Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Amendment
The following Office action in response to communications received April 16, 2026. Claims 1, 5 and 13 have been amended. Claims 14-16 have been canceled. Claim 17 has been added. Therefore, claims 1-13 and 17 are pending and addressed below.
Applicant’s amendments to the claims are sufficient to overcome the 35 USC § 112 and 35 USC § 102 rejections set forth in the previous office action dated December 17, 2025.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-13 and 17 are rejected under 35 U.S.C. § 101 because the claimed invention is directed to a judicial exception without significantly more.
Independent Claims 1, 5, and 13 are directed to an abstract idea consisting of collecting information regarding a target person’s health state and action history, evaluating whether prior actions were successful or unsuccessful in changing the health state, and generating a recommended sleep duration or other recommended action for improving the health state using that evaluation.
Claim 1 recites: “A learning device comprising: at least one memory storing instructions; and at least one processor configured by the instructions to: acquire history information of a target person, the history information including a heart rate history that is acquired from a wearable terminal of the target person, an action history including exercise frequency as an action of the target person, and success / failure information indicating whether or not the action contributed to variation in a health state of the target person; and generate a recommendation model that includes an action policy selection engine and an action imitation engine, wherein the action imitation engine is configured to output information regarding a recommended sleep duration recommended to the target person to improve the health state of the target person based on an input of the history information indicating the history of the action and the health state of the target person, wherein the action policy selection engine includes a success example classifier and a failure example classifier, and wherein the recommendation model is generated by: inputting, from the history information, a past example indicating a history of the action and the health state of the target person, obtaining, using the action imitation engine, an inference result corresponding to the recommended sleep duration for the target person based on the inputted past example, training the action imitation engine and the success example classifier so as to reduce a difference between the inference result and the success information, and training the action imitation engine and the failure example classifier so as to increase a difference between the inference result and the failure information.”
Independent Claim 1, under its broadest reasonable interpretation, recites collecting health-related information and behavior-related information for a target person, evaluating whether prior actions were successful or unsuccessful, and using that information to determine a recommended sleep duration for improving the target person’s health state.
The limitations of Claim 1, as drafted, under their broadest reasonable interpretation, cover the performance of:
Mental processes, including observation of heart rate history and action history, evaluation of success / failure information, judgment regarding whether prior actions contributed to variation in a health state, and decision making regarding a recommended sleep duration.
Certain methods of organizing human activity, including managing personal behavior and health behavior by recommending how long a target person should sleep and by evaluating whether prior exercise frequency and related actions improved the health state of the target person.
Mathematical concepts, because the claim expressly recites “generate a recommendation model,” “obtaining … an inference result,” “training the action imitation engine and the success example classifier so as to reduce a difference,” and “training the action imitation engine and the failure example classifier so as to increase a difference,” which amount to mathematical evaluation, classification, and optimization of model outputs.
But for the recitation of generic computer components, the claim steps are simply collecting information about a person’s heart rate, exercise frequency, and health state, determining whether prior conduct was successful or unsuccessful, and recommending a sleep duration based on that evaluation.
Claim 5 recites: “An action recommendation device comprising: at least one memory storing instructions; and at least one processor configured by the instructions to: acquire heart rate data of a target person from a wearable terminal; generate history information for the target person, the history information including a heart rate history and an action history including an exercise frequency as an action of the target person; determine a recommended sleep duration for the target person based on the history information and a recommendation model; and output the recommended sleep duration to the wearable terminal, wherein the recommendation model includes an action policy selection engine and an action imitation engine and is a pre-trained model that correlates historical data of a plurality of subjects, including their heart rates, with recommended sleep durations, wherein the action policy selection engine includes a success example classifier and a failure example classifier, wherein the recommendation model is pre-trained using adversarial cooperative imitation learning with the action policy selection engine and the action imitation engine, wherein the action imitation engine and the success example classifier are trained so as to reduce a difference between an inference result and success information from the historical data of the plurality of subjects, and wherein the action imitation engine and the failure example classifier are trained so as to increase a difference between the inference result and failure information from the historical data of the plurality of subjects.”
Independent Claim 5, under its broadest reasonable interpretation, recites acquiring health data from a wearable terminal, generating a history from that data, determining a recommended sleep duration using a pre-trained model, and outputting the recommendation to the user.
The limitations of Claim 5, as drafted, under their broadest reasonable interpretation, cover the performance of:
Mental processes, including observation of heart rate data, generation and review of history information, evaluation of historical data of a plurality of subjects, and judgment regarding a recommended sleep duration.
Certain methods of organizing human activity, including managing personal sleep behavior and health behavior through a recommendation delivered to the target person.
