DETAILED ACTION
This communication is a Non-Final Office Action rejection on the merits. Claims 22, 24-27, 31-32, 34-36, 40, and 42-47 are currently pending and have been addressed below.
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
Applicant's arguments filed 02/20/2026 (related to the 112 Rejection) have been fully considered and are persuasive. The amended claim further specifies that the event is selected in based on (i) the first geolocation corresponding to the first one or more locations at which the first type of the events occurred or the second one or more locations at which the second type of the events occurred and (ii) a first time corresponding the timing. Therefore, the 112 Rejection has been withdrawn.
Applicant's arguments filed 02/20/2026 (related to the 103 Rejection) have been fully considered and are persuasive. The combination of Smith et al., Maliwauki, and Kurata et al. does not teach or suggest “a first operation mode configured for automatic generation of a time record event, and a second operation mode configured for generation of the time record event based on approval of the event received via the client device” and “specific feedback provided from the manual entry such as an updated event comprising at least one of a second geolocation to replace the first geolocation or a second time to replace the first time.” Although Kurata discloses a feedback indicative of an accuracy (e.g., whether the predicted event was correct or incorrect), the combination of Smith et al., Maliwauki, and Kurata et al. does not specifically disclose wherein the user feedback includes information indicating the correct location or time (see Applicant’s specification, Paragraph 0035). Therefore, claim 1 has potential allowable subject matter. Claim 32 and 40 recite similar limitations and therefore have Potential Allowable Subject Matter for the same reasons as claim 22. Claims 24-27, 31, 34-36, and 42-47 have Potential Allowable Subject Matter because of their dependency from independent claims 22 and 32.
Applicant's arguments filed on 02/20/2026 (related to the 101 Rejection) have been fully considered but they are not persuasive.
Applicant states, on pages 11-14, that the claims do not recite a mathematical concept and at most simply "involve[] a mathematical concept." For example, the pending claims are directed to machine learning (ML)-based event prediction, identification of geolocations using GPS sensors, detection of operational modes configured for automatic generation of event records (e.g., based on system-determined conditions), transmission and receipt of signals to and from client devices, and retraining of ML models to improve predictive accuracy. Clearly, such features are not directed to any "numerical formula or equation" or a "series of mathematical calculations," and do not recite any "mathematical operation (such as multiplication) or an act of calculating using mathematical methods to determine a variable or number." Further, Applicant respectfully submits that the claims are not directed towards a method of organizing human activity, as alleged by the Office Action. Instead, the claims are directed to a computing architecture that provides specific improvements in reducing network infrastructure load via event record prediction, operational model detection, client device interaction, and model retraining. (See, e.g., Specification, paras. [0021]-[0023], [0037], [0074]). MPEP 2106.04(a)(2) specifies that that "methods of organizing human activity" are "limited to activity that falls within the enumerated sub-groupings of fundamental economic principles or practices, commercial or legal interactions, and managing personal behavior and relationships or interactions between people, and is not to be expanded beyond these enumerated sub-groupings except in rare circumstances." Therefore, the claims do not recite any abstract idea and thus are not directed to a judicial exception.
Examiner respectfully disagrees with Applicant. These claim elements are considered to be abstract ideas because they are directed to a method of organizing human activity which include managing personal behavior. In this case, the limitations of “identifying a first geolocation” and “predicting events according to a sequence of the time record events” is a form of managing personal behavior because the limitations are considering historical patterns to predict behavior of the user (e.g., analyzing user behaviors/patterns). Also, the limitation of “detecting an operation mode” is merely following rules or instructions (e.g., rules for automatic generation of the time record event). If a claim limitation, under its broadest reasonable interpretation, covers managing personal behavior, then it falls within the “method of organizing human activity” grouping of abstract ideas. Further, the limitations of "receiving feedback regarding the event" and "updating the machine learning model, the updated machine learning model configured to predict events with a level of accuracy that is greater than the machine learning model" are directed to mathematical concepts which include mathematical calculations (see 2024 AI Guidance, Example 47, mathematical calculations to iteratively adjust values/patterns). Accordingly, the claim recites an abstract idea.
Applicant further states, on pages 14-18, that Applicant's claimed technology provides an improved machine learning modeling architecture that enhances or improves performance of machine learning models in performing specific tasks. For example, Applicant's claims are directed to a technically improved computing architecture that uses machine learning models trained on different event types and locations (e.g., both first and second event types and corresponding geolocation data) to predict whether particular event types are likely to occur, employs operational mode detection to determine whether automatic event generation modes or manual modes are activated, transmits and receives feedback ( e.g., updated event locations or times) indicative of event record generation accuracy during such automatic generation modes, and retrains or updates the machine learning models to provide an improved model configured to predict events with a greater level of accuracy. Indeed, Applicant's claims are directed to an improved machine learning architecture that uses executable logic to operate autonomously at the system level, thereby addressing technical challenges associated with both network infrastructure load during peak traffic hours as well as machine learning model configurations that fail to accurately predict different event types based on current or historical timing and location data. (Specification, paras. [0021 ]-[0023 ], [003 7], [007 4 ]). Therefore, the claims recite a similar technical improvement found to integrate the claims of Desjardins into a practical application because both the present claims and the claims of Desjardins involve ML architectures that improve model performance itself.
