Deep learning-based person-specific insight generator

Abstract

A method for generating person-specific insights may include receiving sensor data associated with a device; extracting features from the received sensor data; processing the features using a machine learning model to generate machine learning metrics; incorporating the machine learning metrics and the features to generate insight data associated with the device; generating person data using the insight data and the features and mapping the insights to the person data; generating custom insights using the insight data, the person data, and the features, where the custom insights are text-based summaries; and distributing each of the custom insights to a respective person in the person data to place a service order associated with the device.

Description

[Technical Field]

[0001] The present disclosure generally relates to methods and systems for generating person-specific insights. [Background technology]

[0002] Advancements in technology have increased the complexity of Internet of Things (IoT) devices and inter / intra-system automation. In industrial operations and maintenance settings, there are many personnel required to properly maintain devices / machinery, including mechanical engineers, electrical engineers, maintenance technicians, civil engineers, field service technicians, chemical engineers, drilling technicians, industrial technicians, repair technicians, operations technicians, and control technicians.

[0003] Based on their academic training and work experience, such individuals are typically trained in a specific technology / discipline and the technical language associated with that technology / discipline. Additionally, the same technical term may have different meanings across different technologies / disciplines and individuals. For example, the term "filter" may be understood by a water reliability engineer as a water filter, but may mean something completely different to a control engineer, e.g., the term "filter" may be understood as a predictive algorithm.

[0004] Furthermore, technical terminology can be very specific and precise according to technical discipline. For example, while the word "control" is used to refer to benchmarks in the healthcare industry, the same word is not typically used by maintenance technicians. In addition to language and technical terminology, insights based on the same sensor signal need to be differentiated based on person.

[0005] There is an unmet need for solutions to bridge the dilution of human capacity for systems understanding due to siloed personal knowledge and the lack of experienced cross-domain workers due to retirement. Summary of the Invention [Means for solving the problem]

[0006] An aspect of the present disclosure relates to an innovative method for generating person-specific insights, which may include receiving sensor data associated with a device; cleaning the received sensor data to generate cleansed data; processing the cleansed data using a machine learning model to generate machine learning metrics; ingesting the machine learning metrics to generate insight data associated with the device; generating person data using the insight data and mapping the insights to the person data; generating custom insights using the insight data and the person data, where the custom insights are text-based summaries; and distributing each of the custom insights to a respective person in the person data for placing a service order associated with the device.

[0007] An aspect of the present disclosure relates to an innovative non-transitory computer-readable medium storing instructions for generating person-specific insights, which may include receiving sensor data associated with a device; cleaning the received sensor data to generate cleansed data; processing the cleansed data using a machine learning model to generate machine learning metrics; incorporating the machine learning metrics to generate insight data associated with the device; generating person data by using the insight data and mapping the insights to the person data; generating custom insights using the insight data and the person data, where the custom insights are text-based summaries; and distributing each of the custom insights to a respective person in the person data to place a service order associated with the device.

[0008] An aspect of the present disclosure relates to an innovative server system for generating person-specific insights, which may include receiving sensor data associated with a device; cleaning the received sensor data to generate cleansed data; processing the cleansed data using a machine learning model to generate machine learning metrics; ingesting the machine learning metrics to generate insight data associated with the device; generating person data using the insight data and mapping the insights to the person data; generating custom insights using the insight data and the person data, where the custom insights are text-based summaries; and distributing each of the custom insights to a respective person in the person data to place a service order associated with the device.

[0009] An aspect of the present disclosure relates to an innovative system for generating person-specific insights, which may include means for receiving sensor data associated with a device, means for cleaning the received sensor data to generate cleansed data, means for processing the cleansed data using a machine learning model to generate machine learning metrics, means for incorporating the machine learning metrics to generate insight data associated with the device, means for generating person data by using the insight data and mapping the insights to the person data, means for generating custom insights using the insight data and the person data, where the custom insights are text-based summaries, and means for distributing each of the custom insights to a respective person in the person data to place a service order associated with the device. [Brief explanation of the drawings]

[0010] A general architecture implementing various features of the present disclosure will now be described with reference to the accompanying drawings. The accompanying drawings and related description are provided to illustrate example embodiments of the present disclosure and are not intended to limit the scope of the disclosure. Throughout the accompanying drawings, reference numbers are reused to indicate correspondence between referenced elements.

