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/Argument
1. Amendments to claims 2, 5, 7, 9, 12, 14, 16, 18, and 20 overcome the previous rejection under 35 U.S.C. 112(b).
2. Applicant’s argument to the rejection under 35 U.S.C. 101 filed on December 18, 2025 have been fully considered but are not persuasive. Applicant’s arguments on pages 10-11 of the Remarks regarding the Step 2A, Prong One analysis, including the reliance on the August 4, 2025 USPTO memorandum, have been fully considered but are not persuasive. The Examiner respectfully disagrees that the amended claims do not recite a mental process.
Applicant argues that the amended claims are not directed to mental processes because the claims recite execution by at least one processing unit of processor-executable program code on a storage device and allegedly cannot be performed in the human mind. However, the memorandum cited by the Applicant explains that mental process grouping includes limitations that can practically be performed in the human mind, including observations, evaluations, judgements, and opinions. The memorandum further explains that the analysis request determining whether the recited claim limitations can practically be performed in the mentally, not merely whether the claims recite execution by a processor.
Under the broadest reasonable interpretation, the recited limitations such identifying instances of label examples within a data sample, determining an associated label and location of the identified label example, annotating the data sample with associated label and location, and generating an inference in reply to a request involves reviewing information, recognizing patterns or content, associating information with known labels, and determining a conclusion based on the information. These steps correspond to acts of observation, evaluation, and judgement that can practically be performed in the human mind or with the aid of pen and paper or basic computational tools. Similar concepts involving collecting information, analyzing it, and producing results have been identified by the Federal Circuit as mental processes. See Electric Power Group v. Alstom, S.A., 830 F.3d 1350, 1353-54 (Fed. Cir. 2016).
Applicant’s reliance on the recitation of processor-executable program code, a processing unit, and machine learning model execution does not remove the claim from the mental processes grouping. Courts have consistently explained that claims may still recite mental processes even when implemented on a computer or within a computer environment. Thus, implementing the recited information analysis using a processor executing program code does not change the nature of the underlying abstract idea.
Accordingly, the Examiner maintains that the amended claims recite concepts that fall within the mental processes grouping of abstract ideas under Step 2A, Prong One, and Applicant’s arguments do not overcome the rejection.
Applicant’s arguments presented on pages 11-13 of the Remarks regarding the Step 2A, Prong Two analysis has been fully considered but are not persuasive. The Examiner respectfully disagrees that the amended claims integrate the alleged abstract idea into a practical application.
Applicant first argues that the amended claims integrate the judicial exception into a practical application because the claims recite a system, method, and computer readable medium that annotate data samples, train a machine l earning model using annotated data, and generate inferences using the trained machine learning model. Applicant further points to the Specification, including FIG. 1, 4, and 5, as describing example systems that generate annotated data samples for training machine learning models.
However, the additional elements identified by the Applicant merely describes implementing the abstract idea using generic computer components and conventional data processing operations. The recited steps of receiving data samples, identifying instances of label examples within the data sample, determining an associated label and location, annotating the data sample with that information, storing the annotated data, training the machine learning model, and generating an inference represent information analysis performed on data. Implementing these steps using a processing unit executing program code and storing information in a storage device amounts to using a computer as a tool to perform the abstract idea rather than integrating the abstract idea into a practical application.
Additionally, although the Specification may describe example implementations and figures illustrating systems that perform these operations, the eligibility analysis under Step 2A, Prong Two is based on the limitations recited in the claims. The claims themselves do not recite a specific technological improvement to computer functionality or another technical field, but instead recite generic computer implementation of information analysis and labeling activities.
Applicant further argues that the Examiner’s analysis relies on overbroad reasoning and that the claims recite a “particular way to achieve a desired outcome,” citing the August 4, 2025 USPTO memorandum and related guidance. However, the Examiner’s analysis does not oversimplify the claims and instead evaluates the specific limitations recited in the claims as required under the subject matter eligibility guidance. While Applicant contends that the claims recite a particular way of annotating data samples and training machine learning models, the claims themselves recited the desired results of identifying label examples, determining labels and locations, annotating data samples, and generating an inference without specifying a particular technological mechanism for achieving those results. Thus, the claims merely recite applying the abstract idea using a computer as a tool rather than providing a technological improvement to computer functionality or another technical field.
To the extent Applicant relies on the Appeals Review Panel decision cited in the remarks, the Examiner notes that the claims at issue in the decision recited specific limitations reflecting an improvement to the operation of a machine learning model itself. In contrast, the present claims do not recite a specific improvement to the functioning of the machine learning model or another computer technology, but instead recite results-oriented data processing steps associated with annotating data samples and using those annotated data samples for model training and inference.
Applicant additionally argues that the amended claims provide meaningful limitations and therefore do not monopolize the abstract idea of data annotation. However, the additional elements identified by Applicant including the recitation of a processing unit, storage device, processor-executable program code, and, machine learning model operations, represent generic computer components performing routine information processing tasks. These elements do not impose a meaningful limit on the abstract idea but instead merely apply the abstract idea using conventional computer technology.
Accordingly, the additional elements do not integrate the judicial exception into a practical application under Step 2A, Prong Two. Therefore, the claims remain directed to the abstract idea. Accordingly, the rejection of claims 1-20 under 35 U.S.C. § is maintained.
2. Applicant’s arguments filed on December 18, 2025 regarding the 35 U.S.C. 102 and 103 rejections have been fully considered but are not persuasive. Applicant’s arguments presented on pages 13-15 regarding the rejection of claims 1, 3, 8, 10, 15, and 17 under 35 U.S.C. 102 as being anticipated by Bodapati have been fully considered but are not persuasive.
Applicant reproduces amended claim 1 and argues that Bodapati fails to disclose limitations including identifying instances of label examples within a data sample, determining an associated label and location within the data sample, and annotating the data sample with the associated label and location. Applicant further asserts that Bodapati merely receives document files and corresponding label files from a user and therefore does not perform the claimed invention and annotation operations.
However, Applicant’s arguments are not commensurate with the scope of the claims. The rejection is based on the language of the claims under the broadest reasonable interpretation, rather than on Applicant’s characterization of specific embodiments described in the specification. While Applicant focuses on the structures of files described in Bodapati, the claims themselves do not require that the data samples and label examples originate from particular file structures, nor do the claims require that annotation information be embedded within the same file as the data sample. Accordingly, Applicant’s arguments directed to the format or separation of files in Bodapati do not distinguish the claimed invention from the disclosure of the reference.
Applicant’s assertion that Bodapati does not identify instances of label examples within a data sample is not persuasive. Bodapati explicitly describes a process in which the system searches the documents to locate occurrences of labels and determine the locations of those labels within the document (col. 6, lines 11-20). This disclosure directly contradicts Applicant’s contention that the system merely receives documents and label files without identifying label occurrences within the documents. The act of searching through the documents to find occurrences of labels demonstrates that the system identifies label instances within the document itself.
Applicant’s argument that Bodapati does not determine the location of the identified example within the data sample is likewise unpersuasive. Bodapati further explains that entity spans may be represented using identifiers corresponding to positions within the document text, such as starting and stopping character numbers or other span identifiers (col. 12, lines 55-64; col. 13, lines 7-12). These span identifiers represent the positions of labels within the document text and therefore demonstrates that the system determines the location of the identified label example within the data sample.
Applicant also argues that Bodapati fails to disclose annotating the data sample with the associated label and the location. This argument is not persuasive. Bodapati describes associating documents with corresponding labels and using those labeled documents as training data for machine learning models. For example, Bodapati explains that documents and labels may be used to generate document embeddings and label embeddings and that the training system learns associations between documents and corresponding labels (col. 11, lines 51-58; col. 11, lines 62-67; col. 13, lines 1-6). The use of documents and their associated labels as labeled training data reflects that the documents have been labeled with the corresponding label information, which is consistent with annotating the data samples as recited in the claims.