Examiner respectfully disagrees with Applicant. The additional element of a machine learning model is merely used to predict a timing of suggested events according the modeling (Paragraph 0006). In this case, the machine learning model includes inputs (e.g., time record events corresponding to geolocations according to a sequence) and outputs (e.g., predicted events according to a sequence of the time record events). Although the machine learning model receives feedback over time to improve accuracy (Paragraph 0054, dynamically updating the machine learning with the manually updated time record events), the claim does not recite specific details about how the trained machine learning model operates, which is merely claiming the idea of a solution or outcome (MPEP 2106.05a).
The plain meaning of “providing feedback” is merely describing how the machine learning is receiving continuous data to iteratively learn about the user’s behavior (see Example 47, claim 2 of the 2024 AI Guidance). However, “providing feedback to improve a level of accuracy” is a “well known” process in the art of machine learning (MPEP 2106.05d). Also, it’s considered a well-understood, routing, and conventional function since it’s just “performing repetitive calculations” (MPEP 2106.05d).
Further, the additional element of a global positioning sensor is merely used to receive one or more geolocations and identify a match (Paragraph 0034). This is considered “field of use” at Step 2A, Prong 2, since it’s just used to receive geolocation information, but the sensor is not improved (MPEP 2106.05h). At step 2B, this is a conventional computer function of “receiving and transmitting over a network” (MPEP 2106.05d). Examiner notes that the claim does not include any language stating how the current location of the GPS is used to make new predictions.
Therefore, Examiner concludes that the claim fails to recite any improvements to another technology or technical field, improvements to the functioning of the computer itself, use of a particular machine, effecting a transformation or reduction of a particular article to a different state or thing, adding unconventional steps that confine the claim to a particular useful application, and/or meaningful limitations beyond generally linking the use of an abstract idea to a particular environment. See 84 Fed. Reg. 55. Viewed individually or as a whole, these additional claim element does not provide meaningful limitations to transform the abstract idea into a patent eligible application of the abstract idea such that the claim amounts to significantly more than the abstract idea itself. The claim is not patent eligible.
Independent claims 32 and 40 recite similar features and therefore are rejected for the same reasons as independent claim 22. Claims 23-27, 31, 34-36, and 42-47 are rejected for having the same deficiencies as those set forth with respect to the claims that they depend from, independent claims 22 and 32.
Applicant further states, on pages 18-20, that claim 44 identifies a further technical improvement to "how the machine learning model itself operates," analogous to the patent-eligible claims in Desjardins. (See Desjardins, at page 9). For example, claim 44 explains that "the machine learning model comprises a first layer corresponding to the first type of the events and a second layer corresponding to the second type of the events," and "the probability for the timing of each of the first type of the events and the second type of the events is determined using the first layer corresponding to the first type of the events and the second layer corresponding to the second type of the events." (Emphasis added). Accordingly, the technical solutions presented in claim 44 improves the predictive accuracy and efficiency of the machine learning models by using distinct layers (e.g., FCNN or RNN layer groups) to predict event timing for particular event types, thereby improving model accuracy and performance. (Specification, paras. [0081 ]-[0082]).
Examiner respectfully disagrees with Applicant. Although claim 44 further specifies wherein the machine learning may be a fully connected neural network (FCNN) or a recurrent neural network (RNN), the claim and specification do not recite how the FCNN or RNN is improved. In this case, Examiner notes that those models are “well-known” model used to analyze the sequence of the data. Merely stating that the step is performed by a computer component (e.g. FCNN or RNN) results in “apply it” on a computer (MPEP 2106.05f). Thus, the claim is ineligible.
Examiner recommends to follow Example 47, claim 3 of the 2024 AI Guidance (if supported by the specification).
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 22, 24-27, 31-32, 34-36, 40, and 42-47 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e. an abstract idea) without reciting significantly more.
Independent Claim 1
Step One - First, pursuant to step 1 in the January 2019 Revised Patent Subject Matter Eligibility Guidance (“2019 PEG”) on 84 Fed. Reg. 53, the claim 1 is directed to an apparatus which is a statutory category.