[0011] [Figure 1] FIG. 1 illustrates an example process flow for generating person-based insights according to an example embodiment.

[0012] [Figure 2] FIG. 2 illustrates an exemplary process flow for machine learning model development and deployment / production in accordance with an exemplary embodiment.

[0013] [Figure 3] FIG. 3 illustrates an exemplary process flow for model development and deployment / production of insight models according to an exemplary embodiment.

[0014] [Figure 4] FIG. 4 illustrates an example process flow for generating a person via a person model according to an example embodiment.

[0015] [Figure 5] FIG. 5 illustrates an exemplary process flow for Transformer model / text generation model development and deployment according to an exemplary embodiment.

[0016] [Figure 6] FIG. 6 illustrates an exemplary process flow for generating personalized text via a Transformer model according to an exemplary embodiment.

[0017] [Figure 7] FIG. 7 illustrates example outputs from various models according to an example embodiment.

[0018] [Figure 8] FIG. 8 illustrates an exemplary computing environment with exemplary computing devices suitable for use in some exemplary embodiments. DETAILED DESCRIPTION OF THE INVENTION

[0019] The following detailed description provides details of the drawings and implementation examples of the present application. Redundant element reference numbers and descriptions between drawings are omitted for clarity. Terms used throughout the description are provided by way of example and are not intended to be limiting. For example, the use of the term "automatic" may include a fully automatic implementation or a semi-automatic implementation that includes user or administrator control over certain aspects of the implementation, depending on the desired implementation of one skilled in the art practicing the implementation of the present application. Selections may be made by a user via a user interface or other input means, or may be implemented by a desired algorithm. The implementations described herein may be utilized alone or in combination, and the functionality of the implementations may be implemented by any means depending on the desired implementation.

[0020] The exemplary embodiments described herein utilize machine learning / deep learning (autoencoders) as translators trained by IoT sensor signals, system logs, maintenance notes, outage events, violations, and technical notes along with corresponding failure modes to facilitate different personas (personas) working in performing system / device maintenance and upkeep. Such personas may include, but are not limited to, electricians, mechanics, maintenance managers, structural engineers, electrical engineers, mechanical engineers, sustainability managers, operators, business owners, etc.

[0021] Exemplary embodiments decipher IoT event digital signals and their equivalent system behaviors during technical triage and provide functional explanations to various parties. Given the complexities involved, taxonomic complexity across disciplines for describing the same set of digital signals / data needs to be encapsulated within an intelligence system with appropriate interpretations of the same terms. For example, the term "filter" in the digital signal processing space includes noise removal in digital signals, while the term "filter" in the fluid processing space includes contamination removal from liquids.

[0022] For cross-functional triage explanations by many people, the goal is to converge their diagnostic conclusions and outcomes to the same failure mode to be resolved. An exemplary embodiment treats each person and each person's taxonomy as a training pair, with different taxonomies and languages ​​(e.g., English, Japanese, French, etc.) for corresponding common sensor data and events. This results in the same batch of event signals. An additional ensemble model will incorporate the results and processes if two or more failure modes reach ranking and priority.

[0023] 1 shows an example process flow 100 for generating person-based insights according to an example embodiment. The process begins in step S102, where data is collected and received from an IoT system. The sensor data may include, but is not limited to, sensor data, control logs, past maintenance data, etc. In step S104, data cleansing is performed to remove defects from the collected / received data to generate cleansed data.

[0024] In step S106, the cleansed data is provided as input to a machine learning model for processing, and machine learning metrics are generated as output from the machine learning model. The generated machine learning metrics may include, but are not limited to, probability of failure, categorization, risk scores, etc.

[0025] In step S108, the insight model takes in the machine learning metrics to generate insight data such as root causes of failures, suggested actions, etc. In step S110, the person model takes the insight data as input to generate a list of people / person data. Person data may have characteristics including, but not limited to, role, experience, technical knowledge, asset expertise, etc.

[0026] In step S112, the insight data and person data are sent to a person-based text generation model to provide custom insights for each person in the person data. The process then proceeds to step S114, where the custom insights are distributed to people via an interface, such as a dashboard or an application programming interface (API).