Applicant’s argument that the labels and documents are provided in separate files also do not distinguish the claimed invention. The claims do not require the annotation information be embedded within the same file as the data sample, nor do they exclude implementations in which label examples are provided separately and subsequently identified within the documents during processing. Accordingly, the presence of separate document and label files in Bodapati does not negate the disclosure that the system identifies label occurrences within documents and associate these labels with corresponding document data.
For these reasons, Applicant’s arguments do not overcome the rejection. The rejection of claims 1, 3, 8, 10, 15, and 17 under U.S.C. 102(a)(2) as being anticipated by Bodapati is therefore maintained.
Applicant’s arguments regarding the rejection of claims 2, 4-7, 11-14, 16, 18, and 19 under 35 U.S.C. 103 have been fully considered but are not persuasive.
Applicant contends that the combinations of Bodapati in view of Anschel, Siddiqui, and Seiwald fail to render the claims obvious because Bodapati allegedly does not disclose the limitations discussed in the rejection of the independent claims. However, as discussed above with respect to the rejection under 35 U.S.C. 102, Bodapati does disclose the limitations identified by Applicant. Accordingly, Applicant’s arguments premised on the alleged deficiencies of Bodapati are not persuasive. Applicant does not present separate arguments directed to the teachings of Anschel, Siddiqui, or Seiwald, nor does Applicant address the rationale for combining the references as set forth in the Office Action. Therefore, Applicant has not demonstrated error in the Examiner’s findings regarding the cited combinations.
Accordingly, the rejection of claim 2, 4-7,9, 11-14, 16, 18, and 19 under 35 U.S.C. 103 as set forth in the Office Action is maintained.
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 a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more.
101 Subject Matter Eligibility Analysis
Step 1: Claims 1-20 are within the four statutory categories (a process, machine, manufacture or composition of matter).
Step 2A Prong One, Step 2A Prong Two, and Step 2B Analysis:
Step 2A Prong One asks if the claim recites a judicial exception (abstract idea, law of nature, or natural phenomenon). If the claim recites a judicial exception, analysis proceeds to Step 2A Prong Two, which asks if the claim recites additional elements that integrate the abstract idea into a practical application. If the claim does not integrate the judicial exception, analysis proceeds to Step 2B, which asks if the claim amounts to significantly more than the judicial exception. If the claim does not amount to significantly more than the judicial exception, the claim is not eligible subject matter under 35 U.S.C. 101.
None of the claims represent an improvement to technology.
Claims 1-7 and 15-20 are directed to storage mediums and processors which are machines. Claims 8-14 are directed to a method consisting of a series of steps, meaning that it is directed to the statutory category of process.
Regarding claim 1, the following claim elements are abstract ideas:
identify all instances of the plurality of label examples within the data sample (This is an abstract idea of a mental process.” The limitation recites a mental process involving the recognition and identification of specific examples within a data sample. This is the type of analysis that could be practically performed in the human mind with observation and judgement. For example, a person could manually examine a set of data samples and compare each one to a known list of examples, identifying where and how often each example appears within the data. Since it involves steps that can be carried out in the human mind or with the aid of pen and paper, it falls within the mental process grouping of abstract ideas. See MPEP 2106.04(a)(2)(III).);
for each identified label example, determine an associated one of the plurality of labels and a location of the identified label example in the data sample (This is an abstract idea of a “mental process.” The limitation recites a mental process involving associating identified examples with known labels and recording their positions within a data sample. This type of cognitive activity could be performed in the human mind using observation and reasoning. For example, a person could review a document, recognize specific examples (e.g., keywords or patterns), mentally determine the corresponding label (e.g., “spam” or “not spam”), and note the location of each example within the document. Since these steps can be carried out mentally or with the aid of pen and paper, they fall within the mental process grouping of abstract ideas.),
annotate the data sample with the associated one of the plurality of labels and the location (This is an abstract idea of a “mental process.” The limitation recites a mental process involving the act of annotating a data sample by adding a corresponding label and location for each identified example. This is the type of task that can be done mentally or manually by a person. For example, a person could review a document, recognize relevant examples, determine the associated label, and then write marginal notes indicating the label and position of each example within the document. Since this activity can be performed in the human mind or with the aid of pen and paper, it falls within the mental process grouping of abstract ideas. See MPEP 2106.04(a)(2)(III).);
generate…using the annotated data sample in reply to a request from at least one of an application and a service, an inference (This is an abstract idea of a “mental process.” The limitation recites evaluating annotated examples and determining a result or prediction based on those examples. For example, a person could review previously labeled examples and, based on observation and judgement, determine the appropriate label for a new example using patterns learned from the labeled data. This type of evaluation can be performed in the human mind or with the aid of pen and paper, and therefore falls within the mental process grouping of abstract ideas.)
The following claim elements are additional elements which, taken alone or in combination with the other elements, do not integrate the judicial exception into a practical application nor amount to significantly more than the judicial exception:
a storage device (This is a high-level recitation of generic computer components for performing the abstract idea. See MPEP 2106.05.);
at least one processing unit to execute processor-executable program code (This is a high-level recitation of generic computer components for performing the abstract idea. See MPEP 2106.05.);
receive a plurality of data samples for training a machine learning model (This limitation amounts to adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g). Receiving a plurality of data samples (i.e., mere data gathering in conjunction with an abstract idea) is directed to a well understood routine conventional activity of data transmission see MPEP 2106.05(d)(II)(i).);
receive a plurality of label examples associated with each of a plurality of ground truth labels for training the machine learning model (This limitation amounts to adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g). Receiving a plurality of examples (i.e., mere data gathering in conjunction with an abstract idea) is directed to a well understood routine conventional activity of data transmission see MPEP 2106.05(d)(II)(i).);
store the annotated data sample in the storage device (The step of “storing the annotated data sample” is merely a generic data operation that amounts to storing and retrieving information in memory, which is well-understood, routine, and conventional activity. See MPEP 2106.05(d)(II)(iv).)
train the machine learning model based on the annotated data sample associated with the plurality of data samples (This limitation constitutes mere instructions to apply the abstract idea and insignificant extra-solution activity. See MPEP 2106.05(f) and 2106.05(g).):
by an execution of the trained machine learning model (This limitation constitutes mere instructions to apply the abstract idea and insignificant extra-solution activity. See MPEP 2106.05(f) and 2106.05(g).)
Regarding claim 2, the rejection of claim 1 is incorporated herein. Further, claim 2 recites the following abstract ideas:
wherein determination of the location of the identified label example comprises determination of a start index and an end index of the identified label example within the data sample, and wherein the data sample including the identified label example is annotated with the start index and the end index (This is an abstract idea of “mental process.” The limitation recites a mental process involving determining the location of an example within a data sample by identifying its start index and end index, and annotating the data sample with that information. This type of analysis could be performed in the human mind or with pen and paper. For example, a person could read through a document, locate where a particular phrase begins or ends, note the character or word positions, and write those positions as annotations next to the example. Since this activity involves observation, judgement, and manual notation that can be carried out mentally or without the use of specialized technology, it falls within the mental processing grouping of abstract ideas.).
