Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Arguments
Regarding the rejection of amended claims as judicial exceptions to 35 U.S.C. 101, Applicant’s arguments are persuasive and the rejections are withdrawn.
Regarding the rejection of amended claims under 35 U.S.C. 103, Applicant submits that:
“The Examiner erroneously equates Limasanches' disclosure of selecting an existing model with the claimed step of selecting an untrained learning algorithm. Limasanches is fundamentally a system for searching and ‘fine-tuning (retraining)’ a trained learning model that already exists in a database. It utilizes a ‘task description’ and ‘characteristic amount vectors’ to find the most similar completed model to adapt to a new user's data.” (pg. 8 of remarks)
“Because Limasanches relies on fine-tuning an existing model, it is constrained by the architecture and original training of that model. The claimed invention, by selecting the optimal untrained algorithm from a specific group, allows for the creation of a model that is technically optimized for the unique noise and structural properties of the target measurement data (e. g. , X-ray or CT images) from the very start.” (pg. 10)
Examiner respectfully disagrees, and finds that the descriptions given in Applicant’s remarks are not represented in the claims. The relevant claim language is:
a preparation step of preparing a storage in which a plurality of learning algorithm groups is stored in advance, the plurality of learning algorithm groups each including a type of the measurement data, a type of analysis of the measurement data, and a learning algorithm associated with each other […]
a learning algorithm selection step of selecting automatically the learning algorithm to be used for learning from among the plurality of learning algorithm groups based on the input information;
a trained model generation step of generating a trained model based on the training data and the selected learning algorithm
Examiner finds that the claim language does not limit the term “learning algorithm” to an untrained model, or to a method of generating an untrained model; further, nothing in the claim language precludes the “trained model” from being previously trained. Therefore the term “learning algorithm,” in its broadest reasonable interpretation, includes a pre-trained model and the term “trained model” includes a model that has been further trained. The method of Limasanches teaches the selection of a previously trained model (“learning algorithm”) which is subsequently further trained using user data, which meets the limitation “generating a trained model based on the training data and the selected learning algorithm.”
Applicant further submits that Kawaai “does not suggest a system that automatically selects a distinct algorithm type--such as linear regression, random forest, and support vector machines--specifically associated with the physical nature of measurement data.” Examiner has not relied upon Kawaai for any such teaching. Examiner, however, notes that the Applicant’s description of the invention in the remarks is narrower than the language of the claims, which merely states “a learning algorithm selection step of selecting automatically the learning algorithm to be used for learning from among the plurality of learning algorithm groups based on the input information,” which, as noted above, is taught by Limasanches.
The argument is therefore found unpersuasive.
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.
Claims 1–8 rejected under 35 U.S.C. 103 over Limasanches et al., US Pre-Grant Publication No. 2022/0067428 (hereafter Limasanches) in view of Kawaai et al., US Pre-Grant Publication No. 2020/0014761 (hereafter Kawaai).
Regarding claim 1 and analogous claims 7-8:
Limasanches teaches:
“A data analysis method implemented between a system management company and a customer who desires to acquire an analysis result of measurement data acquired from a measuring device, the data analysis method comprising”: Limasanches, paragraph 0005, “A data analysis system configured to perform data analysis between a system management company and a customer who desires to acquire an analysis result of measurement data acquired from a measuring device [A data analysis method implemented between a system management company and a customer who desires to acquire an analysis result of measurement data acquired from a measuring device], the data analysis system comprising: a server configured to generate a trained model to analyze the measurement data; and a data processor configured to request the server to analyze the measurement data; wherein the server includes: a storage configured to store in advance a plurality of learning algorithm groups each including a type of the measurement data, a type of analysis of the measurement data, and a learning algorithm associated with each other; an input receiver configured to receive an input of input information including information about the type of the measurement data and information about the type of analysis of the measurement data, and an input of training data; a learning algorithm selector configured to select the learning algorithm to be used for learning from among the plurality of learning algorithm groups based on the input information; a trained model generator configured to generate the trained model based on the training data and the selected learning algorithm; and an analysis result acquirer configured to analyze the measurement data based on the trained model generated by the trained model generator and acquire the analysis result”; Limasanches, paragraph 0017, “A system disclosed herein may be a physical computer system (one or more physical computers) or may be a system built on a computation resource group (a plurality of computation resources) such as a cloud platform. The computer system or the computation resource group includes one or more interface devices (including, for example, a communication device and an input/output device), one or more storage devices (including, for example, a memory (main storage device) and an auxiliary storage device), and one or more processors.”
