FINAL REJECTION, SECOND DETAILED ACTION
Status of Prosecution
The present application, 18/465,705 filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
The application was filed in the Office on September 12, 2023.
The Office mailed a first detailed action, non-final rejection on May 14, 2026.
Applicant filed remarks and arguments on August 14, 2026.
Claims 1-20 are pending. Claims 1, 12 and 18 are independent.
Status of Claims
Claims 1-4, 12-14 and 18-20 are rejected under 35 U.S.C. § 103 as being unpatentable over Mugali et al., (“Mugali”) United States Patent Application Publication 2018/0329935, published on Nov. 15, 2018 in view of Bodapati et al. (“Bodapati”), United States Patent 11,657,307 B1 published on May 23, 2023.
Claims 5, 7-8 and 15 are rejected under 35 U.S.C. § 103 as being unpatentable over Mugali in view of Bodapati and in further view of Pushkin et al. (“Pushkin”), United States Patent 11,861,039, published on Jan. 2, 2024.
Claim 6 is rejected under 35 U.S.C. § 103 as being unpatentable over Mugali in view of Bodapati and in further view of Brannon et al. (“Brannon”), United States Patent Application Publication 2022/0027479, published on Jan. 27, 2022.
Claims 9-11 and 16-17 are rejected under 35 U.S.C. § 103 as being unpatentable over Mugali in view of Bodapati and in further view of Pushkin and in further view of Veeramachaneni et al. (“Veeramachaneni”), United States Patent Application Publication 2022/0179986, published on June 9, 2022.
Response to Remarks and Arguments
Examiner thanks Applicant for the remarks and arguments.
At the outset, it appears that Applicant’s contention is that the combination of Mugali and Bodapti would not result in the claimed invention and further that a person having ordinary skill in the art would not make the combination and modification to do so.
First, Applicant contends that “Bodapati’s training methodology relies fundamentally on importing external data from a ‘data lake,’ which is entirely incompatible with, and teaches away from, the strict domain isolation and intra-domain training required by Claim 1.” (Remarks: p. 9, internal quotes are the Applicant’s). Examiner respectfully disagrees.
Examiner notes that Bodapati teaches that the sources of data may in fact be external or internal: “Thus, the custom model system 108 may retrieve any stored dataset 122 elements as shown at circle (3), which may be from a storage location within the provider network 100 or external to the provider network 100.” (Bodapati: col. 6, lines 50-54); “While the training data store 760 is depicted as being located external to the model training system 132 and the model hosting system 134, this is not meant to be limiting. For example, in some embodiments not shown, the training data store 760 is located internal to at least one of the model training system 132 or the model hosting system 134.” (Bodapati, col. 28, lines 61-67).
Examiner notes that “[T]he prior art’s mere disclosure of more than one alternative does not constitute a teaching away from any of these alternatives because such disclosure does not criticize, discredit, or otherwise discourage the solution claimed….” MPEP 2141.02 (VI); In re Fulton, 391 F.3d 1195, 1201 (Fed. Cir. 2004).
Consequently, Examiner maintains that the application of Bodapati, despite the teaching of the use of external sources, is not inapposite to the claimed invention’s purpose or functioning.
Next, Applicant contends that, “a person of ordinary skill in the art would not have been motivated to combine the references in the manner proposed by the Examiner because doing so would be contrary to the express purpose of providing isolated domains.” (Remarks: p. 10). As noted above, the characterization of Bodapati is disagreed with and the use of internal data sets is contemplated.
The arguments are not persuasive and the claims stand rejected.
Claims Rejections – 35 USC § 103
The following is a quotation of 35 U.S.C. § 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
A.
Claims 1-4, 12-14 and 18-20 are rejected under 35 U.S.C. § 103 as being unpatentable over Mugali et al., (“Mugali”) United States Patent Application Publication 2018/0329935, published on Nov. 15, 2018 in view of Bodapati et al. (“Bodapati”), United States Patent 11,657,307 B1 published on May 23, 2023.