Regarding claim 3, the rejection of claim 1 is incorporated herein. Further, claim 3 recites the following additional elements, which taken alone or in combination with other elements, do not integrate the judicial exception into a practical application nor amount to significantly more than the judicial exception:
wherein the plurality of data samples are received via a first application programming interface (This limitation amounts to adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g). Receiving a plurality of data samples (i.e., mere data gathering in conjunction with an abstract idea) is directed to a well understood routine conventional activity of data transmission see MPEP 2106.05(d)(II)(i).),
wherein the plurality of examples associated with each of a plurality of labels are received via a second application programming interface (This limitation amounts to adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g). Receiving a plurality of data examples (i.e., mere data gathering in conjunction with an abstract idea) is directed to a well understood routine conventional activity of data transmission see MPEP 2106.05(d)(II)(i).).
application programming interface (This is a high-level recitation of generic computer components for performing the abstract idea. See MPEP 2106.05.);
Regarding claim 4, the rejection of claim 3 is incorporated herein. Further, claim 4 recites the following additional elements, which taken alone or in combination with other elements, do not integrate the judicial exception into a practical application nor amount to significantly more than the judicial exception:
wherein the plurality of data samples are received within a first file (This limitation amounts to adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g). Receiving a plurality of data examples (i.e., mere data gathering in conjunction with an abstract idea) is directed to a well understood routine conventional activity of data transmission see MPEP 2106.05(d)(II)(i).),
wherein the plurality of examples associated with each of the plurality of labels are received in a plurality of files, (This limitation amounts to adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g). Receiving a plurality of data examples (i.e., mere data gathering in conjunction with an abstract idea) is directed to a well understood routine conventional activity of data transmission see MPEP 2106.05(d)(II)(i).),
where each of the plurality of files includes the plurality of examples of only one label (This limitation amounts to adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g). Furthermore, this limitation merely applies an abstract idea, see MPEP 2106.05(f). It specifies a labeling rule without improving computer functionality or providing a technological solution, and simple constrains the format of received data.)
Regarding claim 5, the rejection of claim 4 is incorporated herein. Further, claim 5 recites the following additional elements, which taken alone or in combination with other elements, do not integrate the judicial exception into a practical application nor amount to significantly more than the judicial exception:
wherein a filename of each of the plurality of files including the plurality of examples of only one label comprises the label (This limitation amounts to adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g). Furthermore, this limitation merely applies an abstract idea, see MPEP 2106.05(f). Naming files based on their content is a labelling convention that does not improve technology or computer functionality, and simply constrains how the data is prepared.).
Regarding claim 6, the rejection of claim 1 is incorporated herein. Further, claim 6 recites the following additional elements, which taken alone or in combination with other elements, do not integrate the judicial exception into a practical application nor amount to significantly more than the judicial exception:
wherein the plurality of examples associated with each of the plurality of labels are received in a plurality of files (This limitation amounts to adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g). It also reflects a mere instruction to apply the abstract idea under MPEP 2106.05(f), by requiring the examples to be grouped into files based on labels without improving computer functionality or solving a technological problem. Receiving a labeled data in multiple files (i.e., mere data gathering in conjunction with an abstract idea) is directed to a well understood routine conventional activity of data transmission see MPEP 2106.05(d)(II)(i).).
where each of the plurality of files includes the plurality of examples of only one label (This limitation amounts to adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g). Furthermore, this limitation merely applies an abstract idea, see MPEP 2106.05(f), by imposing a labeling convention without offering any improvement to computer functionality or providing a specific technological solution.)
Regarding claim 7, the rejection of claim 6 is incorporated herein. Further, claim 7 recites the following additional elements, which taken alone or in combination with other elements, do not integrate the judicial exception into a practical application nor amount to significantly more than the judicial exception:
wherein a filename of each of the plurality of files including the plurality of examples of only one label comprises the label (This limitation amounts to adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g). Furthermore, this limitation merely applies an abstract idea, see MPEP 2106.05(f). Naming files based on their content is a labelling convention that does not improve technology or computer functionality, and simply constrains how the data is prepared.).
Regarding claim 8, the following claim elements are abstract ideas:
identifying all instances of the plurality of label examples within the data sample (This is an abstract idea of a mental process.” The limitation recites a mental process involving the recognition and identification of specific examples within a data sample. This is the type of analysis that could be practically performed in the human mind with observation and judgement. For example, a person could manually examine a set of data samples and compare each one to a known list of examples, identifying where and how often each example appears within the data. Since it involves steps that can be carried out in the human mind or with the aid of pen and paper, it falls within the mental process grouping of abstract ideas. See MPEP 2106.04(a)(2)(III).);
for each identified label example, determining an associated one of the plurality of labels and a location of the identified label example in the data sample (This is an abstract idea of a “mental process.” The limitation recites a mental process involving associating identified examples with known labels and recording their positions within a data sample. This type of cognitive activity could be performed in the human mind using observation and reasoning. For example, a person could review a document, recognize specific examples (e.g., keywords or patterns), mentally determine the corresponding label (e.g., “spam” or “not spam”), and note the location of each example within the document. Since these steps can be carried out mentally or with the aid of pen and paper, they fall within the mental process grouping of abstract ideas.),
annotating the data sample with the associated one of the plurality of labels and the location (This is an abstract idea of a “mental process.” The limitation recites a mental process involving the act of annotating a data sample by adding a corresponding label and location for each identified example. This is the type of task that can be done mentally or manually by a person. For example, a person could review a document, recognize relevant examples, determine the associated label, and then write marginal notes indicating the label and position of each example within the document. Since this activity can be performed in the human mind or with the aid of pen and paper, it falls within the mental process grouping of abstract ideas. See MPEP 2106.04(a)(2)(III).);
generating…in reply to a request from at least one of an application and a service, an inference (This is an abstract idea of a “mental process.” The limitation recites reviewing a request and determining a result based on that request. A person could review a request, and based on mental reasoning and judgement, infer what is being asked and determine an appropriate result. This type of evaluation based on mental reasoning and judgement can be performed in the human mind or with the aid of pen and paper, and therefore constitutes an abstract idea of a mental process.)
The following claim elements are additional elements which, taken alone or in combination with the other elements, do not integrate the judicial exception into a practical application nor amount to significantly more than the judicial exception:
receive a plurality of data samples for training a machine learning model and a plurality of label examples associated with each of a plurality of ground truth labels for training the machine learning model (This limitation amounts to adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g). Receiving a plurality of data samples and examples (i.e., mere data gathering in conjunction with an abstract idea) is directed to a well understood routine conventional activity of data transmission see MPEP 2106.05(d)(II)(i).);
storing the annotated data sample in the storage device (The step of “storing the annotated data sample” is merely a generic data operation that amounts to storing and retrieving information in memory, which is well-understood, routine, and conventional activity. See MPEP 2106.05(d)(II)(iv).);
training a machine learning model using the annotated data sample associated with the plurality of data samples (The step of “training” a model is merely an instruction to apply the abstract idea and does provide a meaningful limitation. See MPEP 2106.05(f).)
by an execution of the trained machine learning model (This limitation constitutes mere instructions to apply the abstract idea and insignificant extra-solution activity. See MPEP 2106.05(f) and 2106.05(g).)
Regarding claim 9, the rejection of claim 8 is incorporated herein. The claim recites similar limitations corresponding to claim 2. Therefore, the same subject matter analysis that was utilized for claim 2, as described above, is equally applicable to claim 9.
Therefore, claim 9 is ineligible.
Regarding claim 10, the rejection of claim 8 is incorporated herein. The claim recites similar limitations corresponding to claim 3. Therefore, the same subject matter analysis that was utilized for claim 3, as described above, is equally applicable to claim 9.
Therefore, claim 10 is ineligible.
Regarding claim 11, the rejection of claim 10 is incorporated herein. The claim recites similar limitations corresponding to claim 4. Therefore, the same subject matter analysis that was utilized for claim 4, as described above, is equally applicable to claim 11.