“a preparation step of preparing a storage in which a plurality of learning algorithm groups is stored in advance, the plurality of learning algorithm groups each including a type of the measurement data, a type of analysis of the measurement data, and a learning algorithm associated with each other”: Limasanches, paragraph 0005, “According to an aspect of the present invention, a system selects a learning model for a user task. The system includes one or more processor and one or more storage devices. The one or more storage devices store related information on a plurality of existing learning models [preparing a storage in which a plurality of learning algorithm groups is stored in advance]. The one or more processors acquire information on a detail of a new task, extract a new characteristic amount vector from a new training data set for the new task, reference the related information, acquire information on details of tasks of the plurality of existing models and characteristic amount vectors of training data for the plurality of existing models, and select a candidate learning model for the new task from among the plurality of existing models based on a result of comparing the information on the detail of the new task with information on the tasks of the plurality of existing models and a result of comparing the new characteristic amount vector with the characteristic amount vectors of the existing models”; Limasanches, paragraph 0039, ”The model trainer 107 trains the selected existing learning model using the user's training data set. The model database 108 stores the existing model [a learning algorithm], related information the existing model, the newly trained learning model, and related information on the newly trained learning model. As described later, the related information includes a task description of the learning model [a type of analysis of the measurement data] and an essential characteristic amount vector of training data [a type of the measurement data].”
“the type of the measurement data includes an image, a particle size distribution data, a data indicating the surface shape of an object to be measured, or a chromatogram”: Limasanches, paragraph 0066, “The user enters information of a storage location of the data set in the field 602. In the example illustrated, the user desires to solve the task ‘detection of abnormality in image of public area’. The corresponding data set is a folder storing a plurality of images [the type of the measurement data includes an image] of the public area and labels (indicating that an abnormality is present or not present) associated with the images.”
“the type of analysis of the measurement data includes shape analysis or good/bad determination”: Limasanches, paragraph 0066, “The user enters information of a storage location of the data set in the field 602. In the example illustrated, the user desires to solve the task ‘detection of abnormality in image of public area’. The corresponding data set is a folder storing a plurality of images of the public area and labels (indicating that an abnormality is present or not present [the type of analysis of the measurement data includes … good/bad determination]) associated with the images.”
“a training data receiving step of receiving an input of training data”: Limasanches, paragraph 0020, “In an embodiment, a user inputs, to the system, a simple description of a task (new task) desired by the user to be executed and a training data set for the task [receiving step of receiving an input of training data]. The system extracts an essential characteristic amount from the training data set and extracts related information on the task from the description of the task. The system uses a model, data used for training of the model, the corresponding essential characteristic amount, and the description of the corresponding task to find a related learning model in a database storing the foregoing information. The learning model selected from the database is finely adjusted (retrained) using a user's data set. This enables the model to be adapted to a different user's data set.”
“an input information receiving step of receiving an input of input information including information about the type of the measurement data and information about the type of analysis of the measurement data”: Limasanches, paragraph 0025-0026,“The system according to the embodiment of the present specification includes a task analyzer and an essential characteristic amount extractor. Input to the task analyzer is a description input by a user. Details of a task desired by the user to be achieved are briefly described [information about the type of analysis of the measurement data]. Output from the task analyzer is a task expression in a format that enables a next functional section to acquire an optimal learning model. As an example, the task expression can be in the format of a keyword string or a character string. The task description input by the user and the task expression generated from the task description are information on the details of the task. Input to the essential characteristic amount extractor is a user's training data set that includes a plurality of files and is in a folder format. Each of the files is one sample of the training data set. Output from the essential characteristic amount extractor is one-dimensional characteristic amount vectors corresponding to data samples included in the user's training data set [information about the type of the measurement data]. Each of the one-dimensional characteristic amount vectors can include a plurality of elements.”
