DETAILED ACTION
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA . This action is responsive to pending claims 1-10, 12-20 filed 10/10/2025 .
Claim Interpretation
The following is a quotation of 35 U.S.C. 112(f):
(f) Element in Claim for a Combination. – An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The claims in this application are given their broadest reasonable interpretation using the plain meaning of the claim language in light of the specification as it would be understood by one of ordinary skill in the art. The broadest reasonable interpretation of a claim element (also commonly referred to as a claim limitation) is limited by the description in the specification when 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is invoked.
As explained in MPEP § 2181, subsection I, claim limitations that meet the following three-prong test will be interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph:
(A) the claim limitation uses the term “means” or “step” or a term used as a substitute for “means” that is a generic placeholder (also called a nonce term or a non-structural term having no specific structural meaning) for performing the claimed function;
(B) the term “means” or “step” or the generic placeholder is modified by functional language, typically, but not always linked by the transition word “for” (e.g., “means for”) or another linking word or phrase, such as “configured to” or “so that”; and
(C) the term “means” or “step” or the generic placeholder is not modified by sufficient structure, material, or acts for performing the claimed function.
Use of the word “means” (or “step”) in a claim with functional language creates a rebuttable presumption that the claim limitation is to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites sufficient structure, material, or acts to entirely perform the recited function.
Absence of the word “means” (or “step”) in a claim creates a rebuttable presumption that the claim limitation is not to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is not interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites function without reciting sufficient structure, material or acts to entirely perform the recited function.
Claim limitations in this application that use the word “means” (or “step”) are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action. Conversely, claim limitations in this application that do not use the word “means” (or “step”) are not being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action.
This application includes one or more claim limitations that do not use the word “means,” but are nonetheless being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, because the claim limitation(s) uses a generic placeholder that is coupled with functional language without reciting sufficient structure to perform the recited function and the generic placeholder is not preceded by a structural modifier. Such claim limitation(s) is/are: user daily recorder, user profile storage unit, user data analyzer, user profile generator (claims 10), user daily recorder, data storage unit, data storage determiner (claim 12), information reproduction processor (claim 14), user profile provider (claim 16), reproduction error measurer (claim 18), data combiner, user profile combiner (claim 20).
Because this/these claim limitation(s) is/are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, it/they is/are being interpreted to cover the corresponding structure described in the specification as performing the claimed function, and equivalents thereof.
If applicant does not intend to have this/these limitation(s) interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, applicant may: (1) amend the claim limitation(s) to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph (e.g., by reciting sufficient structure to perform the claimed function); or (2) present a sufficient showing that the claim limitation(s) recite(s) sufficient structure to perform the claimed function so as to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph.
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.
Claim(s) 1-4, 6-15, 17-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to abstract idea without significantly more. We analyze the claims according to the subject matter eligibility flowchart (MPEP 2106). As all claims recite statutory categories (method, hardware apparatus) step 1 is answered affirmatively and we proceed to step 2.
Claim 1 is directed to a method of generating a profile from data, such as sensor data. Semantic and meaningful data is extracted from the collected data, and a user profile is generated and stored. However, these steps are mental processes as they involve observation and judgment with the mind (2a-1). Furthermore, the additional elements (underlined below) do not serve to integrate into a practical application (2a-2) nor constitute significantly more (2b).
In particular: a method for generating a personal profile in a user device, the method comprising:
collecting daily data of a user (Collecting data for analysis is a mental process);
extracting meaningful data from the collected daily data (Distinguishing or extracting data of interest is a mental process) wherein the meaningful data is extracted when a determination result of storing determination model is equal to or greater than a first threshold value (maintaining or considering threshold values can be performed in the mind);
extracting semantic information by analyzing the meaningful data (Discovering the meaning of data is a mental process), wherein the semantic information includes contextual meaning inherent in the data (considering a broad range of semantic information, including contextual data, may be performed in the mind);
generating a current user profile having a single vector form using the meaningful data and the semantic information (Generating a user profile, i.e., user specific data descriptive of a user as a list of numbers or attributes based on the information of interest is a mental process); and
storing the current user profile (Storage of data, such as in the mind or with pen and paper, is a mental process) in a data storage unit (This serves as instructions to implement an abstract idea on a computer and hence does not constitute an integration into a practical application (2a-2). Furthermore, the use computer data storage is well-understood, routine and conventional (WURC) and hence does not constitute significantly more (2b)).
