DETAILED ACTION
Authorization for Internet Communications
The examiner encourages Applicant to submit an authorization to communicate with the examiner via the Internet by making the following statement (from MPEP 502.03):
“Recognizing that Internet communications are not secure, I hereby authorize the USPTO to communicate with the undersigned and practitioners in accordance with 37 CFR 1.33 and 37 CFR 1.34 concerning any subject matter of this application by video conferencing, instant messaging, or electronic mail. I understand that a copy of these communications will be made of record in the application file.”
Please note that the above statement can only be submitted via Central Fax (not Examiner's Fax), Regular postal mail, or EFS Web using PTO/SB/439.
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 .
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 05/30/2025, 08/28/2025 and 12/22/2025 are being considered by the examiner.
Claim Objections
Claims 17 - 20 are objected to because of the following informalities:
Regarding claim 17; the limitation “carrying instructions” should apparently be -- having instructions --.
Claims 18 – 20 are dependent claims and thus also objected.
Appropriate correction is required.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1 – 20 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
The term “most similar” in claims 1 – 2, 9 – 10 and 17 – 20 is a relative term which renders the claim indefinite. The term “most similar” is not defined by the claim, the specification does not provide a standard for ascertaining the requisite degree, and one of ordinary skill in the art would not be reasonably apprised of the scope of the invention because although, the specification discusses examples involving k-nearest neighbor models, the claims are not limited to any particular distance function or similarity computation. The specification likewise fails to identify an objective standard that would inform one of ordinary skill in the art which user profiles qualify as being “most similar” for purposes of the claimed invention. Therefore, the clam fails to provide objective boundaries for this limitation.
Claims 2 – 8, 10 – 16 and 18 – 20 are dependent claims and thus also rejected.
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)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claim(s) 1 – 4, 9 – 12 and 17 - 20 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by the prior art of record, Schoppmann (“Private Nearest Neighbors Classification in Federated Database”) (hereinafter “Schoppmann”) (submitted by the applicant via IDS filed 05/30/2025).
Schoppmann discloses;
Regarding claim 1, a computer-implemented method comprising:
receiving, by a first computing system of a plurality of multi-party computation (MPC) systems, an inference request comprising a first share of a given user profile comprising data about a given user [i.e., see section 2: “Distributed k-NN Classification: This section presents the difference steps required to implement a distributed k-NN functionality and provides an intuitive description of our privacy requirements…A distributed k-NN classifier allows a client holding an unlabeled example x to interact with the servers in order to obtain a label for x”; Section 2. Document Classification: “While in some cases, feature extraction can be done locally by the parties, it depends on the whole dataset for TF-IDF features. We present a protocol for extracting these privately in Section 3. Note that this step has to be performed once for the entire distributed dataset Z and is amortized over many classification queries. Similarity computation is typically the most expensive step of a k-NN classification since it scales with m and needs to be performed for every query”;)];
identifying a number k of nearest neighbor user profiles that are determined to be most similar to the given user profile among a plurality of user profiles, wherein k is a positive integer, and wherein each user profile comprises a respective label indicating one or more user groups to which the user profile is assigned (see Section 2: “Distributed k-NN Classification:… A distributed k-NN classifier allows a client holding an unlabeled example x to interact with the servers in order to obtain a label for x. In particular, the following steps need to be executed: (1) for j..compute the similarity score sj(x) between x and the jth example in the dataset”;);
executing a secure MVPC protocol with one or more second computing systems of the plurality of MPC systems to obtain, for each of the k nearest neighbor profiles, a first secret share of an inference result based on the respective label for each of the k nearest neighbor user profiles, wherein the inference result for each of the k nearest neighbor user profiles indicates whether the given user corresponding to the k nearest neighbor user profile is to be added to a given user group indicated by the label of the k nearest neighbor user profile [i.e., (see Section 3: “Suppose the input dataset is distributed among n parties…; Section 3 – TF-IDF Features: “Our alternative approach is to compute the IDF coefficients in a differentially private way and release them to the parties, who can then locally computer the TF-IDF representation of each document in their respective datasets”;]; and
sending, by the first computing system and to a client device, the first share of the inference result for one or more of the k nearest neighbor user profiles [i.e., (see Section 4-Additive Secret Sharing: “Our secure sparse multiplication protocol (Figure 1) will produce matrices C1, C2 that are chosen uniformly at random under the constraint that C1 + C2 = AB. Since Party I obtains only Ci, and each share individually cannot be distinguished from a random matrix, this does not leak any information about C = AB. Thus, the similarity scores computed by out protocol are not known by a single party, but secret-shared between the holders of both inputs to a generic MPC protocol for selecting the k most similar documents, as described in Section 2. This technique is generally used to hide the values of intermediate results, and it’s the basis of various MPC protocols…”].
