Lifestyle-informed, personalized blood test coverage and risk assessment

Abstract

Systems, apparatus, computer-implemented methods, and non-transitory computer-readable media are provided for facilitating health-based predictions for a patient using a single time-point biological sample of the patient. A regularization component can be integrated into a triple-based objective function associated with a triple-based machine learning model to create a modified triple-based objective function. The regularization component can operate to regularize the distances between pairs of positive and negative data values ​​in a triple-tuple relative to an anchor value in a triple-tuple of a data set. The triple-based machine learning model can be trained using the data set and the modified triple-based objective function to create a trained triple-based machine learning model. The trained triple-based machine learning model can be implemented to output one or more health-based predictions for a patient based at least in part on a single time-point biological sample of the patient.

Description

Priority claim

[0001] This application claims priority to U.S. Provisional Patent Application No. 63 / 406,624, filed on September 14, 2022. Applicants claim priority to and the benefit of said application, and said application is incorporated herein by reference in its entirety. Technical Field

[0002] The present disclosure generally relates to health risk modeling. More specifically, the present disclosure relates to health risk modeling using a single time point biological sample of a patient. Background Art

[0003] Predicting a patient's potential health risk remains a major area of ​​concern and challenge in healthcare. Standardized laboratory tests are widely used for health assessment, differential diagnosis, and treatment of patients. The problem with interpreting these data is the lack of quantitative and personalized metrics. Most current standardized reference values ​​are established based on percentiles for different age and gender groups. However, in most tests, age and gender alone can explain less than 10% of the normal within-test variance. In addition, personalized models based on patient medical history can explain 60% of the variance in half of the tests. These difficulties are further exacerbated if the task is to assess future health risks using only a single time point of blood measurement. In fact, not only do most risk prediction models rely on multiple measurements, but the definition of health risk is also based on multiple time points. Summary of the invention

[0004] Aspects and advantages of embodiments of the present disclosure will be set forth in part in the following description, or may be learned from the description, the attached claims, and / or the accompanying drawings, or may be learned through practice of the embodiments.

[0005] According to an example embodiment, a computing system may include one or more processors. The computing system may also include one or more non-transitory computer-readable media that store instructions together, which, when executed by one or more processors, cause the computing system to perform operations. These operations may include: training a triple-based machine learning model using a data set and a modified triple-based objective function to create a trained triple-based machine learning model. The data set may include multiple data values ​​that can be graphically represented in a multidimensional space. For example, the data set may be a healthcare data set with at least one of a genetic feature, a phenotypic feature, or a lifestyle feature. The operations may further include: receiving a single time point biological sample of a patient, and using a trained triple-based machine learning model and a single time point biological sample of the patient to generate one or more health-based predictions for the patient. The modified triple-based objective function may include a regularization component that can operate to regularize the distance between pairs of positive and negative data values ​​in a triple tuple relative to an anchor value in a triple tuple of the data set.

[0006] According to another example embodiment, a computer-implemented method may include: integrating a regularization component into a triple-based objective function associated with a triple-based machine learning model by a computing system that may include one or more computing devices to create a modified triple-based objective function. The regularization component may operate to regularize the distance between pairs of positive and negative data values ​​in a triple tuple relative to an anchor value in a triple tuple of a data set. The data set may include a plurality of data values ​​that may be graphically represented in a multidimensional space. For example, the data set may be a healthcare data set having at least one of a genetic feature, a phenotypic feature, or a lifestyle feature. The computer-implemented method may further include: training the triple-based machine learning model by the computing system using the data set and the modified triple-based objective function to create a trained triple-based machine learning model. The computer-implemented method may further include: generating one or more health-based predictions for a patient by the computing system using the trained triple-based machine learning model and a single time point biological sample of the patient.

[0007] According to another example embodiment, a computing device may include one or more processors. The computing device may further include one or more non-transitory computer-readable media that collectively store instructions that, when executed by one or more processors, cause the computing device to perform operations. These operations may include: integrating a regularization component into a triple-based objective function associated with a triple-based machine learning model to create a modified triple-based objective function. The regularization component may operate to regularize the distance between pairs of positive and negative data values ​​in a triple-tuple tuple relative to an anchor value in a triple-tuple tuple of a data set. The data set may include multiple data values ​​that can be graphically represented in a multidimensional space. For example, the data set may be a healthcare data set having at least one of a genetic feature, a phenotypic feature, or a lifestyle feature. These operations may further include: training a triple-based machine learning model using the data set and the modified triple-based objective function to create a trained triple-based machine learning model that generates one or more health-based predictions for a patient based on a single time point biological sample of the patient when implemented.

[0008] Other aspects of the disclosure relate to various systems, apparatuses, non-transitory computer-readable media, user interfaces, and electronic devices.

[0009] These and other features, aspects and advantages of various embodiments of the present disclosure will be better understood with reference to the following description and appended claims.The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate example embodiments of the present disclosure and, together with the description, serve to explain the relevant principles. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] A detailed discussion of the embodiments for those skilled in the art is set forth in this specification with reference to the accompanying drawings, in which:

[0011] Figure 1 Depicted is a block diagram of an example non-limiting computing system that can perform health-based predictions for a patient using a single point-in-time biological sample of the patient in accordance with an example embodiment of the present disclosure.

[0012] Figure 2 Depicted is a block diagram of an example non-limiting computing device that can perform health-based predictions for a patient using a single point-in-time biological sample of the patient, according to an example embodiment of the present disclosure.