As to Claim 1, Mugali teaches: A hosted data storage service system, comprising:
a plurality of storage domains implemented by at least one processor and at least one computer-readable storage media (Mugali: par. 0030, Fig. 1, distributed storage systems 125b), each storage domain configured to store a plurality of documents for an entity associated with the storage domain and prevent access to the storage domain by entities unassociated with the storage domain (Mugali: par. 0002, “distributed storage of network session data in hierarchical
data structures stored on multiple servers and/or physical storage devices,”; par. 0027, different users may be assigned to different hierarchical data structures; par. 0036, a rules-based logical layer may be implemented to manage access);
a plurality of machine-learned domain-specific classifiers, each machine-learned domain-specific classifier associated with a respective storage domain (Mugali: par. 0039, different classification algorithms (i.e. classifiers) may be assigned to different data partitions) and configured to generate a classification label for the plurality of documents of the entity associated with the respective storage domain (Mugali: par. 0039, machine learning algorithms are used to generate and populate different classification hierarchy data structures); and
a training system configured to generate the plurality of machine-learned domain-specific classifiers (Mugali: par. 0042, training data used to train the machine learning models).
PNG
media_image1.png
680
874
media_image1.png
Greyscale
Mugali may not explicitly teach: a training system configured to generate the plurality of machine-learned domain-specific classifiers, the training system configured to train a machine-learned domain-specific classifier for a selected storage domain using a subset of annotated documents from the selected storage domain.
Bodapati teaches in general concepts related to lake-based text generation and data augmentation for machine learning training (Bodapati: Abstract). Specifically, Bodapati teaches a training system that trains a model (i.e. classifier) based on document and label embeddings to project the documents and labels into a common embedding space (Bodapati: Fig. 3, [355], col. 11, lines 51 to 56). These techniques are useful for training a custom classifier type model (Bodapati: col. 12, line 32 to 35). A storage service of a multi-tenant provider for documents for documents associated with a first label and a second with a different one as well may be used in conjunction with the training system (Bodapati: col. 16, lines 23 to 38).
PNG
media_image2.png
782
500
media_image2.png
Greyscale
It would have been obvious to a person having ordinary skill in the art at a time before the effective filing date of the application to have modified the Mugali disclosures and teachings by implementing the training system with the annotated documents as taught and suggested by Bodapati. Such a person would have been motivated to do so with a reasonable expectation of success to allow for the different classification methods and techniques of Bodapati would benefit from better classification accuracy of annotated documents for training.
As to Claim 2, Mugali and Bodapati teach the elements of claim 1.
Bodapati further teaches: wherein the machine-learned domain-specific classifier for the selected storage domain comprises:
a similarity-based machine-learned classification model configured to generate the classification label for a selected document by embedding the selected document into an embedding space (Bodapati: Fig. 3, [355], col. 11, lines 51 to 56),
identifying a document cluster in the embedding space as a nearest match to the selected document (Bodapati: Fig. 3, [365], col. 12, lines 13 to 18, the nearby documents), and
applying an associated classification label of the document cluster to the selected document (Bodapati: Fig. 3, [370] col. 12, select documents with corresponding nearest label associated with it. Examiner asserts this would entail applying the label to the selected document).
As to Claim 3, Mugali and Bodapati teach the elements of claim 2.
Bodapati further teaches: wherein the machine-learned domain-specific classifier for the selected storage domain of the plurality of storage domains comprises: an inference-based machine-learned classification model configured to generate the classification label for the selected document (Bodapati: col. 16, lines 17 to 22, at [620], an ML model may generate an inference for a classification task).
As to Claim 4, Mugali and Bodapati teach the elements of claim 3.
Bodapati further teaches: the machine-learned domain-specific classifier for the selected storage domain comprises a similarity-based machine-learned classification model (Bodapati: Fig. 3, [355], col. 11, lines 51 to 56); and
the inference-based machine-learned classification model is trained using classifications generated by the similarity-based machine-learned classification model (Bodapati: col. 7, lines 2 to 12, the training of the custom classifier may make use of various systems and services that can receive inference requests).