“a learning algorithm selection step of selecting automatically the learning algorithm to be used for learning from among the plurality of learning algorithm groups based on the input information”: Limasanches, paragraph 0020, “In an embodiment, a user inputs, to the system, a simple description of a task (new task) desired by the user to be executed and a training data set for the task. The system extracts an essential characteristic amount from the training data set and extracts related information on the task from the description of the task. The system uses a model, data used for training of the model, the corresponding essential characteristic amount, and the description of the corresponding task to find a related learning model in a database storing the foregoing information [selecting automatically the learning algorithm to be used for learning from among the plurality of learning algorithm groups based on the input information]. The learning model selected from the database is finely adjusted (retrained) using a user's data set. This enables the model to be adapted to a different user's data set.”
“a trained model generation step of generating a trained model based on the training data and the selected learning algorithm”: Limasanches, paragraph 0039, ”The model trainer 107 trains the selected existing learning model using the user's training data set [generating a trained model based on the training data and the selected learning algorithm]. The model database 108 stores the existing model, related information the existing model, the newly trained learning model, and related information on the newly trained learning model. As described later, the related information includes a task description of the learning model and an essential characteristic amount vector of training data.”
Limasanches does not explicitly teach:
“and the learning algorithm is a method for generating the trained model, including at least one of linear regression, a random forest, a neural network, or a support vector machine”
“an analysis result acquisition step of analyzing the measurement data based on the trained model and acquiring the analysis result”
Kawaai teaches:
“and the learning algorithm is a method for generating the trained model, including at least one of linear regression, a random forest, a neural network, or a support vector machine”: Kawaai, paragraph 0030, “The matching degree of items is, for example, determined for each item based on whether each item of device data is matched. In some embodiments, matching degree of items may be determined based on the number of matching of items. In some embodiments, if no shared model matching the definition of the environment, conditions, and the like of the device is found, a new model having a neural network structure suitable for the definition may be generated [the learning algorithm is a method for generating … a neural network].”
“an analysis result acquisition step of analyzing the measurement data based on the trained model and acquiring the analysis result”: Kawaai, paragraph 0038, “Finally, in the device 20 ( e.g., the plurality of devices 201, 202, . . . , 20n in FIG. 1), in a state where the shared model or the additional learned model is stored in the learner, inference processing is performed in the learner by using the device data and an inference result as output data is obtained (S25) [analyzing the measurement data based on the trained model and acquiring the analysis result].”
Kawaai and Limasanches are analogous arts as they are both related to automated model selection. It would have been obvious to a person having ordinary skill in the art prior to the effective filing date of the claimed invention to have combined the model inference of Kawaai with the teachings of Limasanches to arrive at the present invention, in order to apply the selected model to get analysis results, as stated in Kawaai, paragraph 0038, “Finally, in the device 20 ( e.g., the plurality of devices 201, 202, . . . , 20n in FIG. 1), in a state where the shared model or the additional learned model is stored in the learner, inference processing is performed in the learner by using the device data and an inference result as output data is obtained (S25).”
Regarding claim 2:
Limasanches as modified by Kawaai teaches “The data analysis method according to claim 1.”
Limasanches further teaches “wherein in the learning algorithm selection step, the learning algorithm corresponding to information about the type of the measurement data and information about the type of analysis of the measurement data is selected from among the plurality of learning algorithm groups”: Limasanches, paragraph 0020, “In an embodiment, a user inputs, to the system, a simple description of a task (new task) desired by the user to be executed and a training data set for the task. The system extracts an essential characteristic amount from the training data set and extracts related information on the task from the description of the task. The system uses a model, data used for training of the model, the corresponding essential characteristic amount [information about the type of the measurement data], and the description of the corresponding task [and information about the type of analysis of the measurement data] to find a related learning model in a database storing the foregoing information [the learning algorithm corresponding to information about the type of the measurement data and information about the type of analysis of the measurement data is selected from among the plurality of learning algorithm groups]. The learning model selected from the database is finely adjusted (retrained) using a user's data set. This enables the model to be adapted to a different user's data set.”