For claim 2: wherein the generating includes:
combining the meaningful data and the semantic information (combining data is a mental process);
generating a user profile having the single vector form from the combined information using a learned network model (Generating a user profile, i.e., user specific data descriptive of a user as a list of numbers or attributes based on the information of interest is a mental process); and
generating the current user profile by combining the user profile generated using the network model and a previous user profile stored in the data storage unit immediately before (Generating a combined profile via a current and immediately prior data is a mental process).
For claim 3: further comprising reproducing the meaningful data and the semantic information using at least one information reproduction model from the current user profile (Reproducing data from a data descriptor is a mental process).
For claim 4: wherein:
the collecting includes collecting data sensed from at least one sensor (collecting data from sensors is a mental process), and
the at least one information reproduction model corresponds to the at least one sensor (The processing of data in relation to sensors is a mental process).
For claim 6: the user profile request includes a type of information representation model, and
the transmitting includes transmitting an information representation model of the type (These limitations merely recite extra-solutional data-gathering steps for transmitting or providing the data once it is constructed, and hence do not meaningfully limit the practice of the mental process (2a-1). Furthermore, the use of client-server data transmission is well-understood, routine and conventional (WURC) and hence does not constitute significantly more (2b)).
For claim 7: measuring an error of the reproduced information (Generating an error, such as between what is expected and what is actual, is a mental process).
For claim 8: learning the parameters of the at least one information reproduction model to minimize the error (Improving upon a reproduction process via updating parameters of the process is a mental process).
For claim 9: wherein the extraction of meaningful data includes:
extracting the meaningful data from the collected daily data using a storing determination model that returns a value between 0 and 1 regarding whether the data is meaningful to the user (Estimating a probability of whether data is meaningful is a mental process);
extracting event data from data determined to be meaningless in the storing determination model using an event storing model that returns a value between 0 and 1 regarding whether it is event data (Extracting an event from data based on an estimate of the probability of an event is a mental process); and
determining the event data as the meaningful data (Data determining or categorizing is a mental process).
For claim 10: an apparatus for generating a personal profile that generates a personal profile from daily life of a user, the apparatus comprising
a user profile storage unit;
a user data analyzer that analyzes the daily data of the user, and extracts semantic information (Analyzing and extracting information is a mental process); and
a user profile generator that generates a current user profile having a single vector form based on the daily data of the user and the semantic information, and stores the current user profile (Generating a user profile, i.e., user specific data descriptive of a user as a list of numbers or attributes based on the information of interest is a mental process) in the user profile storage unit. The additional elements serve as instructions to implement an abstract idea on a computer and hence does not constitute an integration into a practical application (2a-2). Furthermore, the use computer data storage is well-understood, routine and conventional (WURC) and hence does not constitute significantly more (2b)).
For claim 11: a data storage unit; and
a user daily recorder that collects daily data of the user, extracts meaningful data from the collected daily data, and stores the meaningful data (Collecting data for analysis is a mental process; Distinguishing or extracting data of interest is a mental process; storage of data, such as in the mind or with pen and paper, is a mental process) in the data storage unit,
wherein the user data analyzer analyzes meaningful data (analyzing data is a mental process) stored in the data storage unit among the daily data. The additional elements serve as instructions to implement an abstract idea on a computer and hence does not constitute an integration into a practical application (2a-2). Furthermore, the use computer data storage is well-understood, routine and conventional (WURC) and hence does not constitute significantly more (2b)).
