DETAILED ACTION
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 .
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 12/16/2025 has been entered.
Response to Arguments
Rejections under 35 USC 101
Applicant's arguments filed 12/16/2025 with regards to the rejection of claims 1-20 under 35 USC 101 have been fully considered but they are not persuasive.
On pages 8-9 of the Remarks filed 12/16/2025, Applicant argues that the claimed invention as amended is directed to more than an abstract idea because of the required specific physical configuration where the “at least one sensor” is “integrated with the bed.” Applicant argues that this amounts to more than mere data gathering and provides a specific technical solution which provides a new and useful technique for receiving sensor readings used to predict alertness levels.
Applicant further argues that adjusting parameters of a machine learning model based on an alertness rating cannot practically be performed in the human mind and that the human mind cannot practically adjust parameters of a trained machine learning model according to a user input.
Applicant argues that the claimed invention is eligible under Step 2A Prong 2 because the claim reflects a technical improvement to the functioning of the computer system itself, specifically as a technical improvement to the performance of a machine learning model to improve accuracy in determining alertness levels of users.
Examiner respectfully disagrees with Applicant and maintains that the claimed invention is not eligible for patentability under 35 USC 101.
The inclusion of at least one sensor integrated into a bed does not amount to integrating the claimed invention into practical application because it does not offer a new and useful technique for data gathering, and the additional elements still only amount to data gathering. A sensor integrated into a bed to gather data of a sleeping user is known in the field (for example as taught by Sayadi et al US 20200205580 A1). Although the inclusion of the bed and sensor in the claims links the claimed invention to a technological environment and field of use, it still only further defines the extra-solution activity of data gathering for input into an equation/abstract idea.
Adjusting the parameters of the machine learning model based on a user alertness rating does not amount to significantly more than the computer implementation of the abstract idea itself. Iterating the machine learning model through various training models can be considered to be repetitive calculations to reduce error in the predicted alertness level as compared to the user alertness rating, which amounts to extra-solution activity. Additionally, the machine learning model is claimed generically such that it is essentially no different than a generic computer.
The training of the machine learning model does not amount to a technical improvement to the computer system because it uses only the generic capabilities of the computer system, and rather improves its accuracy/lowers error through repetitive calculation as mentioned above.
With this in consideration, the rejection of the claims as being unpatentable under 35 USC 101 is maintained.
Rejections under 35 USC 102/103
Applicant’s arguments with respect to claim(s) 1-20 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument. New grounds of rejection as necessitated by the amended claims detailed below.
Information Disclosure Statement
The Information Disclosure Statements (IDS) filed on 02/28/2023 and 07/12/2024 have been considered by the examiner.
Claim Objections
Claim 1 objected to because of the following informalities: “first sleep sessions” in line 22 should read “first sleep session”. Appropriate correction is required.
Claim 18 objected to because of the following informalities: “first sleep sessions” in line 18 should read “first sleep session”. Appropriate correction is required.
Claim 20 objected to because of the following informalities: “first sleep sessions” in line 20 should read “first sleep session”. Appropriate correction is required.
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 1-10, 13-16, 18, and 20 rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea/mental process without significantly more.
Step 1
Claims 1 and 20 recite a system (machine) and claim 18 recites a method.
Step 2A, Prong 1
Claims 1, 18, and 20 recite the limitations of receiving a first or second predicted alertness level as an output of the first or second version of a machine learning model, which is understood by the Examiner to amount to determining the predicted alertness level of a user. These steps, given their broadest reasonable interpretation, can be directed to a mental process based on judgements made from an observation of outputs from a model. A person could perform the act of determining their predicted alertness level based upon observation and historical data in their mind. Therefore, the claims recite a mental process/abstract idea.
Step 2A, Prong 2
Claims 1, 18, and 20 do not include any additional elements that integrate the abstract idea into a practical application.
Claim 1 includes the additional elements of a bed, at least one sensor, a computer system, a first version of the machine learning model to output predicted alertness levels of a user, and a second version of the machine model to output predicted alertness levels of a user.
The bed having at least one integrated sensor is claimed generically to link the invention to a particular technological environment or field of use.
The at least one sensor is directed to extra-solution activity in the form of data gathering as performing clinical tests on individuals to obtain an input for an equation, where the sensor readings are used to provide input into the alertness prediction model. See MPEP 2106.05(g).