[0013] Figure 3 Depicted is a block diagram of an example non-limiting computing device that can perform health-based predictions for a patient using a single point-in-time biological sample of the patient, according to an example embodiment of the present disclosure.

[0014] Figure 4 Depicted are example non-limiting schematic diagrams of example embodiments according to the present disclosure.

[0015] Figure 5 Depicted are example non-limiting diagrams of training and testing data sets according to example embodiments of the present disclosure.

[0016] Figure 6 Depicted is a flow chart of an example non-limiting computer-implemented method of performing a health-based prediction for a patient using a single time-point biological sample of the patient according to an example embodiment of the present disclosure.

[0017] Figure 7 Depicted is a flow chart of an example non-limiting computer-implemented method of performing a health-based prediction for a patient using a single time-point biological sample of the patient according to an example embodiment of the present disclosure.

[0018] The repeated use of reference numerals and / or numbers in this specification and / or drawings is intended to represent the same or similar features, elements or operations of the present disclosure. For the sake of brevity, repeated descriptions of repeated reference numerals and / or numbers in this specification are omitted. DETAILED DESCRIPTION Overview

[0019] Reference will now be made in detail to embodiments of the present disclosure, one or more examples of which are shown in the accompanying drawings. Each example is provided by way of explanation of the present disclosure rather than by way of limitation. In fact, it will be apparent to those skilled in the art that various modifications and variations may be made in the present disclosure without departing from the scope or spirit of the present disclosure. For example, a feature shown or described as part of one embodiment may be used together with another embodiment to produce yet another embodiment. Therefore, the present disclosure is intended to encompass modifications and variations as within the scope of the appended claims and their equivalents.

[0020] In general, the present disclosure relates to performing health-based predictions for a patient using a single time-point biological sample of the patient. In order to perform such health-based predictions for a patient using a single time-point biological sample of the patient, a triple-based objective function can be modified according to the example embodiments described herein so that when compared to an existing triple-based objective function, the modified triple-based objective function can provide further separation of embeddings in different classes of a data set (e.g., a relatively complex healthcare data set), and thus provide further separation of different classes. In addition, the triple-based objective function can be modified according to the example embodiments described herein so that it can provide a greater density of embeddings in each of the different classes of the data set. As described according to the example embodiments of the present disclosure, by providing such increased separation between embeddings in different classes of the data set, and providing such increased density of embeddings in each of the different classes, the modified triple-based objective function described herein can thereby provide training of a triple-based machine learning model on a relatively complex data set (e.g., a healthcare data set that can include at least one of a genetic feature, a phenotypic feature, or a lifestyle feature), while allowing such a trained model to perform accurate classification using limited data (e.g., a single time-point biological sample of a patient).

[0021] In an example embodiment, a computing system and / or computing device according to an example embodiment of the present disclosure may integrate a regularization component into a triple-based objective function associated with a triple-based machine learning model to create a modified triple-based objective function. In this embodiment and / or another embodiment, the regularization component may operate (e.g., when implemented using a computing system and / or computing device) to regularize the distance between pairs of positive and negative data values ​​in a triple tuple relative to an anchor value in a triple tuple of a data set. For example, in this embodiment and / or another embodiment, the regularization component may operate (e.g., when implemented using a computing system and / or computing device) to regularize the distance between pairs of positive and negative data values ​​in a triple tuple relative to an anchor value in a triple tuple of a data set, so that a first distance between positive and negative data values ​​in a triple tuple of the data set is equal to or greater than a second distance between negative data values ​​in the triple tuple and the anchor value.

[0022] In another example embodiment, a computing system and / or computing device may train a triple-based machine learning model using a data set and a modified triple-based objective function to create a trained triple-based machine learning model. In this embodiment and / or another embodiment, a computing system and / or computing device may use a trained triple-based machine learning model and a single time point biological sample of a patient to generate one or more health-based predictions for a patient. In one embodiment, one or more health-based predictions may constitute and / or include a classification of a patient in, for example, one or more predefined health-based classes (e.g., a healthy lifestyle class, an unhealthy lifestyle class, a defined blood range class, a defined health risk class); a health risk score prediction (e.g., where a score may indicate the likelihood that a patient will be infected or suffer from a defined disease at some time in the future); a diagnostic prediction; a prognostic prediction; and / or another health-based prediction.

[0023] In at least one embodiment, the data set may include a plurality of data values ​​that may be graphically represented in a multidimensional space. In one embodiment, the data set may constitute and / or include a healthcare data set that may include genetic characteristics, phenotypic characteristics, and / or lifestyle characteristics.

[0024] In at least one embodiment, the multidimensional space can constitute and / or include Euclidean space. In this embodiment and / or another embodiment, each distance in the first distance and the second distance, and the distance between the paired positive data values ​​and the negative data values ​​in the triplet tuple can constitute and / or include Euclidean distance.

[0025] In at least one embodiment, the triple-based machine learning model may constitute and / or include a triple-based deep metric learning model. In this embodiment and / or another embodiment, the trained triple-based machine learning model may constitute and / or include a trained triple-based deep metric learning model.

[0026] In one embodiment, the modified triple-based objective function described above may be operable (e.g., when implemented using a computing system and / or computing device) to learn similarity-based embeddings of multiple data values ​​in a dataset during training of a triple-based machine learning model using the dataset and the modified triple-based objective function.