For claim 12: the user daily recorder includes:
at least one sensor for sensing the daily data of the user; and
a data storing determiner that determines data to be stored in the data storage unit as the meaningful data from the user's daily data (Determining data as meaningful is a mental process), and
the data storing determiner includes:
a storing determination model that returns a value between 0 and 1 regarding whether a piece of input data is meaningful (Estimating a probability of whether data is meaningful is a mental process); and
an event storing model that returns a value between 0 and 1 regarding whether data determined to be meaningless in the storing determination model is event data (Extracting an event from data based on an estimate of the probability of an event is a mental process). The additional elements serve as instructions to implement an abstract idea on a computer and hence does not constitute an integration into a practical application (2a-2). Furthermore, the use computer data storage is well-understood, routine and conventional (WURC) and hence does not constitute significantly more (2b)).
For claim 13: the storing determination model is learned based on previous data provided by the user (learning from previous data is a mental process), and
the event storing model is learned based on the event data (learning from previous data is a mental process).
Claims 14-15, 17-20 recite limitations analogous to claims 3-8, 2 and are hence rejected under the same rationale.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1-4, 7-10, 12, 14-15, 18-20 are rejected under 35 U.S.C. 103 as being unpatentable over Malekzadeh ("Replacement autoencoder: A privacy-preserving algorithm for sensory data analysis", published 2/27/2018) in view of Phillips (US 20220249906 A1).
For claim 1, Malekzadeh discloses: a method for generating a personal profile in a user device, the method comprising:
collecting daily data of a user (fig.1, fig.5 contemplates multi-variate time series data from various sensors; see also §V.A, Table 1 (p.6) describing the various datasets being used as experiments comprising multi-sensor data);
extracting meaningful data from the collected daily data (§V.B, fig.5-6: data associated with inference class (i.e., activities) are extracted from data via autoencoder);
extracting semantic information by analyzing the meaningful data (ibid: semantic information such as desired, sensitive, non-nonsensitive is extracted based on classes from the user time-series data), wherein the semantic information includes contextual meaning inherent in the data (§V.B, fig.5: class sensitivity level, such as provided by the user (¶1, fig.5), constitutes semantic information derived from the data);
generating a current user profile having a single vector form using the meaningful data and the semantic information (fig.5-6: user profile vector comprising of feature vector of autoencoder (see fig. 3, §IV.A) is generated using the classes and label categories); and
storing the current user profile in a data storage unit (§III, fig.2 contemplates a mediator that would collect and store transformed data for provision to a server, hence, the user profile feature vector is stored in the mediator for further processing).
Malekzadeh does not disclose: wherein the meaningful data is extracted when a determination result of a storing determination model is equal or greater than a first threshold value.
Phillips discloses: wherein the meaningful data is extracted when a determination result of a storing determination model is equal or greater than a first threshold value (0145-146 contemplates activity threshold, with activity extraction (0147-148) depending on surpassing stillness threshold; see also activity recognition probability (0147) that would determine activity based on confidence threshold, hence, extracting meaningful activity data based on determination model operating on several thresholds, for storing the data to be used).
It would have been obvious before the effective filing date to modify the method of Malekzadeh by incorporating the user activity thresholds of Phillips. Both concern the art of human activity recognition and classification, and the incorporation would have, according to Phillips, allow better determination of free living motion data such as collected by various wearable sensors (0141).
For claim 2, Malekzadeh modified by Phillips discloses the method of claim 1, as described above. Malekzadeh modified by Phillips further discloses: wherein the generating includes:
combining the meaningful data and the semantic information (Malekzadeh §V.B: inference classes and category data is combined in the autoencoder network for feature extraction, see figs. 5-6);
generating a user profile having the single vector form from the combined information using a learned network model (figs.3, 5: the feature vector is generated via the autoencoder); and
generating the current user profile by combining the user profile generated using the network model and a previous user profile stored in the data storage unit immediately before (fig.6 shows a sliding window of a time series in order to generate a time series of user for reconstruction, hence, combining sequential time series data to generate a user profile for transmission to a server for further processing (fig.2)).