The computer system is generically claimed that they amount to generic computer implementation of the abstract idea. See MPEP 2106.05(a).
The first and second versions of the machine learning model also amount to generic computer implementation/automation of the human decision-making process using a computer. The inclusion of the training of the machine learning model based on an alertness rating provided by the user and alertness predictions do not provide an improvement to the computer technology and rather harness the capabilities of a general-purpose computer to predict an alertness level. See MPEP 2106.05(a).
Outputting a predicted alertness level is extra-solution activity of data outputting. See MPEP 2106.05(g).
Claim 18 includes the additional elements present in claim 1.
Claim 20 includes the additional elements present in claim 1 and one or more processors and one or more computer-readable devices. The one or more processors and one or more computer-readable devices are generically claimed such that they amount to generic computer implementation of the abstract idea.
Therefore, none of these elements amount to integrating the abstract idea into a practical application.
Step 2B
Claims 1, 18, and 20 do not include any additional elements that amount to significantly more than the abstract idea.
Claim 1 includes the additional elements of a bed, at least one sensor, a computer system, a first version of the machine learning model to output predicted alertness levels of a user, and a second version of the machine model to output predicted alertness levels of a user.
The bed having at least one integrated sensor is claimed generically to link the invention to a particular technological environment or field of use.
The at least one sensor is directed to extra-solution activity in the form of data gathering as performing clinical tests on individuals to obtain an input for an equation, where the sensor readings are used to provide input into the alertness prediction model. See MPEP 2106.05(g).
The computer system is generically claimed that they amount to generic computer implementation of the abstract idea. See MPEP 2106.05(a).
The first and second versions of the machine learning model do not provide an improvement to the computer technology and rather harness the capabilities of a general-purpose computer to predict an alertness level. See MPEP 2106.05(a).
The inclusion of the training of the machine learning model based on an alertness rating provided by the user and alertness predictions amounts to extra-solution activity of repetitive calculations, by recomputing the predicted alertness level, and only required general computer functionality and therefore does not impose meaningful limits on the scope of the claims. See MPEP 2106.05(d).
Outputting a predicted alertness level is extra-solution activity of data outputting. See MPEP 2106.05(g).
Claim 18 includes the additional elements present in claim 1.
Claim 20 includes the additional elements present in claim 1 and one or more processors and one or more computer-readable devices. The one or more processors and one or more computer-readable devices are generically claimed such that they amount to generic computer implementation of the abstract idea.
The additional elements as described above with regards to data gathering and computer implementation of the abstract idea can be held to be well-understood, routine, and conventional in the art, and they are recited with a high level of generality which does not amount to significantly more than the abstract idea itself.
Therefore, none of these elements amount to significantly more than the abstract idea.
Claims 2, 3, 6-10, 14, and 16 further limit the extra-solution activity of data gathering.
Claims 4 and 10 further limits the extra-solution activity of data outputting.
Claim 5 further defines the abstract idea itself.
Claim 15 further limits the computer implementation of the abstract idea.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 1-3, 6-10, 13, 15, 18, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Shouldice (WO 2021171266 A1), with reference to corresponding US application US 20230128912 A1 for citation purposes, in view of Swift et al (US 20200219616 A1).