[0027] 2 shows an exemplary process flow for machine learning model development and deployment / production according to an exemplary embodiment. Under model development, the process begins in step S202, where historical sensor data is retrieved from a database. In step S204, the sensor data is cleaned to remove defects. In step S206, training features are generated based on the cleaned historical sensor data. In step S208, a machine learning model is generated based on the training features, and the machine learning model is trained using the historical sensor data. The development process then proceeds to step S210, where the trained machine learning model is deployed for inference.

[0028] As shown in FIG. 2 , in real-world use, the process flow begins in step S212, where current sensor data is received from a database or directly from the sensor(s). In step S214, the sensor data is cleaned to remove defects. In step S216, features are generated based on the cleansed sensor data. In step S218, machine learning metrics are generated based on the generated features as well as inferences from the deployed machine learning model. The machine learning metrics may include information such as, but not limited to, probability of failure, probability density vector, anomaly score, local outlier factor, etc. In some exemplary embodiments, a class of machine learning algorithms generates significant contributors or features that lead to the calculation of the metrics. The real-world process then proceeds to step S220, where the generated machine learning metrics are transferred via an API.

[0029] FIG. 3 illustrates an exemplary process flow for model development and deployment / realization of an insight model according to an exemplary embodiment. Under model development, the process begins in step S302, where historical data is searched / retrieved from a database. The historical data may include historical information, including, but not limited to, machine learning metrics, feature data, control system log data, past root cause analyses, past insights, rules from subject matter experts (SMEs), failure modes and effects analysis (FMEA), potential cause data from SMEs, etc. In step S304, the historical data is cleaned to remove defects. In step S306, training features are generated based on the cleansed historical data. In step S308, an insight model is generated based on the training features, and the insight model is trained using the historical data. The insight model may use explainable artificial intelligence (AI) for solution agreement and utilize physics to generate explainability. The development process then proceeds to step S310, where the learned insight model is deployed for inference.

[0030] 3, in actual use, the process flow begins in step S312, where current data is searched / retrieved. The current data may include information including, but not limited to, machine learning metrics, feature data, control system log data, past root cause analyses, past insights, rules from subject matter experts (SMEs), failure mode and effect analyses (FMEAs), potential cause data from SMEs, etc. The machine learning metrics, feature data, and control system log data may be grouped as X-data, while the past root cause analyses, past insights, rules from SMEs, FMEAs, and potential cause data from SMEs may be grouped as Y-data.

[0031] In step S314, the current data is cleaned to remove defects. In step S316, features are generated based on the cleansed current data. In step S318, insight data is generated based on inferences from the developed insight model along with the generated features. The real-world process then proceeds to step S320, where the generated insight data is transferred via an API. In some exemplary embodiments, the insight model is a rule-based algorithm as opposed to an ML model.

[0032] The insight data generated may be worth reviewing by individuals with different areas of expertise / disciplines. For example, a high temperature anomaly in a large motor may be reviewed by a mechanical engineer for potential bearing failure as a result of frictional heating. At the same time, the same temperature anomaly may need to be reviewed by an electrical engineer for potential electrical wiring issues in the motor.

[0033] The insight data can be mapped to different people via person models. In some exemplary embodiments, the person models are expert system models or rule-based engines / algorithms. Expert system models are applied in the design of new systems where historical operational data is not readily available or quality data is severely lacking to build reliable data-driven models. Expert system models may be based on information including, but not limited to, operations manuals, maintenance manuals, user manuals, SME knowledge, best practices, etc. In alternative exemplary embodiments where historical operational data is readily available and insights relate to different people, machine learning models (e.g., collaborative filters, etc.) can be utilized in developing the person models.

[0034] FIG. 4 illustrates an exemplary process flow 400 for person generation via a person model according to an exemplary embodiment. The person model performs the functions of identifying a required operator skill set and generating person data / attributes that identify a person possessing the identified required operator skill set. The operator skill set may include information including, but not limited to, a discipline vector, a level of expertise, asset-specific expertise, etc. On the other hand, the person data / attributes may include information including, but not limited to, a person's addressable role, responsibilities, expertise, assets, etc., which may assist in troubleshooting and repairing problematic devices based on insight data. In some exemplary embodiments, the person model may include two components: a skill generator / skill mapping model and a person generator / person mapping model. The skill generator performs the function of identifying a required operator skill set, while the person generator performs the function of person data generation.