Therefore, claim 11 is ineligible.
Regarding claim 12, the rejection of claim 11 is incorporated herein. The claim recites similar limitations corresponding to claim 5. Therefore, the same subject matter analysis that was utilized for claim 5, as described above, is equally applicable to claim 12.
Therefore, claim 12 is ineligible.
Regarding claim 13, the rejection of claim 8 is incorporated herein. The claim recites similar limitations corresponding to claim 6. Therefore, the same subject matter analysis that was utilized for claim 6, as described above, is equally applicable to claim 13.
Therefore, claim 13 is ineligible.
Regarding claim 14, the rejection of claim 13 is incorporated herein. The claim recites similar limitations corresponding to claim 7. Therefore, the same subject matter analysis that was utilized for claim 7, as described above, is equally applicable to claim 14.
Therefore, claim 14 is ineligible.
Regarding claim 15, the following claim elements are abstract ideas:
identify all instances of the plurality of examples within the data sample (This is an abstract idea of a mental process.” The limitation recites a mental process involving the recognition and identification of specific examples within a data sample. This is the type of analysis that could be practically performed in the human mind with observation and judgement. For example, a person could manually examine a set of data samples and compare each one to a known list of examples, identifying where and how often each example appears within the data. Since it involves steps that can be carried out in the human mind or with the aid of pen and paper, it falls within the mental process grouping of abstract ideas. See MPEP 2106.04(a)(2)(III).);
for each identified example, determine an associated one of the plurality of labels and a location of the identified label example in the data sample (This is an abstract idea of a “mental process.” The limitation recites a mental process involving associating identified examples with known labels and recording their positions within a data sample. This type of cognitive activity could be performed in the human mind using observation and reasoning. For example, a person could review a document, recognize specific examples (e.g., keywords or patterns), mentally determine the corresponding label (e.g., “spam” or “not spam”), and note the location of each example within the document. Since these steps can be carried out mentally or with the aid of pen and paper, they fall within the mental process grouping of abstract ideas.),
annotate the data sample with the associated one of the plurality of labels and the location (This is an abstract idea of a “mental process.” The limitation recites a mental process involving the act of annotating a data sample by adding a corresponding label and location for each identified example. This is the type of task that can be done mentally or manually by a person. For example, a person could review a document, recognize relevant examples, determine the associated label, and then write marginal notes indicating the label and position of each example within the document. Since this activity can be performed in the human mind or with the aid of pen and paper, it falls within the mental process grouping of abstract ideas. See MPEP 2106.04(a)(2)(III).);
generate…in reply to a request from at least one of an application and a service, an inference (This is an abstract idea of a “mental process.” The limitation recites reviewing a request and determining a result based on that request. A person could review a request, and based on mental reasoning and judgement, infer what is being asked and determine an appropriate result. This type of evaluation based on mental reasoning and judgement can be performed in the human mind or with the aid of pen and paper, and therefore constitutes an abstract idea of a mental process.)
The following claim elements are additional elements which, taken alone or in combination with the other elements, do not integrate the judicial exception into a practical application nor amount to significantly more than the judicial exception:
A non-transitory medium storing processor-executable program code (This is a high-level recitation of generic computer components for performing the abstract idea. See MPEP 2106.05.),
program code (This is a high-level recitation of generic computer components for performing the abstract idea. See MPEP 2106.05.)
receive a plurality of data samples for training a machine learning model from a user (This limitation amounts to adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g). Receiving a plurality of data samples (i.e., mere data gathering in conjunction with an abstract idea) is directed to a well understood routine conventional activity of data transmission see MPEP 2106.05(d)(II)(i).);
receive a plurality of label examples associated with each of a plurality of ground truth labels for training the machine learning model from the user (This limitation amounts to adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g). Receiving a plurality of examples (i.e., mere data gathering in conjunction with an abstract idea) is directed to a well understood routine conventional activity of data transmission see MPEP 2106.05(d)(II)(i).);
store the annotated data sample in the storage device (The step of “storing the annotated data sample” is merely a generic data operation that amounts to storing and retrieving information in memory, which is well-understood, routine, and conventional activity. See MPEP 2106.05(d)(II)(iv).)
train the machine learning model using the annotated data sample associated with the plurality of data samples (This limitation constitutes mere instructions to apply the abstract idea and insignificant extra-solution activity. See MPEP 2106.05(f) and 2106.05(g).); and
Regarding claim 16, the rejection of claim 15 is incorporated herein. The claim recites similar limitations corresponding to claim 2. Therefore, the same subject matter analysis that was utilized for claim 2, as described above, is equally applicable to claim 16.
Therefore, claim 16 is ineligible.
Regarding claim 17, the rejection of claim 15 is incorporated herein. The claim recites similar limitations corresponding to claim 3. Therefore, the same subject matter analysis that was utilized for claim 3, as described above, is equally applicable to claim 17.
Therefore, claim 17 is ineligible.
Regarding claim 18, the rejection of claim 17 is incorporated herein. The claim recites similar limitations corresponding to claim 4 and 5. Therefore, the same subject matter analysis that was utilized for claim 4 and 5, as described above, is equally applicable to claim 18.
Therefore, claim 18 is ineligible.
Regarding claim 19, the rejection of claim 15 is incorporated herein. The claim recites similar limitations corresponding to claim 6. Therefore, the same subject matter analysis that was utilized for claim 6, as described above, is equally applicable to claim 19.
Therefore, claim 19 is ineligible.
Regarding claim 20, the rejection of claim 19 is incorporated herein. The claim recites similar limitations corresponding to claim 7. Therefore, the same subject matter analysis that was utilized for claim 7, as described above, is equally applicable to claim 20.
Therefore, claim 20 is ineligible.
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 (i.e., changing from AIA to pre-AIA ) 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)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claims 1, 3, 8, 10, 15, and 17 are rejected under 35 U.S.C. 102(a)(2) as being anticipated by Bodapati et al. (Pat. No.: US 11657307 B1 (Filed: 2019)).