Step 2A, Prong One - Claim 1 recites: A system to: predict, using a model trained using time record events corresponding to geolocations, events according to a sequence of the time record events, wherein the geolocations comprise a first one or more locations at which a first type of the events occurred and a second one or more locations at which a second type of the events occurred; determine, using the model, a probability for a timing of each of the first type of the events and the second type of the events based on the sequence of the time record events; identify a first geolocation associated with a profile; select, according to the model, an event of the events based on (i) the first geolocation corresponding to the first one or more locations at which the first type of the events occurred or the second one or more locations at which the second type of the events occurred and (ii) a first time corresponding the timing; identify a first operation configured for automatic generation of a time record event, and a second operation configured for generation of the time record event based on approval of the event received via the client device; detect that the first operation configured for automatic generation of the time record event is active and the second operation is inactive; transmit, responsive to selection of the event and in accordance with the first operation configured for automatic generation of the time record event, one or more first signals to display the event; generate, subsequent to display of the event and responsive to detection of the first operation, the time record event based on the event; receive, subsequent to generation of the time record event, an updated event comprising at least one of a second geolocation to replace the first geolocation or a second time to replace the first time, the updated event associated with the profile to generate updated time record events, the updated event indicative of an accuracy of the event predicted by the model; retrain, responsive to generation of the updated time record events, the model using the time record events and the updated time record events to provide an updated model, the updated model configured to predict events with a level of accuracy that is greater than the machine learning model; select, responsive to receipt of the updated event, a new event of the events based on the first geolocation and the first time corresponding to the timing; and transmit, responsive to selection of the new event of the events, one or more second signals to display the new event. These claim elements are considered to be abstract ideas because they are directed to a method of organizing human activity which include managing personal behavior. In this case, the limitations of “identifying a first geolocation” and “predicting events according to a sequence of the time record events” is a form of managing personal behavior because the limitations are considering historical patterns to predict behavior of the user (e.g., analyzing user behaviors/patterns). Also, the limitation of “detecting an operation mode” is merely following rules or instructions (e.g., rules for automatic generation of the time record event). If a claim limitation, under its broadest reasonable interpretation, covers managing personal behavior, then it falls within the “method of organizing human activity” grouping of abstract ideas. Further, the limitations of "receiving feedback regarding the event" and "updating the machine learning model, the updated machine learning model configured to predict events with a level of accuracy that is greater than the machine learning model" are directed to mathematical concepts which include mathematical calculations (see 2024 AI Guidance, Example 47, mathematical calculations to iteratively adjust values/patterns). Accordingly, the claim recites an abstract idea.
Step 2A Prong 2 - The judicial exception is not integrated into a practical application. Claim 1 includes additional elements: one or more processors; a memory; a machine learning model; a global positioning sensor associated with the profile; a client device associated with the profile; and a first operation mode and a second operation mode.
The processor is merely used to execute instructions (Paragraph 0095). The memory is merely used to store instructions (Paragraph 0096). The machine learning is merely used to model the time record events and geolocations for each of the number of users (Paragraph 0006). The global positioning sensor is merely used to receive one or more geolocations (Paragraph 0034). The client device is merely used to collect geolocations for user for each of time record events (Paragraph 0034). The first operation mode and the second operation model are merely used to determine whether the user has selected automatic approval of suggested events (Paragraph 0091-0092). Merely stating that the step is performed by a computer component results in “apply it” on a computer (MPEP 2106.05f). These elements of “processor,” “memory,” “machine learning,” “global positioning system,” “client device,” and “operation modes” are recited at a high level of generality such that it amounts no more than mere instructions to apply the exception using a generic computer element. Although the machine learning model receives feedback over time (Paragraph 0054), the claim does not include any specific details of how the machine learning operates or how the predictions are made (see 2024 AI Guidance, Example 47). Also, the global positioning sensor and the client device are considered “field of use” since it’s just used to gather information and/or display information, but the technology is not improved (MPEP 2106.05h). Accordingly, alone and in combination, these additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. Therefore, the claims are directed to an abstract idea.
Step 2B - The claim does not include additional elements that are sufficient to amount significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the claims describe how to generally “apply” the concept of predicting events according to a sequence of the time record events. The specification shows that the processor is merely used to execute instructions (Paragraph 0095). The memory is merely used to store instructions (Paragraph 0096). The machine learning is merely used to model the time record events and geolocations for each of the number of users (Paragraph 0006). The global positioning sensor is merely used to receive one or more geolocations (Paragraph 0034). The client device is merely used to collect geolocations for user for each of time record events (Paragraph 0034). The first operation mode and the second operation model are merely used to determine whether the user has selected automatic approval of suggested events (Paragraph 0091-0092). In this case, “generating event time records based on (i) the first geolocation corresponding to the first one or more locations at which the first type of the events occurred or the second one or more locations at which the second type of the events occurred and (ii) a first time corresponding the timing” is a “well known” process in the art (MPEP 2106.05d). For example, this is similar to the process of generating a time record event in response to detecting that the user geolocation matches the workplace. Also, “providing feedback to improve a level of accuracy” is considered a conventional computer function of “performing repetitive calculations.” Thus, nothing in the claim adds significantly more to the abstract idea. The claim is not patent eligible.
Independent claim 32 is directed to a method at step 1, which is a statutory category. Claim 32 recites similar limitations as claim 1 and is rejected for the same reasons at step 2a, prong one; step 2a, prong 2; and step 2b. The claim is not patent eligible.
Independent claim 40 is directed to an article of manufacture at step 1, which is a statutory category. Claim 40 recites similar limitations as claim 1 and is rejected for the same reasons at step 2a, prong one; step 2a, prong 2; and step 2b. Claim 40 further recites “non-transitory computer-readable medium” – which is treated as just an explicit “processor/computer” for storing instructions and is treated under MPEP 2106.05f in the same manner as claim 1. Accordingly, this additional element of “non-transitory computer-readable medium” is viewed as “apply it on a computer” at step 2a, prong 2 and step 2b. The claim is not patent eligible.