For claim 3, Malekzadeh modified by Phillips discloses the method of claim 1, as described above. Malekzadeh modified by Phillips further discloses: reproducing the meaningful data and the semantic information using at least one information reproduction model from the current user profile (fig.5-6: meaningful data and semantic information, now removed of sensitive data, is reproduced as output from feature vector).
For claim 4, Malekzadeh modified by Phillips discloses the method of claim 3, as described above. Malekzadeh further discloses: the collecting includes collecting data sensed from at least one sensor (fig.1, fig.5 contemplates multi-variate time series data from various sensors; see also §V.A, Table 1 (p.6) describing the various datasets being used as experiments comprising multi-sensor data), and
the at least one information reproduction model corresponds to the at least one sensor (§IV.A, fig.3-4: the sensor data is passed through the decoder for reproducing sensor data).
For claim 7, Malekzadeh modified by Phillips discloses the method of claim 3, as described above. Malekzadeh modified by Phillips further discloses: measuring an error of the reproduced information (§IV.B discloses reconstruction error function, see eq.3).
For claim 8, Malekzadeh modified by Phillips discloses the method of claim 7, as described above. Malekzadeh modified by Phillips further discloses: learning the parameters of the at least one information reproduction model to minimize the error (§IV.B, eq.3: error is minimized via tuning weights).
For claim 9, Malekzadeh modified by Phillips discloses the method of claim 1, as described above. Malekzadeh modified by Phillips further discloses: wherein the extraction of meaningful data includes:
extracting the meaningful data from the collected daily data using a storing determination model that returns a value between 0 and 1 regarding whether the data is meaningful to the user, wherein the data is determined as the meaningful data when the returned value is closer to 1 (Phillips 0146-152 gives overview of several techniques of extracting meaningful data from daily activity sensor data, in particular, 0150-151 discloses filtering techniques that are tied to a range comprising of series of adjacent windows, some of which do not meet the threshold for meaningful activity (such as biking), these thresholds being based on probabilities (see 0149), hence, a number between 0 and 1, hence, extracting meaningful activity data based on a probability determination model);
extracting event data from data determined to be meaningless in the storing determination model using an event storing model that returns a value between 0 and 1 regarding whether it is event data, wherein the data is determined as the event data when the returned value is closer to 1 (Phillips 0152: after determining meaninglessness, a new probabilistic categorization is made for the data, the new data corresponding to a new event data); and
determining the event data as the meaningful data (ibid: the new event data is categorized as meaningful).
For claim 10, Malekzadeh discloses: an apparatus for generating a personal profile that generates a personal profile from daily life of a user, the apparatus comprising:
a data storage unit;
a user daily recorder configured to collect daily data of a user (fig.1, fig.5 contemplates multi-variate time series data from various sensors; see also §V.A, Table 1 (p.6) describing the various datasets being used as experiments comprising multi-sensor data) and to extract meaningful data from the collected daily data (§V.B, fig.5-6: data associated with inference class (i.e., activities) are extracted from data via autoencoder);
a user data analyzer configured to analyze the meaningful data stored in the user daily recorder and to extract semantic information from the meaningful data (ibid: semantic information such as desired, sensitive, non-nonsensitive is extracted based on classes from the user time-series data), wherein the semantic information includes contextual meaning inherent in the data (§V.B, fig.5: class sensitivity level, such as provided by the user (¶1, fig.5), constitutes semantic information derived from the data);
a user profile generator configured to generate a current user profile having a single vector form using the meaningful data and the semantic information (fig.5-6: user profile vector comprising of feature vector of autoencoder (see fig. 3, §IV.A) is generated using the classes and label categories); and
a user profile storage unit configured to store the current user profile (§III, fig.2 contemplates a mediator that would collect and store transformed data for provision to a server, hence, the user profile feature vector is stored in the mediator for further processing).