Regarding claim 1, Shouldice teaches a system comprising:
a bed (232);
at least one sensor (130) integrated with the bed and configured to sense physical phenomenon of a user (see [0063]; at least one of the sensors 130 is positioned generally adjacent to the user during the sleep session, e.g. coupled to the mattress); and
a computer system (control system 110 via interface 119) in communication with the at least one sensor (130), the computer system configured to (see [0029-0032]; electronic interface 119 is configured to receive data from the one or more sensors 130 such that the data can be stored in the memory device 114 and/or analyzed by the processor 112 of the control system 110):
receive, from the at least one sensor, first sensor readings of the user during a first sleep session (see Fig. 7, [0129]; step 702 receive data associated with a user during a sleep session);
provide, as first input, the first sensor readings to a first version of a machine learning model that was trained using machine learning techniques to predict alertness levels of the user (see Fig. 7, [0140]; step 704 determine an alertness level of the user using a machine learning model that takes as input the received data from step 702) and includes one or more parameters estimated based at least in part on physical phenomenon of the user and historic data about at least one of the user and a population of users (see [0129-0133]; receiving data associated with the user and examples thereof, Fig. 8, [0151-0153]; step 802 receive historical sleep-session data associated with at least one person for a plurality of historical sleep sessions); and
receive, as first output from the first version of the machine learning model, first data indicating a first predicted alertness level of the user for a first period of time that starts after the user wakes up from the first sleep session (see Fig. 7, [0144-0147]; step 706; generate a response based at least in part on the determined alertness level);
receive first user input that specifies a first alertness rating reported by the user for the first sleep sessions (see [0103]; user feedback can include an analog rating of a subjective energy level of the user, [0151]; the historical sleep-session data includes subjective feedback from the at least one person, which may be one or more ratings for how the person grades one or more historical sleep sessions);
modify the first version of the machine learning model to create a second version of the machine learning model (see [0160]; the machine learning model is trained in successive iterations such that historical sleep-session data and historical alertness data are used to train the machine learning model in a first iteration),
by adjusting at least one of the one or more parameters of the machine learning model based, at least in part, on the first alertness rating (see [0159-0161]; the control system can update the machine learning model by further training the ML model using the historical sleep-session data and historical alertness associated with the present user);
receive, from the at least one sensor, second sensor readings of the user during a second plurality of sleep session (see [0160-0162]; the machine learning model is trained iteratively using historical sleep-session data associated with a user);
provide, as second input, the second sensor readings to the second version of the machine learning model (see Fig. 7, [0129]; step 702 receive data associated with a user during a sleep session, it can be appreciated that due to the iterative nature of the machine learning model the processes in Fig. 7 and 8 are repeated more than once); and
receive, as second output from the second version of the machine learning model, second data indicating a second predicted alertness level of the user (see Fig. 7, [0140]; step 704 determine an alertness level of the user using a machine learning model that takes as input the received data from step 702) for a second period of time that starts after the user wakes up from the second sleep session (see [0140]; alertness determinations in the future can be obtained approximately the same time after the user wakes from sleep).
Shouldice teaches updating the machine learning model using historical sleep-session data associated with the user in addition to historical sleep-session data associated with a population in order to reduce the error rate of the model as it is trained to tailor predictions for the individual user (Shouldice [0161]). With training, the machine learning model predicts an alertness score that takes into account not just raw alertness data but also health condition and/or demographic data (Shouldice [0162]).
Shouldice is silent regarding wherein the machine learning model determines that the first alertness rating from the user is outside a threshold range of the first predicted alertness level; and
in response to determining that the first alertness rating is outside the threshold range of the first predicted alertness level, modifying the first version of the machine learning model to create a second version of the machine learning model.
Swift teaches a system which uses a machine learning model to predict wakefulness windows for a user (Swift [0020]) using patient data to train/update the model in real time (Swift [0025]). Swift’s model is configured to receive first user input that specifies a first alertness rating reported by the user (see Swift [0045]; the program code tunes the model bason on user feedback),
determine that the first alertness rating from the user is outside a threshold range of the first predicted alertness level (see Swift [0045]; the model may predict a wakefulness window for a user and the patient may not actually be awake, this data can be provided to the program code via a computing interface); and
in response to determining that the first alertness rating is outside the threshold range of the first predicted alertness level (see Swift [0045]; upon receipt of this data, the program code analyzes the discrepancy and revises the model in accordance with this inconsistency; it can be appreciated that the threshold in this case is considered to be the state of wakefulness of the user and when the user is not awake, the rating is outside of the threshold), modifying the first version of the machine learning model to create a second version of the machine learning model (see Swift [0045]; the program code can combine predictions for defined time windows with baseline probabilities that are improved over time, and revise/correct these values by obtaining feedback from individual users).
It would have been obvious for one of ordinary skill in the art prior to the effective filing date of the claimed invention to modify Shouldice’s iterative machine learning model for predicting alertness of a user by comparing a predicted alertness/wakefulness level with feedback from users as taught by Swift. One of ordinary skill in the art would have been motivated to make this modification in order to continuously improve the machine learning model and increase model personalization for the individual user (Swift [0045]).
Regarding claim 2, Shouldice and Swift teach the system of claim 1. Shouldice further teaches wherein the historic data includes, for the user, at least one of sleep data, health metrics, and physical phenomenon (see Shouldice [0151]; historical sleep-session data can include a sleep score, a respiration rate/signal, a sleep state, heart rate, blood pressure, and more).