[0035] The process flow begins at step S402, where insight data is searched / retrieved. In step S404, the insight data is passed to a skills mapping model to generate a required operator skill set. In step S406, the generated required operator skill set and insight data are passed to a person mapping model to generate person data. In some exemplary embodiments, the person data is in the form of a vector. The process then proceeds to step S408, where the person data is transferred via an API.

[0036] Once the insight data is generated, it is transformed into person-specific language via a Transformer model / language model. In some exemplary embodiments, the Transformer model / language model is a deep learning-based Transformer algorithm. The Transformer model / language model is used to perform text translation and translation customization, for example, via an associated encoder and decoder architecture. A text-based summary is generated as output from the Transformer model / language model for the identified / defined person. For example, customized text may be generated for a mechanical engineer, an electrical engineer, etc., via different Transformer models / language models, where each Transformer model / language model provides translation for a different person. In some exemplary embodiments, an audio or video summary may be provided instead of the text-based summary.

[0037] 5 shows an example process flow 500 for Transformer model / text generation model / language model development and deployment according to an example embodiment. In step S502, past insights are used as input feature data. In step S504, a person-based vocabulary is provided as input in generating person-specific text. The person-based vocabulary may include technical terms and definitions uniquely associated with people from different disciplines.

[0038] In step S506, the Transformer model / text generation model / language model is trained by using past insights and the person-based vocabulary. The Transformer model / text generation model / language model is trained by performing correlation identification between relevant insights and the person-based vocabulary to systematically generate customized / personalized text for each identified person. The taxonomic complexity across disciplines for describing the same set of digital signals / data needs to be encapsulated within the intelligence system with the appropriate interpretation of the same terms.

[0039] In some exemplary embodiments, multiple Transformer models / language models may be developed for each person. In some exemplary embodiments, Generative Pre-trained Transformer 3 (GPT3), an autoregressive language model, may be utilized in developing the Transformer models / language models. In step S508, the trained models are deployed for future use. In step S510, the trained models are made available / transferred via an API.

[0040] FIG. 6 shows an example process flow 600 for generating personalized text via Transformer models / language models according to an example embodiment. In step S602, insight data and person data are retrieved / read from a person model. In step S604, the retrieved / read data is provided as input to a Transformer selector. In step S606, the Transformer selector matches and selects one or more Transformer models / language models depending on the person in the person data. In step S608, each of the Transformer models / language models generates customized / personalized text for the respective person. The generated customized / personalized text is then sent to the person via an interface, such as an API, for consumption / perusal.

[0041] FIG. 7 shows exemplary outputs from various models according to exemplary embodiments. Metrics are generated from the ML model through the performance of feature engineering from the input data. FIG. 7 shows exemplary output from the ML model identifying the brake as the problem component with a failure probability of 0.9. In addition, subcomponents such as the brake liner and actuator are identified and a probability density vector is output. The insight model then uses the metrics to derive insights into the inputs. FIG. 7 shows exemplary output from the insight model identifying the potential problem at hand. Specifically, failures related to the cascade, liner, actuator, and operator.

[0042] The person model generates person data that identifies people with unique skill sets through insight mapping. As shown in Figure 7, the person model generates mechanical engineers, electricians, and equipment operators as outputs based on the provided insights. Finally, the transformer model / language model generates customized / personalized text for each person. For example, a mechanical engineer may receive custom text "Brake pads need to be checked for maintenance or replacement," an electrician may receive custom text "Check the actuator force on the brake pads for optimal brake load. Verify the current and voltage for short circuits," an equipment operator may receive custom text "Check if the equipment is too close to other equipment. Check for abnormal noise and vibration when braking," etc.

[0043] The above-described exemplary embodiments may have various benefits and advantages, such as providing differentiated and customized insights / text to each person to enable efficient troubleshooting and maintenance. Customized text based on person-specific vocabulary and discipline areas allows targeted message delivery in technical language understood by each recipient.

[0044] 8 illustrates an exemplary computing environment having an exemplary computing device suitable for use in some exemplary implementations. The computing device 805 in the computing environment 800 may include one or more processing units, cores, or processors 810, memory 815 (e.g., RAM, ROM, and / or the like), internal storage 820 (e.g., magnetic, optical, solid-state storage, and / or organic), and / or I / O interface 825, any of which may be connected over a communication mechanism or bus 830 to communicate information or may be incorporated within the computing device 805. The I / O interface 825 may also be configured to receive images from a camera or provide images to a projector or display, depending on the desired implementation.