As to Claim 12, it is rejected for similar reasons as claim 1. Mugali further teaches a processor and computer readable media (Mugali: pars. 0121-22).
As to Claim 13, it is rejected for similar reasons as claim 2.
As to Claim 14, it is rejected for similar reasons as claim 3.
As to Claim 18, it is rejected for similar reasons as claim 1.
As to Claim 19, Mugali and Bodapasti teach the elements of Claim 18.
Bodpasti further teaches: the machine-learned domain-specific classifier includes a similarity-based machine-learned classification mode (Bodapati: Fig. 3, [355], col. 11, lines 51 to 56); and
modifying the machine-learned domain-specific classifier based on the subset of annotated documents includes:
providing the subset of annotated documents to the similarity-based machine-learned classification model (Bodapati: Fig. 3, [355], col. 11, lines 51 to 56),;
embedding each annotated document into a representation space of the similarity-based machine-learned classification model(Bodapati: Fig. 3, [365], col. 12, lines 13 to 18, the nearby documents); and
storing document embeddings for the subset of annotated documents in the representation space for the similarity-based machine-learned classification model (Bodapati: Fig. 3, [370] col. 12, select documents with corresponding nearest label associated with it. Examiner asserts this would entail applying the label to the selected document).
As to Claim 20, Mugali and Bodapasti teach the elements of Claim 18.
Mugali and Bodpasti further teaches: the machine-learned domain-specific classifier includes an inference-based machine-learned classification model (Bodapati: col. 16, lines 17 to 22, at [620], an ML model may generate an inference for a classification task).; and
modifying the machine-learned domain-specific classifier based on the subset of annotated documents includes:
providing the subset of annotated documents to the inference-based machine-learned classification model (Bodapati: col. 7, lines 21 to 26, annotations to the data may be provided manually or automatically by labelers).
receiving a predicted classification label from the inference-based machine-learned classification model for each of the subset of annotated documents (Bodapati: col. 7, lines 1 to 10, the inferences of the classification);
determining one or more parameters of a loss function based on a difference between a predicted classification label and an annotated label for each of the subset of annotated documents (Mugali: par. 0047, classification algorithms such as L2-regulraized logistic regression, L2—loss and L1-loss may be used); and
modifying at least a portion of the inference-based machine-learned classification model based at least in part on the one or more parameters of the loss function (Examiner asserts that the use of the classification algorithms would be used to improve the classification model iteratively based on the loss function).
B.
Claims 5, 7-8 and 15 are rejected under 35 U.S.C. § 103 as being unpatentable over Mugali et al., (“Mugali”) United States Patent Application Publication 2018/0329935, published on Nov. 15, 2018 in view of Bodapati et al. (“Bodapati”), United States Patent 11,657,307 B1 published on May 23, 2023 and in further view of Pushkin et al. (“Pushkin”), United States Patent 11,861,039, published on Jan. 2, 2024.
As to Claim 5, Mugali and Bodapati teach the elements of claim 1.
Mugali and Bodapati may not explicitly teach: a heuristics engine configured to:
access the subset of annotated documents prior to the plurality of machine-learned domain-specific classifiers;
identify personal information in the subset of annotated documents;
mask the personal information in the subset of annotated documents; and
provide the subset of annotated documents including the masked personal information to the machine-learned domain-specific classifier for the selected storage domain.
Pushkin teaches in general concepts related to identifying sensitive content in data (Pushkin: Abstract). Specifically Pushkin teaches that a storage system stores various content items including documents (Pushkin: col. 17, lines 36 to 37, the storage system [330] documents [336]). Storage system may store documents that were classified with sensitive data that is identified (Puskin: col. 17, lines 40 to 44). A sensitive data discovery component may identify the location of sensitive information within the data items and then redact them and then sending this to the storage system (Pushkin: col. 18, lines 3 to 16).
It would have been obvious to a person having ordinary skill in the art at a time before the effective filing date of the application to have modified the Mugali-Bodapati combination by implementing redaction of personal information from the annotated documents for classified storage as taught and suggested by Pushkin. Such a person would have been motivated to do so with a reasonable expectation of success to protect sensitive data in a manner that is efficient (Pushkin: col. 1, lines 27 to 34).