Regarding claim 1, Bodapati discloses:
A system comprising: a storage device (col. 12, lines 47-51 mentions “ The code is stored on a computer-readable storage medium, for example, in the form of a computer program comprising instructions executable by one or more processors. The computer-readable storage medium is non-transitory.”); and at least one processing unit to execute processor-executable program code stored on the storage device to cause the system to (col. 12, lines 42-47 mentions “Some or all of the operations 400 are performed under the control of one or more computer systems configured with executable instructions and are implemented as code (e.g., executable instructions, one or more computer programs, or one or more applications) executing collectively on one or more processors, by hardware or combinations thereof.”):
receive a plurality of data samples for training a machine learning model (Abstract mentions “Techniques for data lake-based text generation and data augmentation for machine learning training are described.” Col. 15, lines 30-32 further mentions “The operations 600 include, at block 605, receiving, from a computing device of a user, a first plurality of documents and a first plurality of labels.”);
receive a plurality of label examples associated with each of a plurality of ground truth labels for training the machine learning model (Abstract mentions “A user-provided dataset including documents and corresponding label information” Col. 5, lines 6-10 and 15-17 further mentions “The custom model system 108, for example, may additionally or alternatively enable users to build custom text custom classifier model(s) 112 using their domain (or business) specific labels without needing to manage servers, understand ML techniques, etc…The user 109 may provide examples of text for each of the labels they want to use, and the custom model system 108 trains a custom classifier model 112 on those samples.”);
for each of the plurality of data samples (Col. 15, lines 30-32 further mentions “The operations 600 include, at block 605, receiving, from a computing device of a user, a first plurality of documents and a first plurality of labels.”): identify all instances of the plurality of label examples within the data sample (Col. 6, lines 11-20 mentions “Alternatively or additionally, when a user 109 desires to train a custom entity recognition model 114, this file (or files) may include may include a first file providing documents 124 with each document on its own line, and a second corresponding file (e.g., CSV) providing labels 126 where the labels are indicated in a first column with the corresponding entity type in a second column—e.g., “Jane Smith, MANAGER”—and the custom model system 108 may then search through the documents to find the occurrences of the labels and the locations thereof.” – all occurrences correspond to all instances.);
for each identified label example, determine an associated one of the plurality of labels and a location of the identified label example in the data sample (col 12, lines 55-64 and col. 13, lines 7-12 mentions “ a user may provide both exemplary documents as well as entity information, which could be of a variety of forms. For example, a user may provide an identifier of an entity and identifiers of the span(s) (within the document) that correspond to the entity. For example, the document “Aaron Rodgers throws 350th career touchdown pass in Packers win” may include an identifier of the span (e.g., a starting and stopping character number, a starting character number with a length value, etc.) that includes “Aaron Rodgers”…the training system 150 can create its own labels using the entity list(s) by identifying which entities exist within which provided document samples (e.g., via string matching techniques known to those of skill in the art) as well as the locations of these entities.”),
annotate the data sample with the associated one of the plurality of labels and the location (Col. 11, lines 51-58, lines 62-67, and col. 12, lines 1-6 mentions “The operations 350 include, at block 355, train a model based on document embeddings and label embeddings to project the documents and labels into a common embedding space, the model being trained to place embeddings of documents and corresponding labels close, relatively speaking, to one another. For example, in some embodiments, block 355 includes block 357A and generating a document embedding for each user-provided document…In some embodiments, block 355 includes block 357B and generating an embedding for each user-provided label, e.g., in a similar manner With these embeddings, block 355 may include, at block 357C, training a model using a loss function (e.g., a max-margin loss) to separate embeddings of non-associated documents and labels and consolidate the embeddings of associated documents and labels—e.g., the training data could be a combination of a user-provided document embedding, a user-provided label embedding, and an identifier (e.g., 0/1, T/F) of whether the document and label are associated.”)
store the annotated data sample in the storage device (Bodapati, [col. 15, lines 30-36] “The operations 600 include, at block 605, receiving, from a computing device of a user, a first plurality of documents and a first plurality of labels. The receipt may occur at an endpoint of a provider network that is associated with a storage service, which stores the first plurality of documents and the first plurality of labels at one or more storage locations within the provider network.” – teaches that the documents and associated labels used for machine learning training are stored in storage locations within a provider network, which corresponds to storing the annotated data samples in a storage device as recited in the claim.).
train the machine learning model based on the annotated data sample associated with the plurality of data samples (Bodapati, [col. 5, lines 63-67; col. 5, lines 1-10] “ As another example, when a user 109 desires to train a custom entity recognition model 114, this file (or files) may include a first file providing documents 124 with each document on its own line, and a second corresponding file (e.g., CSV) providing labels 126 having one or more of the following columns: a “file identifier” (e.g., the name of the file containing the document), a “line identifier” (e.g., the line number containing the entity, starting with line 0), a “begin offset” identifier (e.g., the character offset in the input text (relative to the beginning of the line) that shows where the entity begins, where the first character is at position 0), an “end offset” identifier (e.g., the character offset in the input text that shows where the entity ends), and/or a “type” identifier” – the documents correspond to claimed data samples, and the labels correspond to the documents constitute annotated data samples. Using those documents and corresponding labels to train the custom classifier model corresponds to training the machine learning model based on the annotated data samples associated with a plurality of data samples, as recited in the claims.):
generate, by an execution of the trained machine learning model using the annotated data sample in reply to a request from at least one of an application and a service, an inference (Bodapati, [col. 7, lines 7-15] “ the hosting system 152 of the custom model system 108 may make use of a hosting system 134 of a machine learning service 130 to deploy a model as a hosted model 136 in association with an endpoint 138 that can receive inference requests from client applications 140A and/or 140B at circle (7), provide the inference requests 160A to the associated hosted model(s) 136, and provide inference results 160B (e.g., predicted classes, predicted entities) back to applications 140A “ – receiving inference requests from client applications and executing the hosted model to produce predicted classes or entities correspond to executing the trained machine learning model in reply to a request from an application or service to generate an inference, as recited in the claim.).
Regarding claim 3, Bodapati discloses:
A system according to Claim 1, wherein the plurality of data samples are received via a first application programming interface (Col. 6, lines 21-42 mentions “the computing device 104 may issue one or more requests (e.g., API calls)… The request may also include one or more of an identifier of a storage location or locations storing the dataset 122 (e.g., an identifier of just the documents 124, an identifier of just the labels 126, an identifier associated with both the documents and labels, etc), which may identify a storage location (e.g., via a Uniform Resource Locator (URL), a bucket/folder identifier, etc.) within the provider network 100 (e.g., as offered by a storage service 116) or external to the provider network 100, a format identifier of the dataset 122, a language identifier of the language of the dataset 122 documents 124, etc. In some embodiments, the request includes the labels 126 themselves within the request, e.g., as part of an entity list for a custom entity recognition model 114.”), and wherein the plurality of examples associated with each of a plurality of labels are received via a second application programming interface (Col. 30, lines 1-6 “ the model hosting system 134 provides the user devices 702 with one or more user interfaces, command-line interfaces (CLI), application programing interfaces (API), and/or other programmatic interfaces for submitting training requests, deployment requests, and/or execution requests. In some embodiments, the user devices 702 can execute a stand-alone application that interacts with the model training system 132 and/or the model hosting system 134 for submitting training requests, deployment requests, and/or execution requests.”).