Regarding claim 3:
Limasanches as modified by Kawaai teaches “The data analysis method according to claim 2.”
Limasanches further teaches:
“a trained model storage step of storing, in association with each other, the trained model generated in the trained model generation step, and the type of the measurement data and the type of analysis of the measurement data, both of which are received in the input information receiving step”: Limasanches, paragraph 0039, ”The model trainer 107 trains the selected existing learning model using the user's training data set. The model database 108 stores the existing model, related information the existing model, the newly trained learning model, and related information on the newly trained learning model. As described later, the related information includes a task description of the learning model and an essential characteristic amount vector of training data [storing, in association with each other, the trained model generated in the trained model generation step, and the type of the measurement data and the type of analysis of the measurement data].”
“an analysis receiving step of receiving the information about the type of the measurement data, the information about the type of analysis of the measurement data, and the measurement data acquired by the measuring device”: Limasanches, paragraph 0020, “In an embodiment, a user inputs, to the system, a simple description of a task (new task) desired by the user [information about the type of analysis of the measurement data] to be executed and a training data set for the task [the measurement data]. The system extracts an essential characteristic amount from the training data set [information about the type of the measurement data] and extracts related information on the task from the description of the task.”
“a trained model selection step of selecting the trained model to be used to analyze the measurement data based on the information about the type of the measurement data and the information about the type of analysis of the measurement data received in the analysis receiving step”: Limasanches, paragraph 0020, “In an embodiment, a user inputs, to the system, a simple description of a task (new task) desired by the user to be executed and a training data set for the task. The system extracts an essential characteristic amount from the training data set and extracts related information on the task from the description of the task. The system uses a model, data used for training of the model, the corresponding essential characteristic amount, and the description of the corresponding task to find a related learning model in a database storing the foregoing information [selecting the trained model to be used to analyze the measurement data based on the information about the type of the measurement data and the information about the type of analysis]. The learning model selected from the database is finely adjusted (retrained) using a user's data set. This enables the model to be adapted to a different user's data set.”
Kawaai further teaches “wherein in the analysis result acquisition step, the measurement data is input to the trained model selected in the trained model selection step, and the analysis result is acquired”: Kawaai, paragraph 0038, “Finally, in the device 20 ( e.g., the plurality of devices 201, 202, . . . , 20n in FIG. 1), in a state where the shared model or the additional learned model is stored in the learner, inference processing is performed in the learner by using the device data and an inference result as output data is obtained (S25) [the measurement data is input to the trained model selected in the trained model selection step, and the analysis result is acquired].”
Kawaai and Limasanches are combinable for the rationale given under claim 1.
Regarding claim 4:
Limasanches as modified by Kawaai teaches “The data analysis method according to claim 3.”
Limasanches further teaches “the data processing program being associated with the type of the measurement data and the type of analysis of the measurement data”: Limasanches, paragraph 0039, ”The model trainer 107 trains the selected existing learning model using the user's training data set. The model database 108 stores the existing model, related information the existing model, the newly trained learning model, and related information on the newly trained learning model [the data processing program being associated with the type of the measurement data and the type of analysis of the measurement data]. As described later, the related information includes a task description of the learning model [type of analysis of the measurement data] and an essential characteristic amount vector of training data [type of the measurement data].”
Kawaai further teaches “a data processing program storage step of storing in advance a plurality of data processing program groups each including a data processing program configured to perform a data process different from a data process of the trained model”: Kawaai, paragraphs 0019-0020, “In addition, in the learned model providing system according to some embodiments of the present disclosure, the device has a function of performing additional learning processing on a shared model. The server device includes an additional learned model management unit configured to receive an additional learned model transmitted from the device to cause a storage unit to store the additional learned model. A target shared model selection unit of the server device is configured to perform selection by including as option, in addition to a shared model, also an additional learned model. In addition, in the learned model providing system according to some embodiments of the present disclosure, the device has a function of performing additional learning processing on a shared model, and includes a storage unit caused to store an additional learned model, and an additional learned model information transmitter configured to transmit information necessary for selecting an additional
learned model to the server device. A target shared model selection unit of the server device is configured to perform selection by including as option, in addition to the shared model, also an additional learned model stored in a storage unit of the device [storing in advance a plurality of data processing program groups each including a data processing program configured to perform a data process different from a data process of the trained model].”