[0027] In at least one embodiment, the modified triple-based objective function described above may be operable (e.g., when implemented using a computing system and / or computing device) to provide improved density of embeddings of data values ​​for each data value class in a data set to provide improved classification results across different classifier models that implement the modified triple-based objective function, as compared to classifier models that implement different objective functions. In one embodiment, the modified triple-based objective function may be operable (e.g., when implemented using a computing system and / or computing device) to provide improved separability between different data value classes in a data set to provide improved classification results across different classifier models that implement the modified triple-based objective function, as compared to classifier models that implement different objective functions. In this embodiment and / or another embodiment, the modified triple-based objective function can be operated (e.g., when implemented using a computing system and / or computing device) to provide improved separability between a first embedding of a first data value of a first data value class in a data set and a second embedding of a second data value of a second data value class in the data set to provide improved classification results across different classifier models that implement the modified triple-based objective function compared to a classifier model that implements another objective function.

[0028] According to example embodiments, a computing system (e.g., described below and in Figure 1 The computing system 100, server computing system 130, and / or automated machine learning computing system 150 shown in the figure may include one or more processors. The computing system may also include one or more non-transitory computer-readable media that collectively store instructions that, when executed by one or more processors, cause the computing system to perform operations. These operations may include: training a triple-based machine learning model using a data set and a modified triple-based objective function to create a trained triple-based machine learning model. The data set may include multiple data values ​​that can be graphically represented in a multidimensional space. In one embodiment, the data set may constitute and / or include a healthcare data set that may include genetic features, phenotypic features, and / or lifestyle features. The operations may further include: receiving a single time point biological sample of a patient, and using a trained triple-based machine learning model and a single time point biological sample of the patient to generate one or more health-based predictions for the patient. The modified triple-based objective function may include a regularization component that may operate to regularize the distance between pairs of positive and negative data values ​​in a triple tuple relative to an anchor value in a triple tuple of the data set.

[0029] According to another example embodiment, a computer-implemented method (e.g., described below and in Figure 7The computer-implemented method 700 shown in may include: integrating a regularization component into a triple-based objective function associated with a triple-based machine learning model by a computing system that may include one or more computing devices to create a modified triple-based objective function. The regularization component may operate to regularize the distance between pairs of positive and negative data values ​​in a triple tuple relative to an anchor value in a triple tuple of a data set. The data set may include multiple data values ​​that can be graphically represented in a multidimensional space. In one embodiment, the data set may constitute and / or include a healthcare data set that may include genetic features, phenotypic features, and / or lifestyle features. The computer-implemented method may further include: training a triple-based machine learning model using the data set and the modified triple-based objective function by the computing system to create a trained triple-based machine learning model. The computer-implemented method may further include: receiving a single time point biological sample of a patient, and using the trained triple-based machine learning model and the single time point biological sample of the patient to generate one or more health-based predictions for the patient.

[0030] According to another example embodiment, a computing device (e.g., described below and respectively in Figure 1 , Figure 2 and Figure 3 The computing device 102, computing device 200, and / or computing device 300 shown in the figure may include one or more processors. The computing device may further include one or more non-transitory computer-readable media that collectively store instructions that, when executed by one or more processors, cause the computing device to perform operations. These operations may include: integrating a regularization component into a triple-based objective function associated with a triple-based machine learning model to create a modified triple-based objective function. The regularization component may operate to regularize the distance between pairs of positive and negative data values ​​in a triple tuple relative to an anchor value in a triple tuple of the data set. The data set may include multiple data values ​​that can be graphically represented in a multidimensional space. In one embodiment, the data set may constitute and / or include a healthcare data set that may include genetic features, phenotypic features, and / or lifestyle features. The operations may further include training a triple-based machine learning model using the dataset and the modified triple-based objective function to create a trained triple-based machine learning model that, when implemented, generates one or more health-based predictions for the patient based on a single time-point biological sample of the patient.

[0031] The computing systems, computing devices, and / or computer-implemented methods of the present disclosure provide many technical effects and benefits. For example, example embodiments of the present disclosure relate to using a single time point biological sample of a patient to perform health-based predictions for a patient. In order to perform such health-based predictions for a patient using a single time point biological sample of a patient, the triple-based objective function can be modified according to the example embodiments described herein, such that when compared to an existing triple-based objective function, the modified version can provide further separation of embeddings in different classes of a data set (e.g., a relatively complex healthcare data set), and thus provide further separation of different classes. In addition, the triple-based objective function can be modified according to the example embodiments described herein, such that when compared to an existing triple-based objective function, the modified version can provide a greater density of embeddings in each class of different classes of a data set.

[0032] In at least one embodiment, by providing such improved separation of embeddings of different classes and such improved density of embeddings of each class in different classes of a data set, a modified version of a triple-based objective function according to the example embodiments described herein can thereby provide improved classification results when compared to existing triple-based objective functions. Specifically, a modified version of a triple-based objective function according to the example embodiments described herein can be used to train a triple-based machine learning model, which can then be implemented to provide accurate health-based predictions for a patient using a single time-point biological sample of the patient, whereas existing triple-based machine learning models require multiple samples to perform the same prediction.

[0033] In at least one embodiment, such modified versions of triple-based objective functions according to the example embodiments described herein can thereby improve processing performance (e.g., accuracy, efficiency), and reduce processing workload and / or computational cost of a processor executing a triple-based machine learning model that has been trained using the modified version of the triple-based objective function, by providing improved and / or accurate classification results while using less input data (e.g., a single time-point biological sample of a patient). In at least one other embodiment, such modified versions of triple-based objective functions according to the example embodiments described herein can thereby improve storage capacity of a memory device storing electronic health records of various patients that can be used by a triple-based machine learning model that has been trained using the modified version of the triple-based objective function, by providing improved and / or accurate classification results while using less input data (e.g., a single time-point biological sample of a patient). Example Apparatus and Systems

[0034] Figure 1 A block diagram of an example non-limiting computing system 100 that can perform health-based predictions for a patient using a single time-point biological sample of the patient according to an example embodiment of the present disclosure is depicted. The system 100 includes a user computing device 102, a server computing system 130, and a training computing system 150 communicatively coupled via a network 180.