Dependent claims 25, and 34 are not directed to any additional claim elements. Rather, these claims offer further descriptive limitations of elements found in the independent claims and addressed above - such as: comprising the one or more processors to display the event on the client device associated with the profile responsive to the first geolocation corresponding to a geolocation of the event. These processes are similar to the abstract idea noted in the independent claim because they further the limitations of the independent claim which are directed to certain methods of organizing human activity which include managing personal behavior. Also, “controlling the timing of the display of acquired content” is still directed to an abstract idea (MPEP 2106.04a). In addition, there are no additional elements to consider at Step 2A Prong 2 and Step 2B. Therefore, the claims still recite an abstract idea that can be grouped into certain methods of organizing human activity.
Dependent claims 24, 26-27, 31, 35-36, and 42-47 are not directed to any additional claim elements. Rather, these claims offer further descriptive limitations of elements found in the independent claims and addressed above - such as: wherein the machine learning model comprises multimodal multi-task learning; wherein the machine learning model trained using the time record events corresponding to the geolocations is trained for each of a plurality of profiles associated with client devices; determine, using the machine learning model, the probability for the timing of each of the events based on the sequence of the time record events and based on the profile; wherein the machine learning model is a recurrent neural network; wherein the machine learning model is trained using hidden states of time record events in a sequence of time record events, the hidden states corresponding to previous geolocations in the sequence of time record events; wherein the machine learning model comprises a first layer corresponding to the first type of the events and a second layer corresponding to the second type of the events, and wherein the probability for the timing of each of the first type of the events and the second type of the events is determined using the firstlayer corresponding to the first type of the events and the second layer corresponding to the second type of the events; wherein the first type of event is a clock-in event and the second type of event is a clock-out event. Although the additional descriptive limitations indicate inputs and outputs of the machine learning, the claims do not describe any specific details of how the machine learning operates (see 2024 AI Guidance, Example 47). Also, claim 44 and claim 46 are merely describing how the data is segmented into different layers (e.g., similar to clustering data which is a mathematical calculation). Merely stating that the step is performed by a computer component (e.g., machine learning comprising a multimodal multi-task learning or a recurrent neural network) results in “apply it” on a computer (MPEP 2106.05f) being applicable at both Step 2A, Prong 2 and Step 2B. Thus, nothing in the claim adds significantly more to the abstract idea. The claim is ineligible.
Potential Allowable Subject Matter
The closest prior art is Smith et al. (US 2013/0297551 A1), Smith et al. discloses a system, comprising: one or more processors, coupled with memory, to (see Figure 1 and related text in Paragraph 0063, The computer-executable component is preferably a general or application specific processor, but any suitable dedicated hardware or hardware/firmware combination device may alternatively or additionally execute the instructions):
predict, using a machine learning model trained using time record events corresponding to geolocations, events according to a sequence of the time record events, wherein the geolocations comprise a first one or more locations at which a first type of the events occurred and a second one or more locations at which a second type of the events occurred (Paragraph 0028, A planning worker module preferably ingests location information from the location log to generate location pattern predictions. Location pattern predictions are preferably location patterns such as times and days of the week when the user is at home, work, school, the gym, or other locations. The patterns also characterize when a user will leave a location and which location they will be going to next. The modeling of the patterns preferably relies on a plurality of pattern models. A pattern model is preferably an algorithmic model that predicts location based on past location information. A pattern model may be geared to a specific type of pattern such as time frame or location type. The pattern models are preferably Markov chains, but alternatively a pattern model may be a neural network, a statistical model, machine learning approach, or any suitable pattern model; Examiner interprets work as the first geolocation and gym as the second geolocation);
determine, using the machine learning model, a probability for a timing of each of first type of the events and the second type of the events based on the sequence of the time record events (Paragraph 0028, The pattern models are preferably Markov chains, but alternatively a pattern model may be a neural network, a statistical model, machine learning approach, or any suitable pattern model. In one preferred implementation, each Markov chain is preferably targeted at different time frames; Paragraph 0029, An event worker module can be configured to identify particular events such as "leaving work", "leaving home", "arriving at work", "arriving home", "going grocery shopping", and other event types. The events are preferably transitional events that signal a time and location parameter. The time parameter preferably indicates if the user is traveling to the location, returning from that location, or in progress of being at that location; Paragraph 0039, The location predictions may be for various times throughout the day, week, or month. Preferably, the location prediction is a combination of data from the maintained location patterns and the extracted event models. Locations of event models preferably represent high probability locations; In this case, Examiner notes that the machine learning may predict at what time the user will arrive to the first geolocation and/or the second geolocation);
identify a first geolocation associated with a profile using a global positioning sensor of a client device associated with the profile (Paragraph 0031, The location aware application preferably transmits location information of the device to the location prediction platform 110. The location information is preferably in the form of latitude and longitude coordinates. The location information is preferably retrieved from a location service of the device, which can use GPS, IP location information, cellular triangulation, and/or any suitable geographic location detection technique. The location information is preferably periodically updated, but based on the predicted locations and times, the application can schedule polling of location services to reduce the polling; Paragraph 0038, Preferably, a set of Markov chains is maintained to predict the likelihood of moving another location based on the current location; Paragraph 0040, The current location is preferably the most recently received location information of the user stored in the location log);
select, according to the machine learning model, an event of the events based on (i) the first geolocation corresponding to the first one or more locations at which the first type of the events occurred or the second one or more locations at which the second type of the events occurred and (ii) a first time corresponding the timing (Paragraph 0028, The pattern models are preferably Markov chains, but alternatively a pattern model may be a neural network, a statistical model, machine learning approach, or any suitable pattern model. In one preferred implementation, each Markov chain is preferably targeted at different time frames; Paragraph 0029, An event worker module can be configured to identify particular events such as "leaving work", "leaving home", "arriving at work", "arriving home", "going grocery shopping", and other event types. The events are preferably transitional events that signal a time and location parameter. The time parameter preferably indicates if the user is traveling to the location, returning from that location, or in progress of being at that location; Paragraph 0038, Preferably, a set of Markov chains is maintained to predict the likelihood of moving another location based on the current location; Paragraph 0039, The location predictions may be for various times throughout the day, week, or month. Preferably, the location prediction is a combination of data from the maintained location patterns and the extracted event models. Locations of event models preferably represent high probability locations);