Regarding claim 8, Bodapati discloses:
A computer-implemented method comprising: receiving a plurality of data samples for training a machine learning model (Abstract mentions “Techniques for data lake-based text generation and data augmentation for machine learning training are described.” Col. 15, lines 30-32 further mentions “The operations 600 include, at block 605, receiving, from a computing device of a user, a first plurality of documents and a first plurality of labels.”) and a plurality of examples associated with each of a plurality of ground truth labels for training the machine learning model (Abstract mentions “A user-provided dataset including documents and corresponding label information” Col. 5, lines 6-10 and 15-17 further mentions “The custom model system 108, for example, may additionally or alternatively enable users to build custom text custom classifier model(s) 112 using their domain (or business) specific labels without needing to manage servers, understand ML techniques, etc…The user 109 may provide examples of text for each of the labels they want to use, and the custom model system 108 trains a custom classifier model 112 on those samples.”);
for each of the plurality of data samples (Abstract mentions “Techniques for data lake-based text generation and data augmentation for machine learning training are described.” Col. 15, lines 30-32 further mentions “The operations 600 include, at block 605, receiving, from a computing device of a user, a first plurality of documents and a first plurality of labels.”): identifying all instances of the plurality of examples within the data sample (Col. 6, lines 11-20 mentions “Alternatively or additionally, when a user 109 desires to train a custom entity recognition model 114, this file (or files) may include may include a first file providing documents 124 with each document on its own line, and a second corresponding file (e.g., CSV) providing labels 126 where the labels are indicated in a first column with the corresponding entity type in a second column—e.g., “Jane Smith, MANAGER”—and the custom model system 108 may then search through the documents to find the occurrences of the labels and the locations thereof.”);
for each identified label example, determining an associated one of the plurality of labels and a location of the identified label example in the data sample (col 12, lines 55-64 and col. 13, lines 7-12 mentions “ a user may provide both exemplary documents as well as entity information, which could be of a variety of forms. For example, a user may provide an identifier of an entity and identifiers of the span(s) (within the document) that correspond to the entity. For example, the document “Aaron Rodgers throws 350th career touchdown pass in Packers win” may include an identifier of the span (e.g., a starting and stopping character number, a starting character number with a length value, etc.) that includes “Aaron Rodgers”…the training system 150 can create its own labels using the entity list(s) by identifying which entities exist within which provided document samples (e.g., via string matching techniques known to those of skill in the art) as well as the locations of these entities.”),
annotating the data sample with the associated one of the plurality of labels and the location (Col. 11, lines 51-58, lines 62-67, and col. 12, lines 1-6 mentions “The operations 350 include, at block 355, train a model based on document embeddings and label embeddings to project the documents and labels into a common embedding space, the model being trained to place embeddings of documents and corresponding labels close, relatively speaking, to one another. For example, in some embodiments, block 355 includes block 357A and generating a document embedding for each user-provided document…In some embodiments, block 355 includes block 357B and generating an embedding for each user-provided label, e.g., in a similar manner With these embeddings, block 355 may include, at block 357C, training a model using a loss function (e.g., a max-margin loss) to separate embeddings of non-associated documents and labels and consolidate the embeddings of associated documents and labels—e.g., the training data could be a combination of a user-provided document embedding, a user-provided label embedding, and an identifier (e.g., 0/1, T/F) of whether the document and label are associated.”);
storing the annotated data sample in the storage device (Bodapati, [col. 15, lines 30-36] “The operations 600 include, at block 605, receiving, from a computing device of a user, a first plurality of documents and a first plurality of labels. The receipt may occur at an endpoint of a provider network that is associated with a storage service, which stores the first plurality of documents and the first plurality of labels at one or more storage locations within the provider network.” – teaches that the documents and associated labels used for machine learning training are stored in storage locations within a provider network, which corresponds to storing the annotated data samples in a storage device as recited in the claim.).;
training a machine learning model using the annotated data sample associated with the plurality of data samples (Col. 6, lines 59-67 and col. 7, lines 1-2 mentions “ Accordingly, the custom model system 108 in some embodiments supplements the user-provided example dataset 122 by generating potentially many additional, high-quality documents with corresponding high-quality labels to yield an augmented dataset 110 (at circle (4)) that can be used for training the custom classifier model 112 or custom entity recognition model 114 (by the training system 150 at circle (5)), which thereafter can be hosted (by the hosting system 152 at circle (6)) and used for synchronous (e.g., real-time) and/or asynchronous (e.g., batch) inference.” – this shows the annotated data (augmented dataset) is used by the training system).
generating, by an execution of the trained machine learning model in reply to a request from at least one of an application and a service, an inference (Bodapati, [col. 7, lines 7-15] “ the hosting system 152 of the custom model system 108 may make use of a hosting system 134 of a machine learning service 130 to deploy a model as a hosted model 136 in association with an endpoint 138 that can receive inference requests from client applications 140A and/or 140B at circle (7), provide the inference requests 160A to the associated hosted model(s) 136, and provide inference results 160B (e.g., predicted classes, predicted entities) back to applications 140A “ – receiving inference requests from client applications and executing the hosted model to produce predicted classes or entities correspond to executing the trained machine learning model in reply to a request from an application or service to generate an inference, as recited in the claim.).
Regarding claim 10, Bodapati discloses:
A method according to Claim 8, wherein the plurality of data samples are received via a first application programming interface (Col. 6, lines 21-42 mentions “the computing device 104 may issue one or more requests (e.g., API calls)… The request may also include one or more of an identifier of a storage location or locations storing the dataset 122 (e.g., an identifier of just the documents 124, an identifier of just the labels 126, an identifier associated with both the documents and labels, etc), which may identify a storage location (e.g., via a Uniform Resource Locator (URL), a bucket/folder identifier, etc.) within the provider network 100 (e.g., as offered by a storage service 116) or external to the provider network 100, a format identifier of the dataset 122, a language identifier of the language of the dataset 122 documents 124, etc. In some embodiments, the request includes the labels 126 themselves within the request, e.g., as part of an entity list for a custom entity recognition model 114.”), and wherein the plurality of examples associated with each of a plurality of labels are received via a second application programming interface (Col. 30, lines 1-6 “ the model hosting system 134 provides the user devices 702 with one or more user interfaces, command-line interfaces (CLI), application programing interfaces (API), and/or other programmatic interfaces for submitting training requests, deployment requests, and/or execution requests. In some embodiments, the user devices 702 can execute a stand-alone application that interacts with the model training system 132 and/or the model hosting system 134 for submitting training requests, deployment requests, and/or execution requests.”).
Regarding claim 15, Bodapati discloses:
A non-transitory medium storing processor-executable program code, the program code executable to cause a system to (col. 12, lines 42-51 mentions “ Some or all of the operations 300, 350 are performed under the control of one or more computer systems configured with executable instructions and are implemented as code (e.g., executable instructions, one or more computer programs, or one or more applications) executing collectively on one or more processors, by hardware or combinations thereof. The code is stored on a computer-readable storage medium, for example, in the form of a computer program comprising instructions executable by one or more processors. The computer-readable storage medium is non-transitory.”):
receive a plurality of data samples for training a machine learning model from a user (Abstract mentions “Techniques for data lake-based text generation and data augmentation for machine learning training are described.” Col. 15, lines 30-32 further mentions “The operations 600 include, at block 605, receiving, from a computing device of a user, a first plurality of documents and a first plurality of labels.”);
receive a plurality of label examples associated with each of a plurality of ground truth labels for training the machine learning model from the user(Abstract mentions “A user-provided dataset including documents and corresponding label information” Col. 5, lines 6-10 and 15-17 further mentions “The custom model system 108, for example, may additionally or alternatively enable users to build custom text custom classifier model(s) 112 using their domain (or business) specific labels without needing to manage servers, understand ML techniques, etc…The user 109 may provide examples of text for each of the labels they want to use, and the custom model system 108 trains a custom classifier model 112 on those samples.”);
for each of the plurality of data samples (Abstract mentions “Techniques for data lake-based text generation and data augmentation for machine learning training are described.” Col. 15, lines 30-32 further mentions “The operations 600 include, at block 605, receiving, from a computing device of a user, a first plurality of documents and a first plurality of labels.”): identifying all instances of the plurality of examples within the data sample (Col. 6, lines 11-20 mentions “Alternatively or additionally, when a user 109 desires to train a custom entity recognition model 114, this file (or files) may include may include a first file providing documents 124 with each document on its own line, and a second corresponding file (e.g., CSV) providing labels 126 where the labels are indicated in a first column with the corresponding entity type in a second column—e.g., “Jane Smith, MANAGER”—and the custom model system 108 may then search through the documents to find the occurrences of the labels and the locations thereof.”);
for each identified example, determining an associated one of the plurality of labels and a location of the identified label example in the data sample (col 12, lines 55-64 and col. 13, lines 7-12 mentions “ a user may provide both exemplary documents as well as entity information, which could be of a variety of forms. For example, a user may provide an identifier of an entity and identifiers of the span(s) (within the document) that correspond to the entity. For example, the document “Aaron Rodgers throws 350th career touchdown pass in Packers win” may include an identifier of the span (e.g., a starting and stopping character number, a starting character number with a length value, etc.) that includes “Aaron Rodgers”…the training system 150 can create its own labels using the entity list(s) by identifying which entities exist within which provided document samples (e.g., via string matching techniques known to those of skill in the art) as well as the locations of these entities.”),
annotate the data sample with the associated one of the plurality of labels and the location (Col. 11, lines 51-58, lines 62-67, and col. 12, lines 1-6 mentions “The operations 350 include, at block 355, train a model based on document embeddings and label embeddings to project the documents and labels into a common embedding space, the model being trained to place embeddings of documents and corresponding labels close, relatively speaking, to one another. For example, in some embodiments, block 355 includes block 357A and generating a document embedding for each user-provided document…In some embodiments, block 355 includes block 357B and generating an embedding for each user-provided label, e.g., in a similar manner With these embeddings, block 355 may include, at block 357C, training a model using a loss function (e.g., a max-margin loss) to separate embeddings of non-associated documents and labels and consolidate the embeddings of associated documents and labels—e.g., the training data could be a combination of a user-provided document embedding, a user-provided label embedding, and an identifier (e.g., 0/1, T/F) of whether the document and label are associated.”)