[0035] The user computing device 102 can be any type of computing device, such as, for example, a personal computing device (e.g., a laptop or desktop computer), a mobile computing device (e.g., a smart phone or tablet computer), a game console or controller, a wearable computing device, an embedded computing device, or any other type of computing device.

[0036] The user computing device 102 includes one or more processors 112 and a memory 114. The one or more processors 112 may be any suitable processing device (e.g., a processor core, a microprocessor, an ASIC, an FPGA, a controller, a microcontroller, etc.) and may be one processor or a plurality of processors operatively connected. The memory 114 may include one or more non-transitory computer-readable storage media, such as RAM, ROM, EEPROM, EPROM, flash memory devices, magnetic disks, etc., and combinations thereof. The memory 114 may store data 116 and instructions 118 that are executed by the processor 112 to cause the user computing device 102 to perform operations.

[0037] In some implementations, the user computing device 102 may store or include one or more triple-based machine learning models 120 (e.g., trained triple-based machine learning models and / or untrained triple-based machine learning models 120). For example, the triple-based machine learning model 120 may constitute and / or include one or more untrained triple-based machine learning models and / or one or more trained triple-based machine learning models that have been trained using a modified triple-based objective function according to one or more embodiments described herein. In one example, the triple-based machine learning model 120 may be or may otherwise include various machine learning models, such as a neural network (e.g., a deep neural network) or other types of machine learning models, including nonlinear models and / or linear models. The neural network may include a feedforward neural network, a recurrent neural network (e.g., a long short-term memory recurrent neural network), a convolutional neural network, or other forms of neural networks. Some example machine learning models may utilize attention mechanisms, such as self-attention. For example, some example machine learning models may include a multi-head self-attention model (e.g., a Transformer model).

[0038] In some implementations, one or more triple-based machine learning models 120 may be received from the server computing system 130 via the network 180, stored in the user computing device memory 114, and then used or otherwise implemented by the one or more processors 112. In some implementations, the user computing device 102 may implement multiple parallel instances of a single triple-based machine learning model 120 (e.g., to perform parallel health-based predictions for different patients across multiple instances of a single time-point biological sample, each corresponding to a different patient).

[0039] More specifically, the triple-based machine learning model 120 can be implemented according to the example embodiments described herein to perform health-based predictions for a patient using a single time-point biological sample of the patient. In order to perform such health-based predictions for a patient using a single time-point biological sample of the patient, the triple-based objective function can be modified according to the example embodiments described herein so that when compared to the existing triple-based objective function, the modified triple-based objective function can provide further separation of embeddings in different classes of a dataset (e.g., a relatively complex healthcare dataset), and therefore provide further separation of different classes. In addition, the triple-based objective function can be modified according to the example embodiments described herein so that it can provide a greater density of embeddings in each of the different classes of the dataset. As described in accordance with example embodiments of the present disclosure, by providing such increased separation between embeddings in different classes of a dataset, and providing such increased density of embeddings in each of the different classes, the modified triple-based objective function described herein can thereby provide for training a triple-based machine learning model 120 on a relatively complex dataset (e.g., a healthcare dataset) while allowing such trained triple-based machine learning model 120 to perform accurate classification using limited data (e.g., a single time-point biological sample of a patient).

[0040] Additionally or alternatively, one or more triple-based machine learning models 140 (e.g., triple-based machine learning models trained using the modified triple-based objective function of the present disclosure) may be included in or otherwise stored and implemented by a server computing system 130 that communicates with the user computing device 102 according to a client-server relationship. For example, the triple-based machine learning model 140 may be implemented by the server computing system 140 as part of a web service (e.g., a classification service for patients and / or a health-based prediction service). Thus, one or more triple-based machine learning models 120 may be stored and implemented at the user computing device 102, and / or one or more models 140 may be stored and implemented at the server computing system 130.

[0041] The user computing device 102 may also include one or more user input components 122 that receive user input. For example, the user input component 122 may be a touch-sensitive component (e.g., a touch-sensitive display screen or a touchpad) that is sensitive to the touch of a user input object (e.g., a finger or a stylus). The touch-sensitive component may be used to implement a virtual keyboard. Other example user input components include a microphone, a traditional keyboard, or other devices by which a user can provide user input.

[0042] The server computing system 130 includes one or more processors 132 and memory 134. The one or more processors 132 may be any suitable processing device (e.g., a processor core, a microprocessor, an ASIC, an FPGA, a controller, a microcontroller, etc.) and may be one processor or a plurality of processors operatively connected. The memory 134 may include one or more non-transitory computer-readable storage media, such as RAM, ROM, EEPROM, EPROM, flash memory devices, magnetic disks, etc., and combinations thereof. The memory 134 may store data 136 and instructions 138 that are executed by the processor 132 to cause the server computing system 130 to perform operations.

[0043] In some implementations, the server computing system 130 includes or is otherwise implemented by one or more server computing devices. Where the server computing system 130 includes multiple server computing devices, such server computing devices may operate according to a sequential computing architecture, a parallel computing architecture, or some combination thereof.