[0045] Computing device 805 may be communicatively connected to input / user interface 835 and output devices / interfaces 840. One or both of input / user interface 835 and output devices / interfaces 840 may be wired or wireless interfaces and may be detachable. Input / user interface 835 may include any physical or virtual device, component, sensor, or interface that can be used to provide input (e.g., buttons, touchscreen interface, keyboard, pointing / cursor control, microphone, camera, Braille, motion sensor, accelerometer, optical reader, and / or the like). Output devices / interfaces 840 may include a display, television, monitor, printer, speaker, Braille, or the like. In some example implementations, input / user interface 835 and output devices / interfaces 840 may be incorporated with or physically connected to computing device 805. In other example implementations, other computing devices may function as or provide the functionality of input / user interface 835 and output device / interface 840 for computing device 805 .

[0046] Examples of computing devices 805 may include, but are not limited to, highly mobile devices (e.g., smartphones, devices in vehicles or other machines, devices carried by humans and animals, and the like), mobile devices (e.g., tablets, notebooks, laptops, personal computers, portable televisions, radios, and the like), and devices not designed for mobility (e.g., desktop computers, other computers, information kiosks, televisions, radios, and the like having one or more processors embedded therein and / or connected thereto).

[0047] Computing device 805 may be communicatively connected (e.g., via I / O interface 825) to external storage 845 and a network 850 for communicating with any number of networked components, devices, and systems, including one or more computing devices of the same or different configurations. Computing device 805 or any connected computing device may function as, provide functionality for, or be referred to as a server, client, thin server, general purpose machine, special purpose machine, or other label.

[0048] I / O interface 825 may include, but is not limited to, wired and / or wireless interfaces using any communication or I / O protocol or standard (e.g., Ethernet, 802.11x, Universal Serial Bus, WiMax, modem, cellular network protocols, and the like) to communicate information to and / or from at least all connected components, devices, and networks in computing environment 800. Network 850 may be any network or combination of networks (e.g., the Internet, a local area network, a wide area network, a telephone network, a cellular network, a satellite network, and the like).

[0049] The computing device 805 may use and / or communicate using computer-usable or computer-readable media, including transitory media and non-transitory media. Transitory media include transmission media (e.g., metallic cables, optical fibers), signals, carrier waves, and the like. Non-transitory media include magnetic media (e.g., disks and tapes), optical media (e.g., CD-ROMs, digital video disks, Blu-ray disks), solid-state media (e.g., RAM, ROM, flash memory, solid-state storage), and other non-volatile storage or memory.

[0050] The computing device 805 can be used to implement techniques, methods, applications, processes, or computer-executable instructions in some exemplary computing environments. The computer-executable instructions can be obtained from transitory media and stored on and retrieved from non-transitory media. The executable instructions can be from one or more of any programming, scripting, and machine language (e.g., C, C++, C#, Java, Visual Basic, Python, Perl, JavaScript, and others).

[0051] The processor 810 may run under any operating system (OS) (not shown) in a native or virtual environment. Along with the OS and other applications (not shown), one or more applications may be deployed, including a logic unit 860, an application programming interface (API) unit 865, an input unit 870, an output unit 875, and an inter-unit communication mechanism 1196 for different units to communicate with each other. The described units and elements may be modified in design, function, configuration, or implementation and are not limited to the provided description. The processor 810 may be in the form of a hardware processor, such as a central processing unit (CPU), or a combination of hardware and software units.

[0052] In some example implementations, information or instructions to execute, once received by API unit 865, may be communicated to one or more other units (e.g., logic unit 860, input unit 870, output unit 875). In some cases, logic unit 860 may be configured to control information flow between units and to control services provided by API unit 865, input unit 870, and output unit 875 in some example implementations described above. For example, the flow of one or more processes or implementations may be controlled by logic unit 860 alone or in cooperation with API unit 865. Input unit 870 may be configured to obtain inputs for computations described in example implementations, and output unit 875 may be configured to provide outputs based on computations described in example implementations.