store the annotated data sample in the storage device (Bodapati, [col. 15, lines 30-36] “The operations 600 include, at block 605, receiving, from a computing device of a user, a first plurality of documents and a first plurality of labels. The receipt may occur at an endpoint of a provider network that is associated with a storage service, which stores the first plurality of documents and the first plurality of labels at one or more storage locations within the provider network.” – teaches that the documents and associated labels used for machine learning training are stored in storage locations within a provider network, which corresponds to storing the annotated data samples in a storage device as recited in the claim.);
train the machine learning model using the annotated data sample associated with the plurality of data samples (Bodapati, [col. 5, lines 63-67; col. 5, lines 1-10] “ As another example, when a user 109 desires to train a custom entity recognition model 114, this file (or files) may include a first file providing documents 124 with each document on its own line, and a second corresponding file (e.g., CSV) providing labels 126 having one or more of the following columns: a “file identifier” (e.g., the name of the file containing the document), a “line identifier” (e.g., the line number containing the entity, starting with line 0), a “begin offset” identifier (e.g., the character offset in the input text (relative to the beginning of the line) that shows where the entity begins, where the first character is at position 0), an “end offset” identifier (e.g., the character offset in the input text that shows where the entity ends), and/or a “type” identifier” – the documents correspond to claimed data samples, and the labels correspond to the documents constitute annotated data samples. Using those documents and corresponding labels to train the custom classifier model corresponds to training the machine learning model based on the annotated data samples associated with a plurality of data samples, as recited in the claims.); and
generate, by an execution of the trained machine learning model in reply to a request from at least one of an application and a service, an inference (Bodapati, [col. 7, lines 7-15] “ the hosting system 152 of the custom model system 108 may make use of a hosting system 134 of a machine learning service 130 to deploy a model as a hosted model 136 in association with an endpoint 138 that can receive inference requests from client applications 140A and/or 140B at circle (7), provide the inference requests 160A to the associated hosted model(s) 136, and provide inference results 160B (e.g., predicted classes, predicted entities) back to applications 140A “ – receiving inference requests from client applications and executing the hosted model to produce predicted classes or entities correspond to executing the trained machine learning model in reply to a request from an application or service to generate an inference, as recited in the claim.).
Regarding claim 17, Bodapati discloses:
A medium according to Claim 15, wherein the plurality of data samples are received via a first application programming interface interface (Col. 6, lines 21-42 mentions “the computing device 104 may issue one or more requests (e.g., API calls)… The request may also include one or more of an identifier of a storage location or locations storing the dataset 122 (e.g., an identifier of just the documents 124, an identifier of just the labels 126, an identifier associated with both the documents and labels, etc), which may identify a storage location (e.g., via a Uniform Resource Locator (URL), a bucket/folder identifier, etc.) within the provider network 100 (e.g., as offered by a storage service 116) or external to the provider network 100, a format identifier of the dataset 122, a language identifier of the language of the dataset 122 documents 124, etc. In some embodiments, the request includes the labels 126 themselves within the request, e.g., as part of an entity list for a custom entity recognition model 114.”), and wherein the plurality of examples associated with each of a plurality of labels are received via a second application programming interface (Col. 30, lines 1-6 “ the model hosting system 134 provides the user devices 702 with one or more user interfaces, command-line interfaces (CLI), application programing interfaces (API), and/or other programmatic interfaces for submitting training requests, deployment requests, and/or execution requests. In some embodiments, the user devices 702 can execute a stand-alone application that interacts with the model training system 132 and/or the model hosting system 134 for submitting training requests, deployment requests, and/or execution requests.”).
Claim Rejections - 35 USC § 103
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 (i.e., changing from AIA to pre-AIA ) 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 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.
Claims 2, 9, and 16 are rejected under the 35 U.S.C. 103 as being unpatentable over Bodapati et al., (Pat. No.: US 11657307 B1 (Filed: 2019)). in view of Anschel et al., (NPL: “Hybrid Search based Enhanced Named Entity Annotation Tool” (Published: February 2022)).
Regarding claim 2, Bodapati teaches all the elements of claim 1, therefore is rejected for the same reasons as those presented for claim 1. However, Bodapati does not teach:
wherein determination of the location of the identified label example comprises determination of a start index and an end index of the identified label example within the data sample, and wherein the data sample including the identified label example is annotated with the start index and the end index.
wherein determination of the location comprises determination of a start index and an end index of the example within the data sample, and wherein the data sample is annotated with the start index and the end index (Section 5 – “We design a standard annotator GUI with the following features for conventional annotation: A context menu to pick entities and select text spans from the start index to the end index.” – this demonstrates a system and method for determining the location of the text span using a start and end index. The annotation is performed within this selected span.).
Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having a combination of Bodapati and Anschel before them, to incorporate determining a start index and an end index of an example within a data sample, as taught by Anschel, into the annotation system of Bodapati. One would have been motivated to make such a combination in order to precisely identify the position of each example within the data sample, thereby enabling accurate annotation, efficient data parsing, and reliable mapping between raw data and labeled segments for use in training machine learning models.
Regarding claim 9, Bodapati teaches all the elements of claim 8, therefore is rejected for the same reasons as those presented for claim 8. The claim recites similar limitations corresponding to claim 2 and is rejected for similar reasons as claim 2 using similar teachings and rationale.
Regarding claim 16, Bodapati teaches all the elements of claim 15, therefore is rejected for the same reasons as those presented for claim 15. The claim recites similar limitations corresponding to claim 2 and is rejected for similar reasons as claim 2 using similar teachings and rationale.
Claims 4, 6, 11, 13, and 19 are rejected under the 35 U.S.C. 103 as being unpatentable over Bodapati et al., (Pat. No.: US 11657307 B1 (Filed: 2019)). in view of Siddiqui et al., (Pub. No.: US 20220276985 A1 (Filed: February 24, 2022)).
Regarding claim 4, Bodapati teaches all the elements of claim 3, therefore is rejected for the same reasons as those presented for claim 3. However, Bodapati does not teach but Bodapati in view of Siddiqui teaches the limitations:
wherein the plurality of data samples are received within a first file, and wherein the plurality of examples associated with each of the plurality of labels are received in a plurality of files, where each of the plurality of files includes the plurality of examples of only one label (paragraph [0033] mentions “For example, with some file types, metadata is stored separately from the data forming the file, and specifically, each tag associated with the data forming the file can be stored in a separate file. The result of this is that each file may actually be the aggregate of a single file of data and tens or even hundreds of files of metadata.” [0034]- “Each tag of a DICOM file must be stored as a separate nested document within the search index to maintain the searchability of that tag by its associated key, composed of DICOM standard fields called the group and element. Thus, each DICOM file may actually include at least one data document, and tens or hundreds of associated metadata documents for each tag.” – the core of the data document can exist in a “single” file. The tags functions as the label are stored in separate files (plurality of files). Finally, it shows that the purpose of the separate file is to store a single “tag” and its “key,” which are the functional equivalents of the label and associated value.