[0044] As described above, the server computing system 130 may store or otherwise include one or more triple-based machine learning models 140. For example, the triple-based machine learning model 140 may be or may otherwise include various machine learning models. Example machine learning models include neural networks or other multi-layer nonlinear models. Example neural networks include feedforward neural networks, deep neural networks, recurrent neural networks, and convolutional neural networks. Some example machine learning models may utilize attention mechanisms, such as self-attention. For example, some example machine learning models may include a multi-head self-attention model (e.g., a Transformer model).

[0045] User computing device 102 and / or server computing system 130 may train models 120 and / or 140 via interaction with training computing system 150 communicatively coupled via network 180. Training computing system 150 may be separate from server computing system 130 or may be part of server computing system 130.

[0046] The training computing system 150 includes one or more processors 152 and memory 154. The one or more processors 152 may be any suitable processing device (e.g., a processor core, a microprocessor, an ASIC, an FPGA, a controller, a microcontroller, etc.) and may be one processor or a plurality of processors operatively connected. The memory 154 may include one or more non-transitory computer-readable storage media, such as RAM, ROM, EEPROM, EPROM, flash memory devices, disks, etc., and combinations thereof. The memory 154 may store data 156 and instructions 158, which are executed by the processor 152 to cause the training computing system 150 to perform operations. In some implementations, the training computing system 150 includes or is otherwise implemented by one or more server computing devices.

[0047] The training computing system 150 may include a model trainer 160 that trains the machine learning models 120 and / or 140 stored at the user computing device 102 and / or the server computing system 130 using various training or learning techniques such as, for example, error back propagation. For example, a loss function may be back propagated through the model to update one or more parameters of the model (e.g., based on the gradient of the loss function). Various loss functions may be used, such as mean squared error, likelihood loss, cross entropy loss, hinge loss, and / or various other loss functions. Gradient descent techniques may be used to iteratively update the parameters in multiple training iterations.

[0048] In some implementations, performing error back-propagation may include performing truncated back-propagation through time.The model trainer 160 may perform a variety of generalization techniques (eg, weight decay, dropout, etc.) to improve the generalization capabilities of the model being trained.

[0049] Specifically, the model trainer 160 can train the triple-based machine learning models 120 and / or 140 based on a set of training data 162 and / or using a modified version of a triple-based objective function that has been modified according to one or more embodiments described herein. The training data 162 can include, for example, a healthcare dataset having genetic features, phenotypic features, lifestyle features, and / or other features that can describe and / or indicate attributes and / or lifestyles of different patients.

[0050] In some implementations, where the user has provided consent, the training examples may be provided by the user computing device 102. Thus, in such implementations, the model 120 provided to the user computing device 102 may be trained by the training computing system 150 based on user-specific data received from the user computing device 102. In some cases, this process may be referred to as personalizing the model.

[0051] The model trainer 160 includes computer logic for providing the desired functionality. The model trainer 160 may be implemented in hardware, firmware, and / or software that controls a general purpose processor. For example, in some implementations, the model trainer 160 includes a program file stored on a storage device, loaded into a memory, and executed by one or more processors. In other implementations, the model trainer 160 includes one or more sets of computer executable instructions stored in a tangible computer readable storage medium such as RAM, a hard disk, or an optical or magnetic medium.

[0052] The network 180 may be any type of communication network, such as a local area network (e.g., an intranet), a wide area network (e.g., the Internet), or some combination thereof, and may include any number of wired or wireless links. In general, communications over the network 180 may be conducted over any type of wired and / or wireless connection using a variety of communication protocols (e.g., TCP / IP, HTTP, SMTP, FTP), encodings or formats (e.g., HTML, XML), and / or protection schemes (e.g., VPN, secure HTTP, SSL).

[0053] The machine learning models described in this specification can be used for a variety of tasks, applications, and / or use cases.

[0054] In some implementations, the input to the machine learning model of the present disclosure may be text or natural language data. The machine learning model may process the text or natural language data to generate an output. As an example, the machine learning model may process the natural language data to generate a language encoding output. As another example, the machine learning model may process the text or natural language data to generate a potential text embedding output. As another example, the machine learning model may process the text or natural language data to generate a translation output. As another example, the machine learning model may process the text or natural language data to generate a classification output. As another example, the machine learning model may process the text or natural language data to generate a text segmentation output. As another example, the machine learning model may process the text or natural language data to generate a semantic intent output. As another example, the machine learning model may process the text or natural language data to generate an amplified text or natural language output (e.g., text or natural language data of higher quality than the input text or natural language, etc.). As another example, the machine learning model may process the text or natural language data to generate a prediction output.

[0055] In some implementations, the input to the machine learning model of the present disclosure may be image data. The machine learning model may process the image data to generate an output. As an example, the machine learning model may process the image data to generate an image recognition output (e.g., recognition of image data, potential embedding of image data, encoded representation of image data, hash of image data, etc.). As another example, the machine learning model may process the image data to generate an image segmentation output. As another example, the machine learning model may process the image data to generate an image classification output. As another example, the machine learning model may process the image data to generate an image data modification output (e.g., a change in image data, etc.). As another example, the machine learning model may process the image data to generate an encoded image data output (e.g., an encoded representation and / or compressed representation of image data, etc.). As another example, the machine learning model may process the image data to generate an amplified image data output. As another example, the machine learning model may process the image data to generate a prediction output.