identify a first operation mode configured for automatic generation of a time record event, … (Paragraph 0038, Block S140, which includes an application worker module maintaining user location patterns from the location log, functions to determine location patterns of a user. A pattern worker module of the location prediction platform preferably continuously updates location patterns for new location information. Maintaining user location patterns preferably includes generating predictions from location history stored in the location log. Maintaining a user location pattern preferably includes implementing an algorithmic pattern model. In one variation, Markov chains are used to generate location predictions based on location history, but the pattern model(s) may alternatively be a neural network, a statistical model, a machine learning approach, or any suitable pattern modeling approach. The appropriate Markov chains are preferably updated to reflect new data in the location log, which functions to constantly evolve the Markov chains. The Markov chains can additionally only use location data over a specified time frame to determine the model, which would enable the Markov chains to update to reflect current patterns if the location patterns of a user changes);
detect that the first operation mode configured for automatic generation of the time record event is active … (Paragraph 0037, In one variation, the location log uses the time of the location information as a key and stores the latitude and longitude data along with an account identifier. An application worker module of the location prediction platform preferably actively works on recent updates to the location log. When new location information is generated, an application worker module preferably begins to process the information. A subset of application worker modules may additionally be configured to work on new data of the location log in a set sequence; Paragraph 0038, A pattern worker module of the location prediction platform preferably continuously updates location patterns for new location information. Maintaining user location patterns preferably includes generating predictions from location history stored in the location log. Maintaining a user location pattern preferable includes implementing an algorithm pattern model; The Markov chains can additionally only use location data over a specified time frame to determine the model, which would enable the Markov chains to update to reflect current patterns if the location patterns of a user changes; Examiner interprets “generating new location patterns based on updated location information received from the user device GPS” as the “automatic generation of a time record event”); …;
generate, … responsive to detection of the first operation mode, the time record event based on the event; receive, … an updated event comprising at least one of a second geolocation to replace the first geolocation or a second time to replace the first time, the updated event received from the client device associated with the profile to generate updated time record events, the updated event indicative of … the event predicted by the machine learning model (Paragraph 0037, In one variation, the location log uses the time of the location information as a key and stores the latitude and longitude data along with an account identifier. An application worker module of the location prediction platform preferably actively works on recent updates to the location log. When new location information is generated, an application worker module preferably begins to process the information. A subset of application worker modules may additionally be configured to work on new data of the location log in a set sequence; Paragraph 0038, A pattern worker module of the location prediction platform preferably continuously updates location patterns for new location information. Maintaining user location patterns preferably includes generating predictions from location history stored in the location log. Maintaining a user location pattern preferable includes implementing an algorithm pattern model. In one variation, Markov chains are used to generate location predictions based on location history, but the pattern model(s) may alternatively be a neural network, a statistical model, a machine learning approach, or any suitable pattern modeling approach; Examiner notes that updated location patterns received from the client device GPS associated with the profile are used to further update the machine learning pattern model);
retrain, responsive to generation of the updated time record events, the machine learning model using the time record events and the updated time record events to provide an updated machine learning model, … (Paragraph 0028, The pattern models are preferably Markov chains, but alternatively a pattern model may be a neural network, a statistical model, machine learning approach, or any suitable pattern model. In one preferred implementation, each Markov chain is preferably targeted at different time frames; Paragraph 0034, periodically receiving location information of at least one mobile device S120; Paragraph 0038, The appropriate Markov chains are preferably updated to reflect new data in the location log, which functions to constantly evolve the Markov chains. The Markov chains can additionally only use location data over a specified time frame to determine the model, which would enable the Markov chains to update to reflect current patterns if the location patterns of a user changes);
select, responsive to receipt of the updated event, a new event of the events based on the first geolocation and the first time corresponding to the timing (Paragraph 0028, The patterns also characterize when a user will leave a location and which location they will be going to next; Paragraph 0037, In one variation, the location log uses the time of the location information as a key and stores the latitude and longitude data along with an account identifier. An application worker module of the location prediction platform preferably actively works on recent updates to the location log. When new location information is generated, an application worker module preferably begins to process the information. A subset of application worker modules may additionally be configured to work on new data of the location log in a set sequence; Paragraph 0038, A pattern worker module of the location prediction platform preferably continuously updates location patterns for new location information. Maintaining user location patterns preferably includes generating predictions from location history stored in the location log. Maintaining a user location pattern preferable includes implementing an algorithm pattern model. In one variation, Markov chains are used to generate location predictions based on location history, but the pattern model(s) may alternatively be a neural network, a statistical model, a machine learning approach, or any suitable pattern modeling approach. Preferably, a set of Markov chains is maintained to predict the likelihood of moving another location based on the current location; Examiner notes that the machine learning may use the updated location patterns and/or current location to make new predictions); …
Although Smith et al. discloses to retrain the machine learning model by receiving updated event data from a user device (Paragraphs 0031, 0034, & 0038, updated location patterns received from a client device GPS), Smith et al. does not specifically disclose wherein the updated event data is a correction received from the user indicating the correct location or time (see Applicant’s specification, Paragraph 0092, generate a time clock event using a known manual process).