Accordingly, it would have been obvious to a person having ordinary skill in the art, before the effective filing date of the claimed invention, having a combination of Bodapati and Siddiqui before them, to incorporate receiving a plurality of data samples within a first file and receiving the plurality of examples for each label in separate files, each containing examples of only one label, as taught by Siddiqui, into the annotation system of Bodapati. One would have been motivated to make a such a combination in order to simplify data organization, reducing labeling errors, and streamline automated processing by ensuring that each file contains examples associated with a single label, thereby facilitating efficient model training and data management.
Regarding claim 6, Bodapati teaches all the elements of claim 1, therefore is rejected for the same reasons as those presented for claim 1. However, Bodapati does not teach but Bodapati in view of Siddiqui teaches the limitations::
wherein the plurality of examples associated with each of the plurality of labels are received in a plurality of files, where each of the plurality of files includes the plurality of examples of only one label (Paragraph [0033] mentions “each tag associated with the data forming the file can be stored in a separate file.” – This shows that tags, which are a form of label, are stored in separate files. Paragraph [0034] further mentions “each DICOM file may actually include at least one data document, and tens or hundreds of associated metadata documents for each tag. – plurality of files being used for tagging. Paragraph [0034] “Each tag of a DICOM file must be stored as a separate nested document within the search index to maintain the searchability of that tag by its associated key, composed of DICOM standard fields called the group and element”. – This shows that the purpose of each separate file is to store a single “tag” and its “key,” which are the functional equivalents of a label and its associated value.)
Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having a combination of Bodapati and Siddiqui before them, to incorporate receiving the plurality of examples for each label in separate files, each containing only examples of one label, as taught by Siddiqui, into the annotation system of Bodapati. One would have been motivated to make such a combination in order to improve data organization and maintain clear separation between classes, thereby reducing the risk of mislabeling, simplifying dataset management, and enabling more efficient batch processing during model training.
Regarding claim 11, Bodapati teaches all the elements of claim 10, therefore is rejected for the same reasons as those presented for claim 10. The claim recites similar limitations corresponding to claim 4 and is rejected for similar reasons as claim 4 using similar teachings and rationale.
Regarding claim 13, Bodapati teaches all the elements of claim 8, therefore is rejected for the same reasons as those presented for claim 8. The claim recites similar limitations corresponding to claim 4 and is rejected for similar reasons as claim 6 using similar teachings and rationale.
Regarding claim 19, Bodapati teaches all the elements of claim 15, therefore is rejected for the same reasons as those presented for claim 15. The claim recites similar limitations corresponding to claim 6 and is rejected for similar reasons as claim 6 using similar teachings and rationale.
Claims 5, 7, 12, 14, 18, and 20 are rejected under the 35 U.S.C. 103 as being unpatentable over Bodapati et al., (Pat. No.: US 11657307 B1 (Filed: 2019)) in view of Siddiqui et al., (Pub. No.: US 20220276985 A1 (Filed: February 24, 2022)) further in view of Seiwald et al., (Pub. No.: US 20140280188 A1 (Filed: 2013)).
Regarding claim 5, Bodapati in view of Siddiqui teaches all the elements of claim 4, therefore is rejected for the same reasons as those presented for claim 4. However, Bodapati in view of Siddiqui does not teach but Bodapati in view of Siddiqui further in view of Seiwald teaches the limitation:
wherein a filename of each of the plurality of files including the plurality of examples of only one label comprises the label (Abstract – “the system includes a tag generating configured to generate a tag based on a filename. And, the system includes a tagging module configured to associate the tag with the file.” – generates a tag (label) based on a filename and then associates the tag with that file.).
Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having a combination of Bodapati, Siddiqui, and Seiwald before them, to incorporate using a filename that includes the label for each file containing only examples of the label, as taught by Seiwald, into the annotation system of Bodapati. One would have been motivated to make such a combination in order to clearly identify the contents of each file, simplify automated dataset labeling, and reduce errors in associating examples with their corresponding labels during model training and evaluation.
Regarding claim 7, Bodapati in view of Siddiqui teaches all the elements of claim 6, therefore is rejected for the same reasons as those presented for claim 6. However, Bodapati in view of Siddiqui does not teach:
wherein a filename of each of the plurality of files including the plurality of examples of only one label comprises the label (Abstract – “the system includes a tag generating configured to generate a tag based on a filename. And, the system includes a tagging module configured to associate the tag with the file.” – generates a tag (label) based on a filename and then associates the tag with that file.).
Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having a combination of Bodapati, Siddiqui, and Seiwald before them, to incorporate using a filename that includes the label for each file containing only examples of the label, as taught by Seiwald, into the annotation system of Bodapati. One would have been motivated to make such a combination in order to clearly identify the contents of each file, simplify automated dataset labeling, and reduce errors in associating examples with their corresponding labels during model training and evaluation.
Claims 12 and 14 recite precisely the method that the systems of Claims 5 and 7, respectively, are configured to perform. Therefore, Claims 12 and 14 are rejected for the reasons set forth in the rejections of Claims 5 and 7, respectively.
Regarding Claim 18, Bodapati teaches the elements of Claim 17 (and thus the rejection of Claim 17 is incorporated). Bodapadi does not teach, but Siddiqui does teach:
wherein the plurality of examples associated with each of the plurality of labels are received in a plurality of files, where each of the plurality of files includes the plurality of examples of only one label (paragraph [0033] mentions “For example, with some file types, metadata is stored separately from the data forming the file, and specifically, each tag associated with the data forming the file can be stored in a separate file. The result of this is that each file may actually be the aggregate of a single file of data and tens or even hundreds of files of metadata.” [0034]- “Each tag of a DICOM file must be stored as a separate nested document within the search index to maintain the searchability of that tag by its associated key, composed of DICOM standard fields called the group and element. Thus, each DICOM file may actually include at least one data document, and tens or hundreds of associated metadata documents for each tag.” – the core of the data document can exist in a “single” file. The tags functions as the label are stored in separate files (plurality of files). Finally, it shows that the purpose of the separate file is to store a single “tag” and its “key,” which are the functional equivalents of the label and associated value.).
Bodapadi in view of Siddiqui does not teach, but Seiwald does teach:
wherein a filename of each of the plurality of files including the plurality of examples of only one label comprises the label. (Abstract – “the system includes a tag generating configured to generate a tag based on a filename. And, the system includes a tagging module configured to associate the tag with the file.” – generates a tag (label) based on a filename and then associates the tag with that file.).
Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having a combination of Bodapati, Siddiqui, and Seiwald before them, to incorporate using a filename that includes the label for each file containing only examples of the label, as taught by Seiwald, into the annotation system of Bodapati. One would have been motivated to make such a combination in order to clearly identify the contents of each file, simplify automated dataset labeling, and reduce errors in associating examples with their corresponding labels during model training and evaluation.
Claim 20 recites a non-transitory computer readable medium comprising the code instructions that is included in the system of Claim 7. Therefore, Claim 20 is rejected for reasons set forth in the rejection of Claim 7.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Daravanh Phakousonh whose telephone number is (571)272-6324. The examiner can normally be reached Mon - Thurs 7 AM - 5 PM, Every other Friday 7 AM - 4PM.
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, Li B Zhen can be reached at 571-272-3768. 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.
/Daravanh Phakousonh/ Examiner, Art Unit 2121
/Li B. Zhen/ Supervisory Patent Examiner, Art Unit 2121
Prosecution Insights