Regarding claim 3, Shouldice and Swift teach the system of claim 1. Shouldice further teaches wherein the historic data includes, for the population of users, at least one of sleep data, health metrics, and physical phenomenon (see Shouldice [0151-0152]; historical sleep-session data can include a sleep score, a respiration rate/signal, a sleep state, heart rate, blood pressure, and more), wherein the population of users is within a particular age group (see Shouldice [0151-0153]; cohorts can be based on demographic information including age group or those with certain health conditions or genetic markers).
Regarding claim 6, Shouldice and Swift teach the system of claim 1. Shouldice further teaches wherein the first period of time is 24 hours from a time at which the user wakes up from the first sleep session (see [0140]; alertness level determinations for the future time of day can be windowed e.g. 24 hours ahead).
Regarding claim 7, Shouldice and Swift teach the system of claim 1. Shouldice further teaches wherein the first period of time is an amount of time that the user is expected to be awake before a next sleep session (see [0140]; alertness level determinations for the future time of day can be windowed e.g. 2 hours ahead, 12 hours ahead, etc. or determined for certain times of the day e.g. 10Am, 2PM, 5PM, 7PM, 10PM).
Regarding claim 8, Shouldice and Swift teach the system of claim 1. Shouldice further teaches wherein the first period of time is based on historic sleep data and historic wake data of the user (see [0140]; determined alertness level includes a trend that provides multiple alertness levels).
Regarding claim 9, Shouldice and Swift teach the system of claim 1. Shouldice further teaches wherein the computer system (control system 110 via interface 119) includes at least one input element (170) configured to receive the first user input from the user of the computer system (see Shouldice [0032]; electronic interface 119 is coupled to or integrated in the user device 170, [0103]; user device 170 can receive the user feedback),
wherein the first user input specifies subjective alertness ratings reported by the user for the first sleep session after waking up from the first sleep session (see Shouldice [0103]; user feedback can include an analog rating of a subjective energy level of the user, [0151]; the historical sleep-session data includes subjective feedback from the at least one person, which may be one or more ratings for how the person grades one or more historical sleep sessions).
Regarding claim 10, Shouldice and Swift teach the system of claim 9. Shouldice further teaches wherein the subjective alertness ratings are ratings of at least one of wakefulness and alertness (see Shouldice [0103]; user feedback can include an analog rating of a subjective energy level of the user, [0151]; the historical sleep-session data includes subjective feedback from the at least one person, which may be one or more ratings for how the person grades one or more historical sleep sessions) selected from a plurality of possible ratings to be selected by the user, wherein the ratings are numeric values (see Shouldice [0170]; the user device 170 can receive user feedback via alphanumeric text).
Regarding claim 13, Shouldice and Swift teach the system of claim 1. Shouldice further teaches wherein the computer system (110) is further configured to present the first output from the first version of the machine learning model to the user (Fig. 7; step 706) based on a determination that the user has woken up from the first sleep session (see Shouldice [0142]; at step 706, a predicted length of the remaining duration of the sleep session is taken into account on whether to generate the response; for example, if the user wakes up for a bathroom break, the response may be generated to notify the user before the user goes back to sleep).
Regarding claim 15, Shouldice and Swift teach the system of claim 1. Shouldice further teaches wherein the computer system comprises at least one of:
a phone device of the user (see Shouldice [0064]; user device 170 can be a mobile device such as a smartphone);
a home-automation hub (see Shouldice [0064]; user device 170 can be a smart home device e.g. a smart speaker such as Google Home, Amazon Echo, Alexa, etc.); or
a server physically separate from the sensor and connected to the sensor by a data network (see Shouldice [0065]; the control system 110 or a portion thereof can be located in a cloud located in one or more servers).