Maliwauki (US 20240177098 A1). Maliwauki discloses to: identify a first geolocation associated with a profile using a … client device associated with the profile (Paragraph 0023, In one or more embodiments, the descriptions may be based on the location of the mobile device. For example, an employer may have multiple business locations, where each business location is associated with a particular set of jobs/tasks. A donut shop has a different set of jobs/tasks than a corporate office. Therefore, the set of descriptions will differ based on the user's location);
select, according to the machine learning model, an event of the events based on on (i) the first geolocation corresponding to the first one or more locations at which the first type of the events occurred or the second one or more locations at which the second type of the events occurred and (ii) a first time corresponding the timing (Paragraph 0012, The set of likely descriptions is generated from historical time entries and current parameters for the time entry. The historical time entries exhibit patterns in, for example, locations from where the time is entered, time stamps for clock-ins and clock-outs, sequence of jobs, and or any other type of historical patterns. In some instances, machine learning models may be used to learn the historical patterns. The current parameters for the time entry include, for example, current location, current time of day, current day of the week, and or any other current parameters; Examiner interprets the generated “likely descriptions” as the “events”); …
transmit, responsive to selection of the event …, one or more first signals to the client device associated with the profile to cause the client device to display the event (Paragraph 0005, or instance, when a user clocks-in using the application, an auto-population engine executed by the mobile device analyzes the history of time entries by the user and generates a list of likely descriptions for the current entry. The list is displayed as selectable objects. When the user selects an object, the corresponding description is automatically added for the current time entry. The generated list may be based on information associated with the current time entry (e.g., time of day, location, etc.) and historical patterns of time entries. One or more machine learning models may be used to learn the historical patterns);
generate, subsequent to display of the event on the client device …, the time record event based on the event (Paragraph 0005, or instance, when a user clocks-in using the application, an auto-population engine executed by the mobile device analyzes the history of time entries by the user and generates a list of likely descriptions for the current entry. The list is displayed as selectable objects. When the user selects an object, the corresponding description is automatically added for the current time entry. The generated list may be based on information associated with the current time entry (e.g., time of day, location, etc.) and historical patterns of time entries. One or more machine learning models may be used to learn the historical patterns; Examiner notes that the time record event is added to the record after the event is displayed to the user and the user selects one of the events);
receive, subsequent to generation of the time record event, an updated event comprising at least one of a second geolocation to replace the first geolocation or a second time to replace the first time, the updated event received from the client device associated with the profile to generate updated time record events, the updated event indicative of … the event [generated] by the machine learning model (Paragraph 0025, In one or more embodiments, the descriptions may be based on one or more machine learning models that learn patterns of the user's time entry behavior. The patterns learned by the machine learning models may include, for example, jobs/tasks associated with the day of the week, time of the day, location, etc. The patterns may further include how the jobs/tasks are sequentially organized, i.e., what type of jobs/tasks would likely follow a job/task or a set of jobs/tasks. Additionally, the learned patterns may also comprise the work behavior of other co-workers. Therefore, any kind of machine learning models that learn the time entry behavior and generate descriptions based on the learned behavior should be considered within the scope of this disclosure; Paragraph 0037, Additionally, the updated GUI 300b displays an overlay 320 that shows likely descriptions (e.g., for the job entry field 312) for the user to select for the current clocked-in session. Generating the overlay 320 greys out one or more other fields, as shown. As shown, the example job selection options provided by the overlay 320 include “Drive time” 322, “Bakery pickup” 324, and “Admin office” 326. The illustrated overlay 320 further provides an option for “Work on something else” 328. Should none of the displayed selection options be applicable to the user, the overlay 320 also provides a cancel button 330, allowing the user to perform a manual entry);
Although the combination of Smith et al. and Maliwauki discloses to retrain the machine learning model by receiving updated event data from a client device GPS or from a user (e.g., a known manual process such as a manual entry), the combination of Smith et al. and Maliwauki does not specifically disclose wherein the updated event data provided to the machine learning model improves a level of accuracy that is greater than the machine learning model.