Regarding claim 2, the computer-implemented method of claim 1, wherein identifying the number k of nearest neighbor user profiles that are determined to be most similar to the given user profile among a plurality of user profiles comprising performing operations of the secure MPC protocol in collaboration with the one or more second computing systems using secret shares of the plurality of user profiles to identify the k nearest neighbor user profiles [i.e., “our secure sparse multiplication protocol (Figure 1) will produce matrices C1, C2 that are chosen uniformly at random under the constraint that C1 + C2 = AB. Since Party i obtains only ci, and each share individually cannot be distinguished from a random matrix, this does not leak any information about C = AB. Thus, the similarity scores computed by our protocol are not known by a single party, but secret-shared between the holders of both input document vectors. The shares are then used as inputs to a generic MPC protocol for selecting the k most similar documents, as described in Section 2. This technique is generally used to hide the values of intermediate results, and it’s the basis of various MPC protocols…we will refer to values shared in this way as additively shared and call the individual parts additive shares”;)].
Regarding claim 3, the computer-implemented method of claim 1, wherein each user group is a user interest group that includes members determined to be interested in a topic [i.e., “our secure sparse multiplication protocol (Figure 1) will produce matrices C1, C2 that are chosen uniformly at random under the constraint that C1 + C2 = AB. Since Party i obtains only ci, and each share individually cannot be distinguished from a random matrix, this does not leak any information about C = AB. Thus, the similarity scores computed by our protocol are not known by a single party, but secret-shared between the holders of both input document vectors. The shares are then used as inputs to a generic MPC protocol for selecting the k most similar documents, as described in Section 2. This technique is generally used to hide the values of intermediate results, and it’s the basis of various MPC protocols…we will refer to values shared in this way as additively shared and call the individual parts additive shares”;)].
Regarding claim 4, the computer-implemented method of claim 1, wherein each second computing system sends, to the client device, a respective second share of the inference result for each of the one or more k nearest neighbor user profiles [i.e., “A key observation about the utility of Algorithm 1 is that in a corpus of documents Z the distribution of the values of…This means there are few very frequent words and lots of infrequent words, which justifies assigning a default value c0 to the words which are not selected by the top-L selection provided by the exponential mechanism. Our experimental evaluation (Section 5) shows that this leads to a small accuracy loss when the privatized IDFs are used for k-NN document classification”; Section 2-Document Classification: “At a high-level, a distributed K-NN classification protocol like the one described above must implement three functionalities: feature extraction, similarity computation and top-k selection. The running example we use throughout the paper is document classification with cosine similarity between TF-IDF representations. This problem illustrates all the challenges involved in a generic k-NN implementation and allow us to focus on scalability by tailoring our MPC protocols to a particular application. While in some cases, feature extraction can be done locally by the parties, it depends on the whole dataset for TF-IDF features. We present a protocol for extracting these privately in Section 3. Note that this step has to be performed once for the entire distributed dataset Z and is amortized over many classification queries. Similarity computation is typically the most expensive step of a k-NN classification since it scales with m and needs to be performed for every query. In section 4 we present a protocol for scoring that can leverage sparsity in the underlying feature representation and an perform many queries in parallel. For top-k selection in a distributed setting with privacy requirements, we can use a generic MPC framework. In our experiments, we show that this last phase contributes little to the overall running time of our classification protocol”)].