[0056] In some implementations, the input to the machine learning model of the present disclosure may be speech data. The machine learning model may process speech data to generate an output. As an example, the machine learning model may process speech data to generate a speech recognition output. As another example, the machine learning model may process speech data to generate a speech conversion output. As another example, the machine learning model may process speech data to generate a potential embedding output. As another example, the machine learning model may process speech data to generate an encoded speech output (e.g., an encoded representation and / or compressed representation of speech data, etc.). As another example, the machine learning model may process speech data to generate an amplified speech output (e.g., speech data of higher quality than the input speech data, etc.). As another example, the machine learning model may process speech data to generate a text representation output (e.g., a text representation of the input speech data, etc.). As another example, the machine learning model may process speech data to generate a predicted output.

[0057] In some implementations, the input to the machine learning model of the present disclosure may be latent coded data (e.g., a latent space representation of an input, etc.). The machine learning model may process the latent coded data to generate an output. As an example, the machine learning model may process the latent coded data to generate a recognition output. As another example, the machine learning model may process the latent coded data to generate a reconstruction output. As another example, the machine learning model may process the latent coded data to generate a search output. As another example, the machine learning model may process the latent coded data to generate a re-clustering output. As another example, the machine learning model may process the latent coded data to generate a prediction output.

[0058] In some implementations, the input to the machine learning model of the present disclosure may be statistical data. Statistical data may be, represent, or otherwise include data calculated and / or computed from some other data source. The machine learning model may process statistical data to generate an output. As an example, the machine learning model may process statistical data to generate an identification output. As another example, the machine learning model may process statistical data to generate a prediction output. As another example, the machine learning model may process statistical data to generate a classification output. As another example, the machine learning model may process statistical data to generate a segmentation output. As another example, the machine learning model may process statistical data to generate a visualization output. As another example, the machine learning model may process statistical data to generate a diagnostic output.

[0059] In some implementations, the input to the machine learning model of the present disclosure may be sensor data. The machine learning model may process the sensor data to generate an output. As an example, the machine learning model may process the sensor data to generate a recognition output. As another example, the machine learning model may process the sensor data to generate a prediction output. As another example, the machine learning model may process the sensor data to generate a classification output. As another example, the machine learning model may process the sensor data to generate a segmentation output. As another example, the machine learning model may process the sensor data to generate a visualization output. As another example, the machine learning model may process the sensor data to generate a diagnostic output. As another example, the machine learning model may process the sensor data to generate a detection output.

[0060] In some cases, the machine learning model can be configured to perform a task that includes encoding input data for reliable and / or efficient transmission or storage (and / or corresponding decoding). For example, the task can be an audio compression task. The input can include audio data, and the output can include compressed audio data. In another example, the input includes visual data (e.g., one or more images or videos), the output includes compressed visual data, and the task is a visual data compression task. In another example, the task can include generating an embedding for the input data (e.g., input audio or visual data).

[0061] In some cases, the input includes visual data, and the task is a computer vision task. In some cases, the input includes pixel data for one or more images, and the task is an image processing task. For example, the image processing task may be image classification, where the output is a set of scores, each score corresponding to a different object class and representing the possibility that the one or more images depict an object belonging to an object class. The image processing task may be object detection, where the image processing output identifies one or more regions in the one or more images, and for each region, identifies the possibility that the region depicts an object of interest. As another example, the image processing task may be image segmentation, where the image processing output defines the corresponding possibility of each category in a set of predetermined categories for each pixel in the one or more images. For example, the group of categories may be foreground and background. As another example, the group of categories may be object classes. As another example, the image processing task may be depth estimation, where the image processing output defines a corresponding depth value for each pixel in the one or more images. As another example, the image processing task may be motion estimation, where the network input includes multiple images, and the image processing output defines the motion of the scene depicted at the pixel between images in the network input for each pixel of one of the input images.

[0062] In some cases, the input includes audio data representing a spoken utterance, and the task is a speech recognition task. The output may include a text output mapped to the spoken utterance. In some cases, the task includes encrypting or decrypting the input data. In some cases, the task includes a microprocessor performance task, such as branch prediction or memory address translation.

[0063] Figure 1 An example computing system that can be used to implement the present disclosure is shown. Other computing systems may also be used. For example, in some implementations, the user computing device 102 may include a model trainer 160 and a training data set 162. In such implementations, the model 120 may be both trained and used locally at the user computing device 102. In some such implementations, the user computing device 102 may implement the model trainer 160 to personalize the model 120 based on user-specific data.

[0064] Figure 2 A block diagram depicts an example non-limiting computing device 200 that can perform health-based predictions for a patient using a single point-in-time biological sample of the patient according to an example embodiment of the present disclosure. The computing device 200 can be a user computing device or a server computing device.

[0065] The computing device 200 includes multiple applications (e.g., applications 1 to N). Each application contains its own machine learning library and machine learning model. For example, each application can include a machine learning model. Example applications include text messaging applications, email applications, dictation applications, virtual keyboard applications, browser applications, etc.

[0066] like Figure 2 As shown, each application can communicate with multiple other components of the computing device, such as, for example, one or more sensors, a context manager, a device state component, and / or additional components. In some implementations, each application can communicate with each device component using an API (e.g., a public API). In some implementations, the API used by each application is specific to the application.

[0067] Figure 3 A block diagram depicts an example non-limiting computing device 300 that can perform health-based predictions for a patient using a single point-in-time biological sample of the patient according to an example embodiment of the present disclosure. The computing device 300 can be a user computing device or a server computing device.

[0068] The computing device 300 includes a plurality of applications (e.g., applications 1 to N). Each application communicates with a central intelligence layer. Example applications include a text messaging application, an email application, a dictation application, a virtual keyboard application, a browser application, etc. In some implementations, each application can communicate with the central intelligence layer (and the models stored therein) using an API (e.g., a public API across all applications).