Malekzadeh does not disclose: wherein the meaningful data is extracted when a determination result of a storing determination model is equal or greater than a first threshold value.
Phillips discloses: wherein the meaningful data is extracted when a determination result of a storing determination model is equal or greater than a first threshold value (0145-146 contemplates activity threshold, with activity extraction (0147-148) depending on surpassing stillness threshold; see also activity recognition probability (0147) that would determine activity based on confidence threshold, hence, extracting meaningful activity data based on determination model operating on several thresholds, for storing the data to be used).
It would have been obvious before the effective filing date to modify the method of Malekzadeh by incorporating the user activity thresholds of Phillips. Both concern the art of human activity recognition and classification, and the incorporation would have, according to Phillips, allow better determination of free living motion data such as collected by various wearable sensors (0141).
For claim 12, Malekzadeh modified by Phillips discloses the method of claim 11, as described above. Malekzadeh modified by Phillips further discloses: wherein:
the user daily recorder includes:
at least one sensor for sensing the daily data of the user (§V.A, Table 1, fig.5); and
a data storing determiner that determines data to be stored in the data storage unit as the meaningful data from the user's daily data (Malekzadeh figs. 3 showing determination of different data types for storage in the mediator unit of fig.2; Phillips 0146-152 gives overview of several techniques of extracting meaningful data from daily activity sensor data for storage as meaningful data);
the data storing determiner includes:
a storing determination model that returns a value between 0 and 1 regarding whether a piece of input data is meaningful, wherein the data is determined as the meaningful data when the returned value is closer to 1 (Phillips 0150-151 discloses filtering techniques that are tied to a range comprising of series of adjacent windows, some of which do not meet the threshold for meaningful activity (such as biking), these thresholds being based on probabilities (see 0149), hence, a number between 0 and 1, hence, extracting meaningful activity data based on a probability determination model); and
an event storing model that returns a value between 0 and 1 regarding whether data determined to be meaningless in the storing determination model is event data, wherein the data is determined as the event data when the returned value is closer to 1, and the event data is determined as meaningful data (Phillips 0152: after determining meaninglessness, a new probabilistic categorization is made for the data, the new data corresponding to a new event data).
For claim 14, Malekzadeh modified by Phillips discloses the method of claim 12, as described above. Malekzadeh further discloses: an information reproduction processor including at least one information reproduction model that reproduces the meaningful data and the semantic information from the current user profile (§IV.A, fig.3: the data and information are reproduced via the decoder).
For claim 15, Malekzadeh modified by Phillips discloses the method of claim 14, as described above. Malekzadeh further discloses: wherein:
the daily data includes data sensed from at least one sensor (fig.1, fig.5 contemplates multi-variate time series data from various sensors; see also §V.A, Table 1 (p.6) describing the various datasets being used as experiments comprising multi-sensor data), and
the at least one information reproduction model corresponds to the at least one sensor (§IV.A, fig.3-4: the sensor data is passed through the decoder for reproducing sensor data).
For claim 18, Malekzadeh modified by Phillips discloses the method of claim 14, as described above. Malekzadeh further discloses: a reproduction error measurer that measures an error of information reproduced by the at least one information reproduction model (§IV.B discloses reconstruction error function, see eq.3).
For claim 19, Malekzadeh modified by Phillips discloses the method of claim 18, as described above. Malekzadeh further discloses: wherein the information reproduction processor learns parameters of the at least one information reproduction model to minimize the error (§IV.B, eq.3: error is minimized via tuning weights).