Regarding claim 9, a system comprising:
one or more processors of a first computing system; and
one or more storage devices storing instructions that, when executed by the one or more processors, cause the one or more processors to perform operations comprising:
receiving, by the first computing system of a plurality of multi-party computation(MPC) systems, an inference request comprising a first share of a given user profile [i.e., see section 2: “Distributed k-NN Classification: This section presents the difference steps required to implement a distributed k-NN functionality and provides an intuitive description of our privacy requirements…A distributed k-NN classifier allows a client holding an unlabeled example x to interact with the servers in order to obtain a label for x”; Section 2. Document Classification: “While in some cases, feature extraction can be done locally by the parties, it depends on the whole dataset for TF-IDF features. We present a protocol for extracting these privately in Section 3. Note that this step has to be performed once for the entire distributed dataset Z and is amortized over many classification queries. Similarity computation is typically the most expensive step of a k-NN classification since it scales with m and needs to be performed for every query”;)];
identifying a number k of nearest neighbor user profiles that are determined to be most similar to the given user profile among a plurality of user profiles, wherein k is a positive integer, and wherein each user profile comprises a respective label indicating one or more user groups to which the user profile is assigned (see Section 2: “Distributed k-NN Classification:… A distributed k-NN classifier allows a client holding an unlabeled example x to interact with the servers in order to obtain a label for x. In particular, the following steps need to be executed: (1) for j..compute the similarity score sj(x) between x and the jth example in the dataset”;):
executing a secure MVPC protocol with one or more second computing systems of the plurality of MPC systems to obtain, for each of the k nearest neighbor profiles, a first secret share of an inference result based on the respective label for each of the k nearest neighbor user profiles, wherein the inference result for each of the k nearest neighbor user profiles indicates whether the given user corresponding to the k nearest neighbor user profile is to be added to a given user group indicated by the label of the k nearest neighbor user profile [i.e., (see Section 3: “Suppose the input dataset is distributed among n parties…; Section 3 – TF-IDF Features: “Our alternative approach is to compute the IDF coefficients in a differentially private way and release them to the parties, who can then locally computer the TF-IDF representation of each document in their respective datasets”;]; and
sending, by the first computing system and to a client device, the first share of the inference result for one or more of the k nearest neighbor user profiles [i.e., (see Section 4-Additive Secret Sharing: “Our secure sparse multiplication protocol (Figure 1) will produce matrices C1, C2 that are chosen uniformly at random under the constraint that C1 + C2 = AB. Since Party I obtains only Ci, and each share individually cannot be distinguished from a random matrix, this does not leak any information about C = AB. Thus, the similarity scores computed by out protocol are not known by a single party, but secret-shared between the holders of both inputs to a generic MPC protocol for selecting the k most similar documents, as described in Section 2. This technique is generally used to hide the values of intermediate results, and it’s the basis of various MPC protocols…”].
Regarding claim 10, the system of claim 9, wherein identifying the number k of nearest neighbor user profiles that are determined to be most similar to the given user profile among a plurality of user profiles comprising performing operations of the secure MPC protocol in collaboration with the one or more second computing systems using secret shares of the plurality of user profiles to identify the k nearest neighbor user profiles [i.e., “our secure sparse multiplication protocol (Figure 1) will produce matrices C1, C2 that are chosen uniformly at random under the constraint that C1 + C2 = AB. Since Party i obtains only ci, and each share individually cannot be distinguished from a random matrix, this does not leak any information about C = AB. Thus, the similarity scores computed by our protocol are not known by a single party, but secret-shared between the holders of both input document vectors. The shares are then used as inputs to a generic MPC protocol for selecting the k most similar documents, as described in Section 2. This technique is generally used to hide the values of intermediate results, and it’s the basis of various MPC protocols…we will refer to values shared in this way as additively shared and call the individual parts additive shares”;)].
Regarding claim 11, the system of claim 9, wherein each user group is a user interest group that includes members determined to be interested in a topic [i.e., “our secure sparse multiplication protocol (Figure 1) will produce matrices C1, C2 that are chosen uniformly at random under the constraint that C1 + C2 = AB. Since Party i obtains only ci, and each share individually cannot be distinguished from a random matrix, this does not leak any information about C = AB. Thus, the similarity scores computed by our protocol are not known by a single party, but secret-shared between the holders of both input document vectors. The shares are then used as inputs to a generic MPC protocol for selecting the k most similar documents, as described in Section 2. This technique is generally used to hide the values of intermediate results, and it’s the basis of various MPC protocols…we will refer to values shared in this way as additively shared and call the individual parts additive shares”;)].