Kawaai and Limasanches are analogous arts as they are both related to automated model selection. It would have been obvious to a person having ordinary skill in the art prior to the effective filing date of the claimed invention to have combined the additional processing step of Kawaai with the teachings of Limasanches to arrive at the present invention, in order to provide further model specialization, as stated in Kawaai, paragraph 0021, “In addition, providing an additional learning processing function allows an additional learned model more specialized in the environment and conditions of the device to be obtained, so that it is possible to additionally perform highly accurate inference processing in the device.”
Regarding claim 5:
Limasanches as modified by Kawaai teaches “The data analysis method according to claim 4.”
Limasanches further teaches (bold only) “a data processing program addition receiving step of receiving addition of another data processing program associated with the type of the measurement data and the type of analysis of the measurement data to the plurality of data processing program groups stored in advance“: Limasanches, paragraph 0039, ”The model trainer 107 trains the selected existing learning model using the user's training data set. The model database 108 stores the existing model, related information the existing model, the newly trained learning model, and related information on the newly trained learning model [the data processing program being associated with the type of the measurement data and the type of analysis of the measurement data]. As described later, the related information includes a task description of the learning model [type of analysis of the measurement data] and an essential characteristic amount vector of training data [type of the measurement data].”
Kawaai further teaches (bold only) “a data processing program addition receiving step of receiving addition of another data processing program associated with the type of the measurement data and the type of analysis of the measurement data to the plurality of data processing program groups stored in advance”: Kawaai, paragraph 0035, “A shared model is selected or a learning model is newly generated, and then additional learning is performed by a learner on the shared model or the new learning model (S16). The additional learning is performed by using sample data for performing additional learning, collected from the device 20. After the additional learning is completed, the generated additional learned model is stored in the storage unit 15 (S17) [receiving addition of another data processing program … to the plurality of data processing program groups stored in advance]. The server device 10 may transmit the generated additional learned model to the device 20.”
Kawaai and Limasanches are analogous arts as they are both related to automated model selection. It would have been obvious to a person having ordinary skill in the art prior to the effective filing date of the claimed invention to have combined the storage of the additional processing step of Kawaai with the teachings of Limasanches to arrive at the present invention, in order to reuse the additional model in providing further model specialization, as stated in Kawaai, paragraph 0021, “In addition, providing an additional learning processing function allows an additional learned model more specialized in the environment and conditions of the device to be obtained, so that it is possible to additionally perform highly accurate inference processing in the device.”
Regarding claim 6:
Limasanches as modified by Kawaai teaches “The data analysis method according to claim 1.”
Limasanches further teaches “a learning algorithm addition receiving step of receiving addition of another learning algorithm associated with the type of the measurement data and the type of analysis of the measurement data to the plurality of learning algorithm groups stored in advance”: Limasanches, paragraph 0051, “When the ratio of the harmful sample is smaller than the threshold (YES in step S106), the model trainer 107 trains the selected learning model using the user data set (S109). Input to the learning model for the training is the essential characteristic amount vector extracted from the user data set. After that, the trained learning model, the essential characteristic amount vector of the training data, and the task description are stored in the model database 108 and can be used for the future (S110) [receiving addition of another learning algorithm associated with the type of the measurement data and the type of analysis of the measurement data to the plurality of learning algorithm groups stored in advance].”
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Jia et al., US Pre-Grant Publication No. 2020/0210708, discloses a method for selecting a convolution-based neural network model for classification of an image, based on characteristics of the image.
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to VINCENT SPRAUL whose telephone number is (703) 756-1511. The examiner can normally be reached M-F 9:00 am - 5:00 pm.
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, MICHAEL HUNTLEY can be reached at (303) 297-4307. 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.
/VAS/
Examiner, Art Unit 2129
/MICHAEL J HUNTLEY/Supervisory Patent Examiner, Art Unit 2129
Prosecution Insights