For claim 20, Malekzadeh modified by Phillips discloses the method of claim 10, as described above. Malekzadeh further discloses: wherein the user profile generator includes:
a data combiner for combining the meaningful data and the semantic information (§IV.A, fig.3-4: the meaningful class data and the semantic category (§V.B) is combined via a for training and processing data);
a vector generation model for generating a user profile having the single vector form from the combined information (ibid: the feature vector is generated via the autoencoder); and
a user profile combiner for generating the current user profile by combining the user profile generated from the vector generation model and a previous user profile stored in the data storage unit immediately before (fig.6 shows a sliding window of a time series in order to generate a time series of user for reconstruction, hence, combining sequential time series data to generate a user profile for transmission to a server for further processing (fig.2)).
Claim(s) 5, 16 are rejected under 35 U.S.C. 103 as being unpatentable over Malekzadeh ("Replacement autoencoder: A privacy-preserving algorithm for sensory data analysis", published 2/27/2018) in view of Phillips (US 20220249906 A1) in view of Li ("Edge AI: On-demand accelerating deep neural network inference via edge computing", published 2019).
For claim 5, Malekzadeh modified by Phillips discloses the method of claim 3, as described above. Malekzadeh further discloses: receiving a user profile request from a server providing a service (§III, fig.2(b) receiving desired information from server).
Malekzadeh modified by Phillips does not disclose: transmitting the user profile and the at least one information reproduction model to the server according to the user profile request.
Li discloses: transmitting the user profile and the at least one information reproduction model to the server (fig.5 discloses a split inference technique where a deployed neural network is transmitted from a deployment device to an edge server for co-inference during deployment. Hence, combination with the autoencoder networks of Malekzadeh modified by Phillips disclosing a technique wherein user profile data comprising an intermediate state is transmitted during operation and the later parts of the autoencoder network are transmitted to a server for generating inferences, the combination with Malekzadeh (§III, fig.2(b) disclosing such a deployment occurring in response to a service request).
It would have been obvious before the effective filing date to modify the method of Malekzadeh modified by Phillips by incorporating the co-inference technique of Li. Both concern the art of neural network optimization, and the incorporation would have, according to Li, optimize accuracy and latency in mobile ANN applications (§1 ¶-last-2 (“To summarize …)).
For claim 16, Malekzadeh modified by Phillips discloses the method of claim 14, as described above. Malekzadeh further discloses: a user profile provider that provides to the server (§III, fig.2(d)) in response to a user profile request from a server providing a service (§III, fig.2(b) receiving desired information from server).
Malekzadeh modified by Phillips does not disclose: wherein the provides includes the user profile and the at least one information reproduction model.
Li discloses: wherein the provides includes the user profile and the at least one information reproduction model to the server (fig.5 discloses a split inference technique where a deployed neural network is transmitted from a deployment device to an edge server for co-inference during deployment. Hence, combination with the autoencoder networks of Malekzadeh modified by Phillips disclosing a technique wherein user profile data comprising an intermediate state is transmitted during operation and the later parts of the autoencoder network are transmitted to a server for generating inferences, the combination with Malekzadeh (§III, fig.2(b) disclosing such a deployment occurring in response to a service request).
It would have been obvious before the effective filing date to modify the method of Malekzadeh modified by Phillips by incorporating the co-inference technique of Li. Both concern the art of neural network optimization, and the incorporation would have, according to Li, optimize accuracy and latency in mobile ANN applications (§1 ¶-last-2 (“To summarize …)).
Claim(s) 6, 17 are rejected under 35 U.S.C. 103 as being unpatentable over Malekzadeh ("Replacement autoencoder: A privacy-preserving algorithm for sensory data analysis", published 2/27/2018) in view of Phillips (US 20220249906 A1) in view of Li ("Edge AI: On-demand accelerating deep neural network inference via edge computing", published 2019) in view of Zhou (US 20200204643 A1).