[0069] The central intelligence layer includes multiple machine learning models. For example, Figure 3 As shown, a corresponding machine learning model can be provided for each application, and the corresponding machine learning model is managed by the central intelligence layer. In other implementations, two or more applications can share a single machine learning model. For example, in some implementations, the central intelligence layer can provide a single model for all applications. In some implementations, the central intelligence layer is included in the operating system of the computing device 300 or is otherwise implemented by the operating system.

[0070] The central intelligence layer may communicate with a central device data layer. The central device data layer may be a centralized data repository for computing device 300. Figure 3 As shown, the central device data layer can communicate with multiple other components of the computing device, such as, for example, one or more sensors, a context manager, a device state component, and / or additional components. In some implementations, the central device data layer can communicate with each device component using an API (e.g., a private API).

[0071] Figure 4 An example non-limiting diagram 400 is depicted according to an example embodiment of the present disclosure. Figure 4 The diagram 400 shown in the example embodiment depicted in can provide an illustration of how the modified triple-based objective function of the present disclosure and / or the regularization component that may be integrated therein may operate (e.g., when implemented using the computing system and / or computing device described herein) to regularize the distances between pairs of positive and negative data values ​​in a triple tuple relative to an anchor value in the triple tuple of a data set. For example, in Figure 4 In the embodiment depicted in , the modified triple-based objective function and / or its regularization component can operate (e.g., when implemented using the computing system and / or computing device described herein) to regularize the distances between pairs of positive and negative data values ​​in the triple tuples of the data set relative to the anchor values ​​in the triple tuples of the data set, so that the first distance 402a between the positive data value 404 and the negative data value 406 in the triple tuples of the data set is equal to or greater than the second distance 402b between the negative data value 406 and the anchor value 408 in the triple tuples. Figure 4 As shown in the embodiments depicted in , the triple-based machine learning models 120 and / or 140 that have been trained using the modified triple-based objective function of the present disclosure can map positive pairs (e.g., positive data value 404, other positive data values, anchor value 408) within a marginal ball 410 associated with the anchor data value 408.

[0072] Figure 5 Depicted are example non-limiting diagrams of training datasets 502a, 502b, 502c and test datasets 504a, 504b, 504c according to example embodiments of the present disclosure. Compared to existing triple-based machine learning models, triple-based machine learning models 120 and / or 140 that have been trained using the modified triple-based objective function of the present disclosure can learn embeddings that are relatively more compact within a class of the dataset and further separated from other classes of the dataset, thereby producing better classification results, as shown by Figure 5 As shown in the training data sets 502a, 502b, 502c and the test data sets 504a, 504b, 504c described in FIG.

[0073] exist Figure 5 In the example shown, training data sets 502a and 502b correspond to implementations performed using existing triple-based machine learning models, while training data set 502c corresponds to implementations performed using triple-based machine learning models 120 and / or 140 according to example embodiments of the present disclosure. Figure 5In the example shown, test data sets 504a, 504b correspond to an implementation performed using an existing triple-based machine learning model, while test data set 504c corresponds to an implementation performed using a triple-based machine learning model 120 and / or 140 according to an example embodiment of the present disclosure. Example Method

[0074] Figure 6 A flowchart is shown of an example non-limiting computer-implemented method 600 for performing a health-based prediction for a patient using a single time-point biological sample of the patient according to one or more example embodiments of the present disclosure. Figure 1 , Figure 2 and Figure 3 The computer-implemented method 600 is implemented by the computing device 200, computing device 300, computing system 100, server computing system 130, and / or automated machine learning computing system 150 shown in FIG. For purposes of illustration and discussion, Figure 6 The example embodiments shown in the drawings depict operations performed in a particular order. One of ordinary skill in the art, using the disclosure provided herein, will appreciate that the various operations or steps of the computer-implemented method 600 or any of the other methods disclosed herein may be adjusted, modified, rearranged, performed simultaneously, include operations not shown, and / or changed in various ways without departing from the scope of the present disclosure.

[0075] At 602, the computing system can obtain (e.g., receive) patient laboratory visit sample data 602 (e.g., a single time point biological sample of a patient). At 604 and 606, the computing system can perform one or more computing and / or machine learning operations (e.g., a patient representation 604 and a single mathematical definition of health risk 606) in accordance with one or more embodiments to facilitate one or more clinical and / or consumer health operations, such as, for example: predicting a patient's health risk 608 using, for example, one or more triple-based machine learning models 120 and / or 140; and / or additional testing or intervention 610.

[0076] Figure 7 A flowchart is shown of an example non-limiting computer-implemented method 700 for performing a health-based prediction for a patient using a single time-point biological sample of the patient according to one or more example embodiments of the present disclosure. Figure 1 , Figure 2 and Figure 3The computer-implemented method 700 is implemented by the computing device 200, computing device 300, computing system 100, server computing system 130, and / or automated machine learning computing system 150 shown in FIG. For purposes of illustration and discussion, Figure 7 The example embodiments shown in the drawings depict operations performed in a particular order. One of ordinary skill in the art, using the disclosure provided herein, will appreciate that the various operations or steps of the computer-implemented method 700 or any of the other methods disclosed herein may be adjusted, modified, rearranged, performed simultaneously, include operations not shown, and / or changed in various ways without departing from the scope of the present disclosure.