As to Claim 7, Mugali and Bodapati teach the elements of claim 1.
Mugali and Bodapati may not explicitly teach: the classification label identifies one of a plurality of security classifications.
Pushkin teaches in general concepts related to identifying sensitive content in data (Pushkin: Abstract). Specifically Pushkin teaches that a storage system stores various content items including documents (Pushkin: col. 17, lines 36 to 37, the storage system [330] documents [336]). Storage system may store documents that were classified with sensitive data that is identified (Puskin: col. 17, lines 40 to 44). A sensitive data discovery component may identify the location of sensitive information within the data items and then redact them and then sending this to the storage system (Pushkin: col. 18, lines 3 to 16). The system may also store information regarding whether individual docuemnst were classified with sensitive data along with the documents (Puskin: col. 18, lines 46 to 50, these are security classifications).
It would have been obvious to a person having ordinary skill in the art at a time before the effective filing date of the application to have modified the Mugali-Bodapati combination by implementing identification of personal information from the annotated documents for classified storage as taught and suggested by Pushkin. Such a person would have been motivated to do so with a reasonable expectation of success to protect sensitive data in a manner that is efficient (Pushkin: col. 1, lines 27 to 34).
As to Claim 8, Mugali, Bodapati and Pushkin teach the elements of claim 7.
Pushkin further teaches: for a first storage domain and first machine-learned domain-specific classifier, the classification label identifies one of a first plurality of security classifications; and
for a second storage domain and second machine-learned domain-specific classifier, the classification label identifies one of a second plurality of security classifications, wherein at least one of the second plurality of security classifications is different from the first plurality of security classifications (Examiner asserts that there is nothing to suggest that he different classifications may be different from each other, including types of sensitive data as taught by Pushkin: col. 8, lines 20 to 26).
As to Claim 15, it is rejected for similar reasons as claim 5.
C.
Claim 6 is rejected under 35 U.S.C. § 103 as being unpatentable over Mugali et al., (“Mugali”) United States Patent Application Publication 2018/0329935, published on Nov. 15, 2018 in view of Bodapati et al. (“Bodapati”), United States Patent 11,657,307 B1 published on May 23, 2023 and in further view of Brannon et al. (“Brannon”), United States Patent Application Publication 2022/0027479, published on Jan. 27, 2022.
As to Claim 6, Mugali and Bodapati teach the elements of claim 1.
Mugali and Bodapati may not explicitly teach: each storage domain is isolated from other storage domains via one or more access restrictions.
Brannon teaches in general concepts related to a system that analyzes data assets to calculate a risk score for transferring those data assets (Brannon: Abstract). Specifically Brannon teaches that an automatic classification system may classify one or more pieces of personal information in one or more documents (Brannon: par. 0098). Data access and privacy may be addressed by using a data model particular to the data storage system, limiting who can access the storage portions (Brannon: pars. 0138-39).
It would have been obvious to a person having ordinary skill in the art at a time before the effective filing date of the application to have modified the Mugali-Bodapati disclosures and teachings by restricting access to the storage domains as taught and suggested by Brannon. Such a person would have been motivated to do so with a reasonable expectation of success to protect the data accordingly from unauthorized access, intentional or not.
D.
Claims 9-11 and 16-17 are rejected under 35 U.S.C. § 103 as being unpatentable over Mugali et al., (“Mugali”) United States Patent Application Publication 2018/0329935, published on Nov. 15, 2018 in view of Bodapati et al. (“Bodapati”), United States Patent 11,657,307 B1 published on May 23, 2023 and in further view of Pushkin et al. (“Pushkin”), United States Patent 11,861,039, published on Jan. 2, 2024 and in further view of Veeramachaneni et al. (“Veeramachaneni”), United States Patent Application Publication 2022/0179986, published on June 9, 2022.
As to Claim 9, Mugali and Bodapati teach the elements of claim 1.
Bodapati further teaches: data identifying a subset of documents to be used for training the machine-learned domain-specific classifier for the selected domain (Bodapati: col. 7, lines 27 to 30, user provided data set [122] is provided for training).