Kurata et al. (US 2011/0081634 A1). Kurata et al. discloses to: transmit, responsive to selection of the event … generation of the time record event, one or more first signals to the client device associated with the profile to cause the client device to display the event (Paragraph 0110, the score in the score map SM shows a probability of a user being presumed to act according to the behaviour/situation pattern corresponding to the score. That is, the score map SM shows a score distribution of the behaviour/situation patterns according to which a user is presumed to act under the situation of the current location indicated by the geo category code; Paragraph 0111, For example, the probability of a user in a department store at three o'clock on Sunday doing "shopping" is presumed to be high. However, the probability of a user in the same department store at around seven o'clock in the evening "having a meal" is also presumed to be high. As described, the score map SM (to be more precise, a score map SM group) shows the score distribution of a user's behaviour/situation patterns at certain times at certain locations. For example, the score map SM may be input in advance by the user or a third party, may be obtained by using machine learning, or may be built by using other statistical method. Also, the score map SM may be optimised by personal profile information PR or a behaviour/situation feedback FB obtained from the user. The profile information PR includes age, sex, occupation, information on home, and information on workplace, for example. Furthermore, the behaviour/situation feedback FB includes information indicating whether a behaviour/situation pattern that is output is correct or not); …
receive, subsequent to [predicting], an updated [feedback] received from the client device associated with the profile to generate updated [feedback], the updated [feedback] indicative of an accuracy of the event predicted by the machine learning model; retrain, responsive to generation of the updated [feedback], the machine learning model using the time record events and the updated [feedback] to provide an updated machine learning model, the updated machine learning model configured to predict events with a level of accuracy that is greater than the machine learning model (Paragraph 0215, In the example of FIG. 21, the behaviour/situation patterns to be taken into consideration are “sport,” “walk.” “recreation.” “shopping. . . . , “work.” “viewing.” and “sleeping.” However, the types of behaviour/ situation patterns are not limited to the above, and various behaviour/situation patterns as shown in FIG. 26 can be taken into consideration, for example; Paragraph 0233, The determination model created in this manner is expected to output a correct or almost correct behaviour/situation pattern for a feature vector of the same format that is arbitrarily input. Thus, the behaviour/situation recognition unit 112 inputs, to the created determination model, a feature vector formed from sensor data or the like actually observed, and detects a behaviour/situation pattern. If sufficient learning has been performed, a behaviour/situation pattern can be detected by this method with high accuracy. However, the process of creating a determination model by a machine learning algorithm is a process for which the amount of computation is extremely large. Therefore, as has been described with reference to FIGS. 2 to 6, a system configuration has to be modified in a case of using the learning model determination. Furthermore, a method of using the rule-based determination and the learning model determination in combination can also be conceived; Paragraph 0234, Heretofore, the behaviour/situation pattern detection method that uses the learning model determination has been described. As described above, when using the learning model determination, a behaviour/situation pattern can be detected with high accuracy if sufficient learning has been performed. Also, by rebuilding the determination model by taking a feedback from a user into consideration, a determination model capable of detecting a behaviour/situation pattern with further improved accuracy can be created. Accordingly, using the learning model determination is beneficial for improving the accuracy of behaviour/situation pattern detection; Paragraph 0267, As has been described, by using the behaviour/situation pattern detection method according to the present embodiment, it becomes possible to detect a behaviour/situation pattern relating to a user's daily behaviour (HC behaviour) as illustrated in FIG. 26. As a result, it becomes possible to use a user's daily behaviour which is hard to predict from a behaviour history expressed by an accumulation of LC behaviours; Paragraph 0289, The behaviour prediction unit 208 reads information on the behaviour history and location history accumulated in the history storage unit 202, and predicts the behaviour/situation pattern and the location information for the future based on the pieces of information that have been read. For example, the behaviour prediction unit 208 uses the behaviour history and the like read from the history storage unit 202, and predicts the next behaviour/situation pattern of the user based on a stochastic location transition model. As the stochastic location transition model, a method of estimating a transition probability on the basis of location clustering described later is used, for example);
select, responsive to receipt of the updated [feedback], a new event of the events based on the [feedback] (Paragraph 0233, The process of creating a determination model by a machine learning algorithm is a process for which the amount of computation is extremely large. Therefore, as has been described with reference to FIGS. 2 to 6, a system configuration has to be modified in a case of using the learning model determination; Paragraph 0234, Heretofore, the behaviour/situation pattern detection method that uses the learning model determination has been described. As described above, when using the learning model determination, a behaviour/situation pattern can be detected with high accuracy if sufficient learning has been performed. Also, by rebuilding the determination model by taking a feedback from a user into consideration, a determination model capable of detecting a behaviour/situation pattern with further improved accuracy can be created. Accordingly, using the learning model determination is beneficial for improving the accuracy of behaviour/situation pattern detection; Paragraph 0289; The behaviour prediction unit 208 reads information on the behaviour history and location history accumulated in the history storage unit 202, and predicts the behaviour/situation pattern and the location information for the future based on the pieces of information that have been read; Examiner notes that the machine learning may use the updated feedback to make new predictions);
and transmit, responsive to selection of the new event of the events, one or more second signals to the client device to cause the client device to display the new event on the client device associated with the profile (Paragraph 0280, Application display unit 210; Paragraph 0289; The behaviour prediction unit 208 reads information on the behaviour history and location history accumulated in the history storage unit 202, and predicts the behaviour/situation pattern and the location information for the future based on the pieces of information that have been read).