[0053] The processor 810 may be configured to receive sensor data associated with a device, as shown in FIG. 1. The processor 810 may be configured to cleanse the received sensor data to generate cleansed data, as shown in FIG. 1. The processor 810 may be configured to perform processing to generate machine learning metrics using a machine learning model, as shown in FIG. 1. The processor 810 may be configured to ingest the machine learning metrics to generate insight data associated with the device, as shown in FIG. 1. The processor 810 may be configured to generate people data using the insight data and map the insight to the people data, as shown in FIG. 1. The processor 810 may be configured to generate custom insights using the insight data and the people data, where the custom insights are text-based summaries, as shown in FIG. 1. The processor 810 may be configured to disseminate each custom insight to each person in the people data to place a service order associated with the device, as shown in FIG. 1.

[0054] Some portions of the detailed description are presented in terms of algorithms and symbolic representations of operations within a computer. These algorithmic descriptions and symbolic representations are the means used by those skilled in the data processing arts to convey the essence of their innovations to others skilled in the art. An algorithm is a sequence of defined steps leading to a desired end state or result. In one implementation, the steps performed require physical manipulations of tangible quantities to achieve a tangible result.

[0055] Unless otherwise specified, as will be apparent from the description, the description utilizing terms such as "processing," "computing," "calculating," "determining," "displaying," and the like throughout the description will be understood to include the actions and processes of a computer system or other information processing device that manipulates and converts data represented as physical (electronic) quantities in the registers and memory of the computer system into other data similarly represented as physical quantities in the memory or registers of the computer system, or other information storage, transmission, or display device.

[0056] Implementations may also relate to apparatuses for performing the operations described herein. This apparatus may be specially constructed for the required purposes, or it may comprise one or more general-purpose computers selectively activated or reconfigured by one or more computer programs. Such computer programs may be stored in a computer-readable medium, such as a computer-readable storage medium or a computer-readable signal medium. Computer-readable storage media may include tangible media, such as, but not limited to, optical disks, magnetic disks, read-only memory, random-access memory, solid-state devices and drives, or any other type of tangible or non-transitory medium suitable for storing electronic information. Computer-readable signal media may include media such as carrier waves. The algorithms and displays presented herein are not inherently related to any particular computer or other apparatus. A computer program may include a pure software implementation containing instructions for performing the operations of a desired implementation.

[0057] Various general-purpose systems may be used with the programs and modules according to the examples herein, or it may prove convenient to construct more specialized apparatus to perform the desired method steps. Additionally, the implementations are not described with reference to any particular programming language. It will be understood that a variety of programming languages ​​may be used to implement the teachings of the implementations described herein. Instructions in the programming language may be executed by one or more processing units, such as a central processing unit (CPU), processor, or controller.

[0058] As is known in the art, the operations described above may be performed by hardware, software, or some combination of software and hardware. Various aspects of the implementations may be implemented using circuits and logic (hardware), while other aspects may be implemented using instructions stored on a machine-readable medium (software), which, when executed by a processor, cause the processor to perform methods for carrying out the implementations of the present application. Furthermore, some implementations of the present application may be performed exclusively in hardware, while other implementations may be performed exclusively in software. Furthermore, the various functions described may be performed in a single unit or distributed among several components in any number of ways. When performed by software, the methods may be executed by a processor, such as a general-purpose computer, based on instructions stored on a computer-readable medium. If desired, the instructions may be stored on the medium in compressed and / or encrypted format.

[0059] Additionally, other implementations of the present application will be apparent to those skilled in the art upon consideration of this specification and practice of the present techniques. Various aspects and / or components of the described implementations may be used alone or in any combination. It is intended that the specification and implementations be considered as examples only, with the true scope and spirit of the present application being indicated by the following claims.

Claims

1. 1. A method for generating person-specific insights, comprising: receiving sensor data associated with the device; cleaning the received sensor data to generate cleansed data; processing the cleansed data using a machine learning model to generate machine learning metrics; Incorporating the machine learning metrics to generate insight data related to the device; generating people data using the insight data and mapping the insights to the people data; generating custom insights using the insight data and the person data, the custom insights being text-based summaries; and Distributing each of the custom insights to each person in the person data for placing a service order related to the device. A method comprising:

2. The method of claim 1 , wherein the custom insights are delivered via at least one of a dashboard or an application programming interface (API).

3. The method of claim 1 , wherein the machine learning metric comprises at least one of a failure probability, a probability density vector, an anomaly score, or a local outlier factor.