Mathematical concepts, because the claim recites a “pre-trained model,” “correlates historical data,” “adversarial cooperative imitation learning,” “reduce a difference,” and “increase a difference,” all of which describe mathematical modeling and training operations for producing a recommendation.
But for the recitation of generic computer components, the claim steps are simply receiving health data, comparing that data with prior subject histories, and outputting advice on how long the person should sleep.
Claim 13 recites: “A learning method executed by a computer, the learning method comprising: acquiring history information of a target person, the history information including a heart rate history that is acquired from a wearable terminal of the target person, an action history including exercise frequency as an action of the target person, and success / failure information indicating whether or not the action contributed to variation in a health state of the target person; and generating a recommendation model that includes an action policy selection engine and an action imitation engine, wherein the action imitation engine is configured to output information regarding a recommended sleep duration recommended to the target person to improve the health state of the target person based on an input of the history information indicating the history of the action and the health state of the target person, wherein the action policy selection engine includes a success example classifier and a failure example classifier, and wherein the generating the model comprises: inputting, from the history information, a past example indicating the history of the action and the health state of the target person, obtaining, using the action imitation engine, an inference result corresponding to the recommended sleep duration for the target person based on the inputted past example, training the action imitation engine and the success example classifier so as to reduce a difference between the inference result and the success information, and training the action imitation engine and the failure example classifier so as to increase a difference between the inference result and the failure information.”
Independent Claim 13 recites substantially the same abstract idea as Claim 1 in method form, namely collecting information regarding health state and action history, evaluating success / failure information, and generating a recommended sleep duration using model-based analysis of that information.
The limitations of Claim 13, as drafted, under their broadest reasonable interpretation, cover the performance of mental processes, certain methods of organizing human activity, and mathematical concepts for the same reasons discussed above with respect to Claim 1.
The claims recite additional elements such as:
at least one processor;
at least one memory storing instructions;
a wearable terminal;
a recommendation model;
an action policy selection engine;
an action imitation engine;
a success example classifier;
a failure example classifier; and
output of recommended sleep duration or other recommended action information.
These elements are recited at a high level of generality and merely use generic computer components to perform generic computer functions.
Looking to the specification, the hardware is described as conventional computing hardware, including a processor 11, memory 12, and interface 13 for the learning device, and a processor 21, memory 22, and interface 23 for the action recommendation device. The specification further explains that examples of the processor include a CPU, GPU, or TPU, and the memory includes RAM, ROM, and flash memory, indicating that the claimed computing environment is conventional rather than a technological improvement to computer functionality.
The specification also describes the sensor 6 and wearable terminal at a high level of generality as sources of ordinary biometric signals such as heart rate, EEG, pulse wave, blood pressure, body temperature, respiration rate, acceleration, and sleep time, again reflecting use of generic sensing and communication components for data gathering.
The additional elements do not integrate the judicial exception into a practical application. The claims do not improve the functioning of a computer, processor, memory, wearable terminal, sensor, network, or other technology. Instead, the claims use generic processors, memory, a wearable terminal, and model components as tools to collect information, analyze the information, and output a recommendation about a target person’s sleep duration or other action.
The recited “action policy selection engine,” “action imitation engine,” “success example classifier,” and “failure example classifier” do not change this conclusion. Even accepting these limitations as claimed, they merely describe different logical components of a model used to evaluate information and generate a recommendation, i.e., different labels for mathematical analysis and decision-making steps implemented on a computer.
The machine learning and training limitations likewise do not impose a meaningful limit on the judicial exception. The claim language directed to “obtaining … an inference result,” “reduce a difference,” “increase a difference,” “pre-trained model,” and “adversarial cooperative imitation learning” merely describes how the abstract idea is mathematically modeled and optimized, rather than an improvement to computer technology itself.
Further, under MPEP 2106.05(f), mere instructions to implement an abstract idea on a computer, or to use generic computer components as tools to perform the abstract idea, do not integrate the exception into a practical application. Here, the claims merely invoke generic processor, memory, wearable, model, classifier, and output components to carry out the collection, evaluation, training, and recommendation activities. The recited ML training operations are therefore no more than instructions to apply the abstract idea using generic computing technology and mathematical modeling.
To the extent the claims recite outputting the recommended sleep duration to a wearable terminal, generating recommended action promotion information, generating basis information, determining a timing for recommendation, acquiring target person data from a terminal device, or generating history information from diagnosis data of a medical checkup, those features merely gather additional data, present the result of the abstract analysis, or limit the abstract idea to a particular field of use in health management, and therefore do not integrate the exception into a practical application.
The ordered combination of claim elements adds nothing significantly more than the abstract idea itself. The use of generic processors, generic memory, generic wearable terminals, generic sensors, data storage, and model execution to receive data, generate history information, classify success and failure, train a recommendation model, and output a recommended sleep duration is conventional and routine as described in the specification. The specification repeatedly presents these components as ordinary computing and sensing infrastructure used to carry out the desired recommendation logic.