For claim 6, Malekzadeh modified by Phillips modified by Li discloses the method of claim 5, as described above. Malekzadeh does not disclose the limitations of claim 6. Zhou discloses: the user profile request includes a type of information representation model (fig.6:602, 0112-114), and
the transmitting includes transmitting an information representation model of the type (fig.6:603-604, 0115-120).
It would have been obvious before the effective filing date to modify the method of Malekzadeh by incorporating the user profile typing technique of Zhou. Both concern the art of user profile generation and management, and the incorporation would have, according to Zhou, allow for a plurality of types of user profile according to application needs (0113, 0118-119), improve data security (0003).
For claim 17, Malekzadeh modified by Phillips modified by Li discloses the method of claim 16, as described above. Malekzadeh does not disclose the limitations of claim 17. Zhou discloses: the user profile request includes a type of information representation model (fig.6:602, 0112-114), and
the user profile provider provides the type of information reproduction model (fig.6:603-604, 0115-120).
It would have been obvious before the effective filing date to modify the method of Malekzadeh by incorporating the user profile typing technique of Zhou. Both concern the art of user profile generation and management, and the incorporation would have, according to Zhou, allow for a plurality of types of user profile according to application needs (0113, 0118-119), improve data security (0003).
Claim(s) 13 are rejected under 35 U.S.C. 103 as being unpatentable over Malekzadeh ("Replacement autoencoder: A privacy-preserving algorithm for sensory data analysis", published 2/27/2018) in view of Phillips (US 20220249906 A1) in view of Qian (US 20220269816 A1).
For claim 13, Malekzadeh modified by Phillips discloses the method of claim 12, as described above. Malekzadeh modified by Phillips does not disclose the limitations of claim 13.
Qian discloses: a model is learned based on previous data provided by the user (c.8 ¶3: training models via altered user data; c.7 ¶2: training models on user data). Hence, combination with Qian yielding a technique wherein the stored determination models and the event storing models for classifying user activity data are trained on user data.
It would have been obvious before the effective filing date to modify the method of Malekzadeh modified by Phillips by incorporating the user data training technique of Qian. Both concern the art of user data management for machine learning, and the incorporation would have, according to Qian, allow increases in personalization based on user data while preventing privacy violations (c.1: Background).
Response to Arguments
Applicant’s arguments have been fully considered. In the remarks, Applicant argued:
1. The amended claims are integrated into a practical application, as it recites limitations directed to the selecting and storing of meaningful data above a threshold, and the combining of meaningful with semantic information to generate a user profile in vector form.
Examiner respectfully disagrees for the reasons given above in the rejection, i.e., as all these steps may be performed purely in the mind.
2. Regarding the prior art claims, the art of record and Malekzadeh in particular does not disclose the extracting of meaningful data and semantic information bases on a threshold.
Applicant’s argument is moot in view of newly cited art.
3. Furthermore, the prior art is silent as to a single-vector form, as Malekzadeh merely compresses the input data into a latent feature vector, nor does it disclose storing such a profile for future use.
Examiner respectfully disagrees. The latent representation in an autoencoder would comprise a single vector form, as contemplated in Malekzadeh fig.3. Furthermore, such a latent vector would need to be stored at least for passage onto later stages of the neural net.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Haerterich (US 20220070150 A1) discloses an technique of privatizing user mobile data via an autoencoder.
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 extension fee 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 date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to LIANG LI whose telephone number is (303)297-4263. The examiner can normally be reached Mon-Fri 9-12p, 3-11p MT (11-2p, 5-1a ET).
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor Jennifer Welch can be reached on (571)272-7212. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from Patent Center and the Private Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from Patent Center or Private PAIR. Status information for unpublished applications is available through Patent Center or Private PAIR to authorized users only. Should you have questions about access to Patent Center or the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free).
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. The examiner is available for interviews Mon-Fri 6-11a, 2-7p MT (8-1p, 4-9p ET).
/LIANG LI/
Primary examiner AU 2143