Regarding claim 18, Shouldice teaches a method for determining alertness levels of a user, the method comprising:
receiving, by a computing system (control system 110 via interface 119) and from at least one sensor (130) integrated with a bed (232, see [0063]; at least one of the sensors 130 is positioned generally adjacent to the user during the sleep session, e.g. coupled to the mattress), first sensor readings of a user during a first sleep session (see Fig. 7, [0129]; step 702 receive data associated with a user during a sleep session);
providing, by the computing system and as first input, the first sensor readings to a first version of a machine learning model that was trained using machine learning techniques to predict alertness levels of the user (see Fig. 7, [0140]; step 704 determine an alertness level of the user using a machine learning model that takes as input the received data from step 702) and includes one or more parameters estimated based at least in part on physical phenomenon of the user and historic data about at least one of the user and a population of users (see [0129-0133]; receiving data associated with the user and examples thereof, Fig. 8, [0151-0153]; step 802 receive historical sleep-session data associated with at least one person for a plurality of historical sleep sessions);
receiving, by the computing system and as first output from the first version of the machine learning model, first data indicating a first predicted alertness level of the user for a first period of time that starts after the user wakes up from the first sleep session (see (see Fig. 7, [0144-0147]; step 706; generate a response based at least in part on the determined alertness level);
receiving, first user input that specifies a first alertness rating reported by the user for the first sleep sessions (see [0103]; user feedback can include an analog rating of a subjective energy level of the user, [0151]; the historical sleep-session data includes subjective feedback from the at least one person, which may be one or more ratings for how the person grades one or more historical sleep sessions);
modifying the first version of the machine learning model to create a second version of the machine learning model (see [0160]; the machine learning model is trained in successive iterations such that historical sleep-session data and historical alertness data are used to train the machine learning model in a first iteration),
by adjusting at least one of the one or more parameters of the machine learning model based, at least in part, on the first alertness rating (see [0159-0161]; the control system can update the machine learning model by further training the ML model using the historical sleep-session data and historical alertness associated with the present user); and
receiving, from the at least one sensor (130), second sensor readings of the user during a second sleep session (see [0160-0162]; the machine learning model is trained iteratively using historical sleep-session data associated with a user);
providing, as second input, the second sensor readings to the second version of the machine learning model (see Fig. 7, [0129]; step 702 receive data associated with a user during a sleep session, it can be appreciated that due to the iterative nature of the machine learning model the processes in Fig. 7 and 8 are repeated more than once); and
receiving, as second output from the second version of the machine learning model, second data indicating a second predicted alertness level of the user (see Fig. 7, [0140]; step 704 determine an alertness level of the user using a machine learning model that takes as input the received data from step 702) for a second period of time that starts after the user wakes up from the second sleep session (see [0140]; alertness determinations in the future can be obtained approximately the same time after the user wakes from sleep).
Shouldice teaches updating the machine learning model using historical sleep-session data associated with the user in addition to historical sleep-session data associated with a population in order to reduce the error rate of the model as it is trained to tailor predictions for the individual user (Shouldice [0161]). With training, the machine learning model predicts an alertness score that takes into account not just raw alertness data but also health condition and/or demographic data (Shouldice [0162]).
Shouldice is silent regarding wherein the machine learning model determines that the first alertness rating from the user is outside a threshold range of the first predicted alertness level; and
in response to determining that the first alertness rating is outside the threshold range of the first predicted alertness level, modifying the first version of the machine learning model to create a second version of the machine learning model.
Swift teaches a system which uses a machine learning model to predict wakefulness windows for a user (Swift [0020]) using patient data to train/update the model in real time (Swift [0025]). Swift’s model is configured to receive first user input that specifies a first alertness rating reported by the user (see Swift [0045]; the program code tunes the model based on on user feedback),
determine that the first alertness rating from the user is outside a threshold range of the first predicted alertness level (see Swift [0045]; the model may predict a wakefulness window for a user and the patient may not actually be awake, this data can be provided to the program code via a computing interface); and
in response to determining that the first alertness rating is outside the threshold range of the first predicted alertness level (see Swift [0045]; upon receipt of this data, the program code analyzes the discrepancy and revises the model in accordance with this inconsistency; it can be appreciated that the threshold in this case is considered to be the state of wakefulness of the user and when the user is not awake, the rating is outside of the threshold), modifying the first version of the machine learning model to create a second version of the machine learning model (see Swift [0045]; the program code can combine predictions for defined time windows with baseline probabilities that are improved over time, and revise/correct these values by obtaining feedback from individual users).
It would have been obvious for one of ordinary skill in the art prior to the effective filing date of the claimed invention to modify Shouldice’s iterative machine learning model for predicting alertness of a user by comparing a predicted alertness/wakefulness level with feedback from users as taught by Swift. One of ordinary skill in the art would have been motivated to make this modification in order to continuously improve the machine learning model and increase model personalization for the individual user (Swift [0045]).