Regarding claim 12, the system of claim 9, wherein each second computing system sends, to the client device, a respective second share of the inference result for each of the one or more k nearest neighbor user profiles [i.e., “A key observation about the utility of Algorithm 1 is that in a corpus of documents Z the distribution of the values of…This means there are few very frequent words and lots of infrequent words, which justifies assigning a default value c0 to the words which are not selected by the top-L selection provided by the exponential mechanism. Our experimental evaluation (Section 5) shows that this leads to a small accuracy loss when the privatized IDFs are used for k-NN document classification”; Section 2-Document Classification: “At a high-level, a distributed K-NN classification protocol like the one described above must implement three functionalities: feature extraction, similarity computation and top-k selection. The running example we use throughout the paper is document classification with cosine similarity between TF-IDF representations. This problem illustrates all the challenges involved in a generic k-NN implementation and allow us to focus on scalability by tailoring our MPC protocols to a particular application. While in some cases, feature extraction can be done locally by the parties, it depends on the whole dataset for TF-IDF features. We present a protocol for extracting these privately in Section 3. Note that this step has to be performed once for the entire distributed dataset Z and is amortized over many classification queries. Similarity computation is typically the most expensive step of a k-NN classification since it scales with m and needs to be performed for every query. In section 4 we present a protocol for scoring that can leverage sparsity in the underlying feature representation and an perform many queries in parallel. For top-k selection in a distributed setting with privacy requirements, we can use a generic MPC framework. In our experiments, we show that this last phase contributes little to the overall running time of our classification protocol”)].
Regarding claim 17, a non-transitory computer readable storage medium carrying instructions that, when executed by one or more processors of a first computing system, cause the one or more processors to perform operations comprising:
receiving, by the first computing system of a plurality of multi-party computation (MPC) systems, an inference request comprising a first share of a given user profile [i.e., see section 2: “Distributed k-NN Classification: This section presents the difference steps required to implement a distributed k-NN functionality and provides an intuitive description of our privacy requirements…A distributed k-NN classifier allows a client holding an unlabeled example x to interact with the servers in order to obtain a label for x”; Section 2. Document Classification: “While in some cases, feature extraction can be done locally by the parties, it depends on the whole dataset for TF-IDF features. We present a protocol for extracting these privately in Section 3. Note that this step has to be performed once for the entire distributed dataset Z and is amortized over many classification queries. Similarity computation is typically the most expensive step of a k-NN classification since it scales with m and needs to be performed for every query”;)];
identifying a number k of nearest neighbor user profiles that are determined to be most similar to the given user profile among a plurality of user profiles, wherein k is a positive integer, and wherein each user profile comprises a respective label indicating one or more user groups to which the user profile is assigned (see Section 2: “Distributed k-NN Classification:… A distributed k-NN classifier allows a client holding an unlabeled example x to interact with the servers in order to obtain a label for x. In particular, the following steps need to be executed: (1) for j..compute the similarity score sj(x) between x and the jth example in the dataset”;);
executing a secure MVPC protocol with one or more second computing systems of the plurality of MPC systems to obtain, for each of the k nearest neighbor profiles, a first secret share of an inference result based on the respective label for each of the k nearest neighbor user profiles, wherein the inference result for each of the k nearest neighbor user profiles indicates whether the given user corresponding to the k nearest neighbor user profile is to be added to a given user group indicated by the label of the k nearest neighbor user profile [i.e., (see Section 3: “Suppose the input dataset is distributed among n parties…; Section 3 – TF-IDF Features: “Our alternative approach is to compute the IDF coefficients in a differentially private way and release them to the parties, who can then locally computer the TF-IDF representation of each document in their respective datasets”;]; and
sending, by the first computing system and to a client device, the first share of the inference result for one or more of the k nearest neighbor user profiles [i.e., (see Section 4-Additive Secret Sharing: “Our secure sparse multiplication protocol (Figure 1) will produce matrices C1, C2 that are chosen uniformly at random under the constraint that C1 + C2 = AB. Since Party I obtains only Ci, and each share individually cannot be distinguished from a random matrix, this does not leak any information about C = AB. Thus, the similarity scores computed by out protocol are not known by a single party, but secret-shared between the holders of both inputs to a generic MPC protocol for selecting the k most similar documents, as described in Section 2. This technique is generally used to hide the values of intermediate results, and it’s the basis of various MPC protocols…”].