Mugali and Bodapati may not explicitly teach: data indicative of a security classification taxonomy of an entity associated with the selected domain;
data identifying one or more authorized users authorized to annotate the subset of documents according to the security classification taxonomy for generating the subset of annotated documents of the selected domain.
Pushkin teaches in general concepts related to identifying sensitive content in data (Pushkin: Abstract). Specifically Pushkin teaches that a storage system stores various content items including documents (Pushkin: col. 17, lines 36 to 37, the storage system [330] documents [336]). Storage system may store documents that were classified with sensitive data that is identified (Puskin: col. 17, lines 40 to 44). A sensitive data discovery component may identify the location of sensitive information within the data items and then redact them and then sending this to the storage system (Pushkin: col. 18, lines 3 to 16). The system may also store information regarding whether individual documents were classified with sensitive data along with the documents (Puskin: col. 18, lines 46 to 50, these are security classifications). Different security classifications for different sensitive information (i.e. a taxonomy) may be employed. (Pushkin: col. 8, lines 20 to 26).
It would have been obvious to a person having ordinary skill in the art at a time before the effective filing date of the application to have modified the Mugali-Bodapati combination by implementing identification of personal information from the annotated documents for classified storage as taught and suggested by Pushkin. Such a person would have been motivated to do so with a reasonable expectation of success to protect sensitive data in a manner that is efficient (Pushkin: col. 1, lines 27 to 34).
Mugali, Bodapati and Pushkin may not explicitly teach: a settings user interface configured to receive, from an administrator of a selected domain the data.
Veeramachaneni teaches in general concepts related to remotely managing user permissions for data objects on a file server (Veeramachaneni: Abstract). Specifically Veeramachaneni teaches that a user interface may be used to configure the permissions for different data repositories (folders for instance) (Veeramachaneni: Fig. 4).
PNG
media_image3.png
758
968
media_image3.png
Greyscale
It would have been obvious to a person having ordinary skill in the art at a time before the effective filing date of the application to have modified the Mugali-Bodapati-Pushkin combination by allowing the configuration of the different data elements as taught and suggested by Veeramachaneni. Such a person would have been motivated to do so with a reasonable expectation of success to allow for the reduction of cognitive burden on the user by allowing for the use of a graphical user interface.
As to Claim 10, Mugali, Bodapati Pushkin and Veeramachaneni teach the elements of claim 9.
Veeramachaneni further teaches: an editor user interface configured to receive, from the one or more authorized users for the selected domain: data indicative of one or more security classification labels to be applied to each of the subset of documents of the selected domain (Veeramachaneni: Fig. 4, the different typers of permissions [412]).
As to Claim 11, Mugali, Bodapati Pushkin and Veeramachaneni teach the elements of claim 10.
Veeramachaneni further teaches: the editor user interface is configured to receive, from the one or more authorized users for the selected domain: data indicative of an acceptance of or a correction to the classification label generated by the machine-learned domain-specific classifier for one or more of the plurality of documents (Veeramachaneni: Fig. 4, the user is able to review changes and thus indicate acceptance).
As to Claim 16, it is rejected for similar reasons as claim 9.
As to Claim 17, it is rejected for similar reasons as claim 10.
Conclusion
Additional relevant prior art made of the record:
Kvernik et al. (“Kvernik”), United States Patent Application Publication 2016/0203416, published on July 14, 2016 (analyzing data storage characteristics and patterns);
Patil et al. (“Patil”), United States Patent Application Publication 2015/0278764, published on Oct. 1, 2015 (classifying and storing into different work domains);
Carr et al. (“Carr”), United States Patent Application Publication 2023/0199089, published on June 22, 2023 (versioning and domain interconnection in multi-tenant systems).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JAMES T TSAI whose telephone number is (571)270-3916. The examiner can normally be reached M-F 8-5 Eastern.
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, Viker Lamardo can be reached at 571-270-5871. 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.
/JAMES T TSAI/ Primary Examiner, Art Unit 2147
Prosecution Insights