The ML and training limitations also do not amount to an inventive concept. Recitations such as “generate a recommendation model,” “action policy selection engine,” “action imitation engine,” “success example classifier,” “failure example classifier,” “obtain … an inference result,” “reduce a difference,” “increase a difference,” and “pre-trained using adversarial cooperative imitation learning” merely set out the particular mathematical manner in which the abstract idea is evaluated and optimized, which is not enough to supply significantly more under § 101.
Under MPEP 2106.05(f), the mere use of a computer as a tool to perform an abstract idea, including evaluating data, classifying data, optimizing outputs, and generating recommendations, does not provide an inventive concept. Here, the additional elements simply automate the abstract idea of assessing prior actions and health outcomes and recommending a future action or sleep duration. Any storing, transmitting, displaying, or outputting of the recommendation is insignificant extra-solution activity.
It is worth noting that the above analysis already encompasses each of the current dependent claims (i.e., claims 2-4, 6-12, and 17). Particularly, each of the dependent claims also fails to amount to “significantly more’ than the abstract idea since each dependent claim is directed to a further abstract idea, and/or a further conventional computer element/function utilized to facilitate the abstract idea. Accordingly, none of the current claims implements an element—or a combination of elements—directed to an inventive concept (e.g., none of the current claims is reciting an element—or a combination of elements—that provides a technological improvement over the existing/conventional technology). These information characteristics do not change the fundamental analogy to the abstract idea of groupings, and when viewed individually or as a whole, they do not add anything substantial beyond the abstract idea. Furthermore, the combination of elements does not indicate a significant improvement to the functioning of a computer or any other technology. Therefore, the claims when taken as a whole are ineligible for the same reasons as the independent claims.
Accordingly, the claims are directed to an abstract idea without significantly more and therefore are not patent eligible under 35 U.S.C. § 101.
"Subject Matter Free of Prior Art"
Examiner notates below the reasons why the claims overcome the prior art. The limitations most likely to distinguish over the cited Bostic et al. reference are:
Claim 1
“Generate a recommendation model that includes an action policy selection engine and an action imitation engine, wherein the action policy selection engine includes a success example classifier and a failure example classifier.”
“Training the action imitation engine and the success example classifier so as to reduce a difference between the inference result and the success information.”
“Training the action imitation engine and the failure example classifier so as to increase a difference between the inference result and the failure information.”
Claim 5
“Wherein the recommendation model includes an action policy selection engine and an action imitation engine and is a pre-trained model that correlates historical data of a plurality of subjects, including their heart rates, with recommended sleep durations.”
“Wherein the action policy selection engine includes a success example classifier and a failure example classifier.”
“Wherein the recommendation model is pre-trained using adversarial cooperative imitation learning with the action policy selection engine and the action imitation engine, wherein the action imitation engine and the success example classifier are trained so as to reduce a difference between an inference result and success information … and … the failure example classifier are trained so as to increase a difference between the inference result and failure information from the historical data of the plurality of subjects.”
Claim 13
“Generating a recommendation model that includes an action policy selection engine and an action imitation engine, wherein the action policy selection engine includes a success example classifier and a failure example classifier.”
“Obtaining, using the action imitation engine, an inference result corresponding to the recommended sleep duration for the target person based on the inputted past example.”
“Training the action imitation engine and the success example classifier so as to reduce a difference between the inference result and the success information and training the action imitation engine and the failure example classifier so as to increase a difference between the inference result and the failure information.”
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Pub. No.: US 20210005324 A1 to Bostic et al.; A computerized method for healthcare data management generally includes receiving a symptom report from an individual, and receiving, at a healthcare data system computing device, continuous health information including data related to an individual's health state and data related to a health state of a population of patients; calculating, at the healthcare data system computing device, a risk score, wherein the risk score is based on at least the individual's symptom report, on data related to an individual's health state, and on the health state of a population of patients, wherein the population of patients share an attribute with the individual; and sending, from the healthcare data system computing device, a return-to-work recommendation to an entity based at least in part on the calculated risk score.
Pub. No.: US 20210202103 A1; Systems and methods are provided for simulating a patient health state by determining one or more relationships between patient data and historical data, creating enriched data elements based on the determined relationships, and using a machine learning module to compute a current health state for a patient and to simulate a future health state of the patient.
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 EDWARD B WINSTON III whose telephone number is (571)270-7780. The examiner can normally be reached M-F 1030 to 1830.
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, Robert Morgan can be reached at (571) 272-6773. 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.
/E.B.W/ Examiner, Art Unit 3683
/ROBERT W MORGAN/ Supervisory Patent Examiner, Art Unit 3683
Prosecution Insights