Regarding claim 20, Shouldice teaches a computer-implemented system (100), comprising:
one or more processors (112); and
one or more computer-readable devices including instructions that, when executed by the one or more processors, cause the computer-implemented system to perform operations (see [0031]; the memory device 114 stores machine-readable instructions that are executable by the processor 112 of the control system 110) that include:
receiving, from at least one sensor (130) integrated with a bed (232, see [0063]; at least one of the sensors 130 is positioned generally adjacent to the user during the sleep session, e.g. coupled to the mattress), first sensor readings of a user during a first sleep session (see Fig. 7, [0129]; step 702 receive data associated with a user during a sleep session);
providing, as first input, the first sensor readings to a first version of a machine learning model that was trained using machine learning techniques to predict alertness levels of the user (see Fig. 7, [0140]; step 704 determine an alertness level of the user using a machine learning model that takes as input the received data from step 702) and includes one or more parameters estimated based at least in part on physical phenomenon of the user and historic data about at least one of the user and a population of users (see [0129-0133]; receiving data associated with the user and examples thereof, Fig. 8, [0151-0153]; step 802 receive historical sleep-session data associated with at least one person for a plurality of historical sleep sessions);
receiving, as first output from the first version of the machine learning model, first data indicating a first predicted alertness level of the user for a first period of time that starts after the user wakes up from the first sleep session (see (see Fig. 7, [0144-0147]; step 706; generate a response based at least in part on the determined alertness level);
receiving first user input that specifies a first alertness rating reported by the user for the first sleep sessions (see [0103]; user feedback can include an analog rating of a subjective energy level of the user, [0151]; the historical sleep-session data includes subjective feedback from the at least one person, which may be one or more ratings for how the person grades one or more historical sleep sessions);
modifying the first version of the machine learning model to create a second version of the machine learning model (see [0160]; the machine learning model is trained in successive iterations such that historical sleep-session data and historical alertness data are used to train the machine learning model in a first iteration),
by adjusting at least one of the one or more parameters of the machine learning model based, at least in part, on the first alertness rating (see [0159-0161]; the control system can update the machine learning model by further training the ML model using the historical sleep-session data and historical alertness associated with the present user); and
receiving, from the at least one sensor (130), second sensor readings of the user during a second sleep session (see [0160-0162]; the machine learning model is trained iteratively using historical sleep-session data associated with a user);
providing, as second input, the second sensor readings to the second version of the machine learning model (see Fig. 7, [0129]; step 702 receive data associated with a user during a sleep session, it can be appreciated that due to the iterative nature of the machine learning model the processes in Fig. 7 and 8 are repeated more than once); and
receiving, as second output from the second version of the machine learning model, second data indicating a second predicted alertness level of the user (see Fig. 7, [0140]; step 704 determine an alertness level of the user using a machine learning model that takes as input the received data from step 702) for a second period of time that starts after the user wakes up from the second sleep session (see [0140]; alertness determinations in the future can be obtained approximately the same time after the user wakes from sleep).
Shouldice teaches updating the machine learning model using historical sleep-session data associated with the user in addition to historical sleep-session data associated with a population in order to reduce the error rate of the model as it is trained to tailor predictions for the individual user (Shouldice [0161]). With training, the machine learning model predicts an alertness score that takes into account not just raw alertness data but also health condition and/or demographic data (Shouldice [0162]).
Shouldice is silent regarding wherein the machine learning model determines that the first alertness rating from the user is outside a threshold range of the first predicted alertness level; and
in response to determining that the first alertness rating is outside the threshold range of the first predicted alertness level, modifying the first version of the machine learning model to create a second version of the machine learning model.
Swift teaches a system which uses a machine learning model to predict wakefulness windows for a user (Swift [0020]) using patient data to train/update the model in real time (Swift [0025]). Swift’s model is configured to receive first user input that specifies a first alertness rating reported by the user (see Swift [0045]; the program code tunes the model bason on user feedback),
determine that the first alertness rating from the user is outside a threshold range of the first predicted alertness level (see Swift [0045]; the model may predict a wakefulness window for a user and the patient may not actually be awake, this data can be provided to the program code via a computing interface); and
in response to determining that the first alertness rating is outside the threshold range of the first predicted alertness level (see Swift [0045]; upon receipt of this data, the program code analyzes the discrepancy and revises the model in accordance with this inconsistency; it can be appreciated that the threshold in this case is considered to be the state of wakefulness of the user and when the user is not awake, the rating is outside of the threshold), modifying the first version of the machine learning model to create a second version of the machine learning model (see Swift [0045]; the program code can combine predictions for defined time windows with baseline probabilities that are improved over time, and revise/correct these values by obtaining feedback from individual users).