4. The machine learning model: retrieving historical sensor data forming a database; cleaning the past sensor data from defects; generating training features based on the cleaned historical sensor data; generating the machine learning model based on the training features and training the machine learning model using the historical sensor data; and Deploying the trained machine learning model. The method of claim 1 , wherein the eigenvalue is derived by:

5. The capturing of the machine learning metrics to generate the insight data related to the device includes: receiving input data to an insight model, the input data including the sensor data, the machine learning metrics, and potential cause data; and generating said insight data as output from said insight model; The method of claim 1 , comprising:

6. The method of claim 5 , wherein the insight model comprises one of a machine learning model or a rule-based algorithm.

7. Using the insight data to generate people data includes: Processing the insight data on a skills mapping model to generate required operator skill sets; and processing the insight data and the required operator skill sets on a person mapping model to generate the person data; Including, The method of claim 1 , wherein the person data is a list of people who possess the required operator skill set.

8. The method of claim 7 , wherein the person data is a list of people who possess the required operator skill set.

9. Generating the custom insights using the insight data and the person data includes: receiving the insight data and the person data as inputs to a transformer selector for performing language model selection from a plurality of language models by matching each person in the person data with at least one language model of the plurality of language models; and The method of claim 1 , comprising generating at least one custom insight derived from each Transformer model matched to each person in the person data.

10. The plurality of language models: using past insight data and a person vocabulary as inputs for training the plurality of language models, the person vocabulary including technical terms and definitions uniquely associated with people from different fields of expertise, and training the plurality of language models by identifying correlations between the past insight data and the person vocabulary; and deploying the trained language models. The method of claim 9 , wherein the value is derived by:

11. 1. A non-transitory computer-readable medium storing instructions for generating person-specific insights, the instructions comprising: receiving sensor data associated with the device; cleaning the received sensor data to generate cleansed data; processing the cleansed data using a machine learning model to generate machine learning metrics; Incorporating the machine learning metrics to generate insight data related to the device; generating people data using the insight data and mapping the insights to the people data; generating custom insights using the insight data and the person data, the custom insights being text-based summaries; and Distributing each of the custom insights to each person in the person data for placing a service order related to the device.

1. A non-transitory computer-readable medium comprising:

12. The non-transitory computer-readable medium of claim 11 , wherein the custom insights are delivered via at least one of a dashboard or an application programming interface (API).

13. The non-transitory computer-readable medium of claim 11 , wherein the machine learning metric comprises at least one of a failure probability, a probability density vector, an anomaly score, or a local outlier factor.

14. The machine learning model: retrieving historical sensor data forming a database; cleaning the past sensor data from defects; generating training features based on the cleaned historical sensor data; generating the machine learning model based on the training features and training the machine learning model using the historical sensor data; and Deploying the trained machine learning model.

12. The non-transitory computer-readable medium of claim 11, derived by:

15. The capturing of the machine learning metrics to generate the insight data related to the device includes: receiving input data to an insight model, the input data including the sensor data, the machine learning metrics, and potential cause data; and generating said insight data as output from said insight model; 12. The non-transitory computer-readable medium of claim 11, comprising:

16. The non-transitory computer-readable medium of claim 15 , wherein the insight model comprises one of a machine learning model or a rule-based algorithm.

17. Using the insight data to generate people data includes: Processing the insight data on a skills mapping model to generate required operator skill sets; and processing the insight data and the required operator skill sets on a person mapping model to generate the person data; Including, The non-transitory computer-readable medium of claim 11 , wherein the person data is a list of people who possess the required operator skill set.

18. 20. The non-transitory computer-readable medium of claim 17, wherein the person data is a list of people who possess the required operator skill set.

19. Generating the custom insights using the insight data and the person data includes: receiving the insight data and the person data as inputs to a transformer selector for performing language model selection from a plurality of language models by matching each person in the person data with at least one language model of the plurality of language models; and generating at least one custom insight derived from each Transformer model matched to each person in the person data; 12. The non-transitory computer-readable medium of claim 11, comprising:

20. The plurality of language models: using past insight data and a person vocabulary as inputs for training the plurality of language models, the person vocabulary including technical terms and definitions uniquely associated with people from different fields of expertise, and training the plurality of language models by identifying correlations between the past insight data and the person vocabulary; and deploying the trained language models.

20. The non-transitory computer-readable medium of claim 19, derived by:

Images