using the second machine learning model are representative of the true eligible population. Moreover, Knuff at [0078] teaches the machine learning aspect of the system arriving at a higher confidence decision to remove those patients and potentially other patients who share similar traits and biomarkers faster than it took the sites in the first example to share information, thus better protecting human life and producing successful trials, at [0084] teaches input a list of biomarkers that will be measured during screening or continuously throughout a clinical trial for each patient, at [0155] teaches during training of the neural network machine learning model with inputs of a 3D grid 1802 of Gaussian-like atom type densities, the weights are iteratively modified in order to minimize the losses 1804, which is some measure of the goodness of fit of the model outputs to the training data. In an embodiment, the procedure is performed using some variation of gradient descent, where the changes applied to each weight during the update step are proportional in some way to the gradient of the loss with respect to the weight in question and at [0193] teaches sponsor and trial site edge machine learning apps may be updated with relevant longitudinal patient data results from Phase 1 trials to the Phase 2 trials, then to the Phase 3 trials. This is interpreted by Examiner as adjusting one or more layers or one or more weights of the one or more of the first machine learning model or the second machine learning model to increase an accuracy of one or more of a next target population identified or a next plurality of categories generated, respectively, for future sampling.
Given the broadest reasonable interpretation, the cited reference teaches the claimed features.
Conclusion
The prior art made of record though not relied upon in the present basis of rejection are noted in the attached PTO 892 and include:
Birnbaum (US 2018/0300640) discloses systems and methods for model-assisted cohort selection. Will (US 2020/0243167) discloses predicting clinical trial eligibility based on cohort trends.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to LIZA TONY KANAAN whose telephone number is (571)272-4664. The examiner can normally be reached on Mon-Thu 9:00am-6:00pm ET.
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, Robert Morgan can be reached on 571-272-6773. 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 the 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/docs 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.
/LIZA TONY KANAAN/Examiner, Art Unit 3683
Prosecution Insights