It would have been obvious for one of ordinary skill in the art prior to the effective filing date of the claimed invention to modify Shouldice’s iterative machine learning model for predicting alertness of a user by comparing a predicted alertness/wakefulness level with feedback from users as taught by Swift. One of ordinary skill in the art would have been motivated to make this modification in order to continuously improve the machine learning model and increase model personalization for the individual user (Swift [0045]).
Claims 4 and 5 are rejected under 35 U.S.C. 103 as being unpatentable over Shouldice (WO 2021171266 A1), with reference to corresponding US application US 20230128912 A1 for citation purposes, in view of Swift et al (US 20200219616 A1) and Kenyon et al (US 20170238868 A1).
Regarding claim 4, Shouldice and Swift teach the system of claim 1, wherein the first and second predicted alertness levels are numeric values on a scale of 1-100 (see Shouldice [0144]; alertness score is 80/100). They are silent regarding wherein the numeric values on a scale of 1-10, wherein a numeric value of 1 represents a highest level of alertness and a numeric value of 10 represents a lowest level of alertness.
Kenyon teaches an alertness prediction system using physiological data and user feedback (Kenyon, Abstract) wherein the predicted alertness levels are numeric values on a scale of 1-10, wherein a numeric value of 1 represents a highest level of alertness and a numeric value of 10 represents a lowest level of alertness (see Kenyon [0075]; alertness risk scale from 0-10, wherein 10 represents the highest fatigue risk).
It would have been obvious for one of ordinary skill in the art prior to the effective filing date of the claimed invention to modify Shouldice’s alertness prediction system with the numerical scale to quantify fatigue risk as taught by Kenyon. One of ordinary skill in the art would have been motivated to make this modification in order to present the predicted alertness of a user in an easily discernable fashion using numerical scales.
Regarding claim 5, Shouldice and Swift teach the system of claim 1. They are silent regarding wherein the model is a two-process model (TPM).
Kenyon teaches wherein the model is a two-process model (see Kenyon Fig. 6; two-process algorithmic model 618).
It would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system for predicting alertness levels of Shouldice by using a two-process model as described by Kenyon. One of ordinary skill in the art would have been motivated to make this modification in order to refine the predicted alertness levels by weighing parameters in a non-linear manner using pattern recognition or machine learning techniques via a two-process algorithmic model (Kenyon [0074]).
Claims 14 and 16 is/are rejected under 35 U.S.C. 103 as being unpatentable over Shouldice (WO 2021171266 A1), with reference to corresponding US application US 20230128912 A1 for citation purposes, in view of Swift et al (US 20200219616 A1) and Sayadi et al (US 20200205580).
Regarding claims 14 and 16, Shouldice and Swift teach the system of claim 1. Shouldice teaches wherein the one or more sensors (130) may include a pressure sensor (132), and wherein at least one of the one or more sensors (130) is coupled to the mattress (Shouldice [0063]). However, they are silent explicitly regarding wherein the integrated sensor is a pressure sensor of the bed; and
the system further comprising a mattress with at least one air chamber, wherein the at least one sensor is a pressure sensor in fluid communication with the air chamber.
Sayadi teaches a system for sensing sleep readings of a user and using machine learning to analyze and predict adjustments for the user’s comfort (Sayadi Abstract, [0004], [0031]) comprising a mattress (112) with at least one air chamber (114A/114B) wherein the at least one sensor is a pressure sensor (146) of the bed/mattress (see Sayadi [0038]; pressure transducers can be incorporated into the air bed system) in fluid communication with the air chamber (see Sayadi [0038-0042]; the pressure transducer(s) are associated with the air chamber(s) and can send pressure readings to the processor during deflation/inflation)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Shouldice with a mattress with at least one air chamber, and a pressure sensor in fluid communication with the air chamber as described by Sayadi. One of ordinary skill in the art would have been motivated to make this modification in order to collect information about the user which can be analyzed to determine various states of the user lying on the bed including awake, light sleep, deep sleep, etc. to be used in subsequent machine learning models (Sayadi [0042]).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ALISHA J SIRCAR whose telephone number is (571)272-0450. The examiner can normally be reached Monday - Thursday 9-6:30, Friday 9-5:30 CT.
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, Benjamin Klein can be reached at 571-270-5213. 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.
/A.J.S./Examiner, Art Unit 3792
/MICHAEL W KAHELIN/Primary Examiner, Art Unit 3792