However, the cited art, alone or in any combination, fails to teach or suggest at least: a system, comprising: one or more processors, coupled with memory, to: predict, using a machine learning model trained using time record events corresponding to geolocations, events according to a sequence of the time record events, wherein the geolocations comprise a first one or more locations at which a first type of the events occurred and a second one or more locations at which a second type of the events occurred; determine, using the machine learning model, a probability for a timing of each of the first type of the events and the second type of the events based on the sequence of the time record events; identify a first geolocation associated with a profile using a global positioning sensor of a client device associated with the profile; select, according to the machine learning model, an event of the events based on (i) the first geolocation corresponding to the first one or more locations at which the first type of the events occurred or the second one or more locations at which the second type of the events occurred and (ii) a first time corresponding the timing; identify a first operation mode configured for automatic generation of a time record event, and a second operation mode configured for generation of the time record event based on approval of the event received via the client device; detect that the first operation mode configured for automatic generation of the time record event is active and the second operation mode is inactive; transmit, responsive to selection of the event and in accordance with the first operation mode configured for automatic generation of the time record event, one or more first signals to the client device associated with the profile to cause the client device to display the event; generate, subsequent to display of the event on the client device and responsive to detection of the first operation mode, the time record event based on the event; receive, subsequent to generation of the time record event, an updated event comprising at least one of a second geolocation to replace the first geolocation or a second time to replace the first time, the updated event received from the client device associated with the profile to generate updated time record events, the updated event indicative of an accuracy of the event predicted by the machine learning model; retrain, responsive to generation of the updated time record events, the machine learning model using the time record events and the updated time record events to provide an updated machine learning model, the updated machine learning model configured to predict events with a level of accuracy that is greater than the machine learning model; select, responsive to receipt of the updated event, a new event of the events based on the first geolocation and the first time corresponding to the timing; and transmit, responsive to selection of the new event of the events, one or more second signals to the client device to cause the client device to display the new event on the client device associated with the profile.
Nor does the remaining prior art of record remedy the deficiencies found in the cited prior art. Furthermore, neither the prior art, the nature of the problem, nor knowledge of a person having ordinary skill in the art provides for any predictable or reasonable rationale to combine prior art teachings.
Claims 32 and 40 recite similar limitations and therefore have Potential Allowable Subject Matter for the same reasons as claim 22. Claims 24-27, 31, 34-36, and 42-47 have Potential Allowable Subject Matter because of their dependency from independent claims 22 and 32.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant’s disclosure.
Carzoli et al. (US 2022/0301357 A1) - discloses a geo-fence location technology to verify the geo-location of the mobile device is within a geo-fenced clock-in/clock-out verification area of the remote job location. Once geo-fence location verification is complete and a current time is before or within a predetermined amount of time from a job start time, the employee is logged into the remote employment management system 200. This geo-fenced remote clock-in/clock-out verification eliminates manual time entry and excess paper use, which saves time and money. In addition, the remote employment management system 200 automatically updates a labor record (e.g., Salesforce Labor) with punch times of the remote employees (see at least Paragraph 0056).
Priness et al. (US 2016/0321551 A1) – discloses to predict related contextual or semantic information, such as when a user will likely arrive; how long a user will likely stay; likely activity of the user associated with the location (e.g., exercising at a gym or grocery shopping); or other contextual information. Using information about the user's current location with historical observations about the user, expected user events (e.g., an expected flight or an appointment), and/or other lasting information or ephemeral information (e.g., holidays and traffic, respectively), a prediction of one or more future semantic locations and corresponding confidences may be determined and may be used for providing a personalized computing experience to the user (see at least Paragraph 0004).
Liao (Liao, D., Liu, W., Zhong, Y., Li, J. and Wang, G., 2018, July. Predicting Activity and Location with Multi-task Context Aware Recurrent Neural Network. In IJCAI (pp. 3435-3441)) – discloses a multi-task learning neural network to predict users’ activities and locations simultaneously; Page 3437, 4.1 Overview, In this section, we introduce the Multi-task Context Aware Recurrent Neural Network (MCARNN). The overall framework is presented in Fig. 1. We employ a multi-task learning framework to capture the spatial-activity topic in shared hidden layers for activity and location prediction (see at least Page 3437).
Eggleston (CA 2926074 C) – discloses tracking workers' time has been a necessity since workers began working on an hourly basis, businesses charged on an hourly basis, and businesses tracked labour costs on an hourly basis. However, today, employees, contractors, temporary staff etc. perform multiple tasks on multiple projects each day and may perform them in different locations. As such, for the supervisory role within a business it becomes extremely difficult to verify that a particular individual did actually perform the task they say they did and took the length of time that they say they did or were where they were supposed to be for the allotted time(s). Accordingly, there is a need for a time and location tracking system that addresses these issues by automatically leveraging user location relative to geo-fences established in relation to specific worksites (see at least abstract).
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 MARJORIE PUJOLS-CRUZ whose telephone number is (571)272-4668. The examiner can normally be reached Mon-Thru 7:30 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, Patricia H Munson can be reached at (571)270-5396. 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.
/M.P./
Examiner, Art Unit 3624 /PATRICIA H MUNSON/Supervisory Patent Examiner, Art Unit 3624