Regarding claim 18, the non-transitory computer-readable storage medium of claim 17, wherein identifying the number k of nearest neighbor user profiles that are determined to be most similar to the given user profile among a plurality of user profiles comprising performing operations of the secure MPC protocol in collaboration with the one or more second computing systems using secret shares of the plurality of user profiles to identify the k nearest neighbor user profiles [i.e., “our secure sparse multiplication protocol (Figure 1) will produce matrices C1, C2 that are chosen uniformly at random under the constraint that C1 + C2 = AB. Since Party i obtains only ci, and each share individually cannot be distinguished from a random matrix, this does not leak any information about C = AB. Thus, the similarity scores computed by our protocol are not known by a single party, but secret-shared between the holders of both input document vectors. The shares are then used as inputs to a generic MPC protocol for selecting the k most similar documents, as described in Section 2. This technique is generally used to hide the values of intermediate results, and it’s the basis of various MPC protocols…we will refer to values shared in this way as additively shared and call the individual parts additive shares”;)].
Regarding claim 19, the non-transitory computer-readable storage medium of claim 17, wherein each user group is a user interest group that includes members determined to be interested in a topic [i.e., “our secure sparse multiplication protocol (Figure 1) will produce matrices C1, C2 that are chosen uniformly at random under the constraint that C1 + C2 = AB. Since Party i obtains only ci, and each share individually cannot be distinguished from a random matrix, this does not leak any information about C = AB. Thus, the similarity scores computed by our protocol are not known by a single party, but secret-shared between the holders of both input document vectors. The shares are then used as inputs to a generic MPC protocol for selecting the k most similar documents, as described in Section 2. This technique is generally used to hide the values of intermediate results, and it’s the basis of various MPC protocols…we will refer to values shared in this way as additively shared and call the individual parts additive shares”;)].
Regarding claim 20, the non-transitory computer-readable storage medium of claim 17, wherein each second computing system sends, to the client device, a respective second share of the inference result for each of the one or more k nearest neighbor user profiles [i.e., “A key observation about the utility of Algorithm 1 is that in a corpus of documents Z the distribution of the values of…This means there are few very frequent words and lots of infrequent words, which justifies assigning a default value c0 to the words which are not selected by the top-L selection provided by the exponential mechanism. Our experimental evaluation (Section 5) shows that this leads to a small accuracy loss when the privatized IDFs are used for k-NN document classification”; Section 2-Document Classification: “At a high-level, a distributed K-NN classification protocol like the one described above must implement three functionalities: feature extraction, similarity computation and top-k selection. The running example we use throughout the paper is document classification with cosine similarity between TF-IDF representations. This problem illustrates all the challenges involved in a generic k-NN implementation and allow us to focus on scalability by tailoring our MPC protocols to a particular application. While in some cases, feature extraction can be done locally by the parties, it depends on the whole dataset for TF-IDF features. We present a protocol for extracting these privately in Section 3. Note that this step has to be performed once for the entire distributed dataset Z and is amortized over many classification queries. Similarity computation is typically the most expensive step of a k-NN classification since it scales with m and needs to be performed for every query. In section 4 we present a protocol for scoring that can leverage sparsity in the underlying feature representation and an perform many queries in parallel. For top-k selection in a distributed setting with privacy requirements, we can use a generic MPC framework. In our experiments, we show that this last phase contributes little to the overall running time of our classification protocol”)].
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Wang (US2016/0283735 A1) discloses a computer program product for generating a classification model using original data that is sensitive or private to a data owner, the computer program product comprising a tangible storage medium readable by a processing circuit and storing instructions run by the processing circuit for performing a method, the method comprising: receiving, from one or more first entities, a masked data set, each data set from an entity having masked data corresponding to the original sensitive data, the masked data set further including a masked feature label set for use in classifying the masked data contents; forming a shared data collection of said masked data and the masked feature label sets received from said first entities; and training, by a second entity, a classification model from the shared masked data and feature label sets, said model being a classification model configured to classify original sensitive data contained in masked data sets received from said entities, wherein the classification model learned from the shared masked data and feature label sets is the same as the model learned from the original sensitive data.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SYED A RONI whose telephone number is (571)270-7806. The examiner can normally be reached M-F 9:00-5:00 pm (EST).
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Jeffrey L Nickerson can be reached at (469) 295-9235. 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.
/SYED A RONI/Primary Examiner, Art Unit 2432
Prosecution Insights