[0077] At 702, the computer-implemented method 700 may include: integrating, by a computing system (e.g., computing system 100, server computing system 130, automated machine learning computing system 150) including one or more computing devices (e.g., computing device 200, computing device 300, computing device 102) a regularization component into a triple-based objective function associated with a triple-based machine learning model to create a modified triple-based objective function. In some embodiments, the regularization component may operate (e.g., when implemented by computing system 100) to regularize distances (e.g., distances 402a, 402b) between pairs of positive data values ​​(e.g., positive data value 404) and negative data values ​​(e.g., negative data value 406) in a triple tuple of a data set (e.g., a healthcare data set) relative to an anchor value (e.g., anchor value 408) in the triple tuple. In these or other embodiments, the data set may include a plurality of data values ​​graphically represented in a multidimensional space (e.g., Euclidean space). For example, a dataset may constitute and / or include a healthcare dataset, which may include genetic characteristics, phenotypic characteristics, and / or lifestyle characteristics.

[0078] At 704, the computer-implemented method 700 may include: training a triple-based machine learning model by a computing system (e.g., by computing system 100 and / or automated machine learning computing system 150 using model trainer 161) using a data set and a modified triple-based objective function to create a trained triple-based machine learning model (e.g., trained triple-based machine learning model 120 and / or 140).

[0079] At 706 , the computer-implemented method 700 may include generating, by the computing system, one or more health-based predictions for the patient using the trained triple-based machine learning model and the single time-point biological sample of the patient. Additional public content

[0080] As referenced herein, the term "entity" refers to a human, a user, an end-user, a consumer, a computing device and / or a program (e.g., a processor, computing hardware and / or software, an application, etc.), an agent, a machine learning (ML) and / or artificial intelligence (AI) algorithm, a model, a system and / or an application, and / or another type of entity that can implement one or more embodiments of the present disclosure as described herein, shown in the drawings and / or included in the appended claims and / or facilitate the implementation of the one or more embodiments. As referenced herein, the term "comprising" is intended to represent inclusion in a manner similar to the term "including". As referenced herein, the terms "or" and "and / or" are generally intended to be inclusive, that is, (i.e.), "A or B" or "A and / or B" are each intended to mean "A or B or both". As referenced herein, the terms "first", "second", "third", etc. are used interchangeably to distinguish one component or entity from another component or entity, and are not intended to represent the location, functionality or importance of a separate component or entity. As referred to herein, the term "coupled" refers to chemical coupling (e.g., chemical bonding), communicative coupling, electrical and / or electromagnetic coupling (e.g., capacitive coupling, inductive coupling, direct and / or connection coupling, etc.), mechanical coupling, operational coupling, optical coupling, and / or physical coupling.

[0081] Many details are set forth in the foregoing description. However, it will be apparent to those of ordinary skill in the art who benefit from the present disclosure that the present disclosure may be practiced without these specific details. In some cases, structures and devices are shown in block diagram form rather than in detail to avoid obscuring the present disclosure.

[0082] Some parts of the specific embodiments have been presented with respect to the process and symbolic representation of the operation of the data bits in the computer memory. Here, the process may include a self-consistent sequence of steps that produce a result. The steps may include those steps that require physical manipulation of physical quantities. These quantities may take the form of electrical or magnetic signals that can be stored, transmitted, combined, compared, and otherwise manipulated. These signals may be referred to as bits, values, symbols, characters, items, numbers, etc. These terms and similar terms may be associated with physical quantities and may represent labels applied to these quantities. Terms including "analyze", "access", "determine", "identify", "adjust", "modify", "send", "receive", "process", "generate", etc. may refer to the actions and processes of a computer system, a computing device, or a similar electronic computing device, which manipulate data represented as physical (e.g., electronic) quantities in the registers and memories of a computer system and transform the data into other data that may be similarly represented as physical quantities in the memory, registers, or other information storage devices, data transmission devices, or data processing devices of a computer system.

[0083] Certain examples of the present disclosure may relate to devices that can be implemented to perform the operations described herein. The device may include a computing device that can be activated or reconfigured by a computer program including electronic instructions stored in the computing device. Such a computer program may be stored in a computer-readable storage medium (e.g., a non-transitory computer-readable storage medium), which may include any type of storage. For example, storage may include a hard drive, a solid-state drive, a floppy disk, an optical disk, a read-only memory (ROM), a compact disk read-only memory (CD-ROM), a magneto-optical disk, a random access memory (RAM), a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), a magnetic card or an optical card, or any type of medium suitable for storing electronic instructions.

[0084] The technology discussed herein relates to servers, databases, software applications and other computer-based systems, as well as the actions performed by such systems and the information transmitted to and from such systems. The inherent flexibility of computer-based systems allows for various possible configurations, combinations and divisions of tasks and functionality between and among components. For example, a single device or component or multiple devices or components working in combination can be used to implement the processes discussed herein. Databases and applications can be distributed on a single system or across multiple systems. Distributed components can operate sequentially or in parallel.

[0085] Although the present subject matter has been described in detail with respect to various specific example embodiments of the present subject matter, each example is provided by way of explanation rather than limitation of the present disclosure. Those skilled in the art may easily produce changes, variations, and equivalents to such embodiments after understanding the foregoing. Therefore, the present disclosure does not exclude such modifications, variations, and / or additions to the present subject matter that would be readily apparent to those of ordinary skill in the art. For example, a feature shown or described as part of one embodiment may be used together with another embodiment to produce yet another embodiment. Therefore, the present disclosure is intended to encompass these changes, variations, and equivalents.

[0086] The above description is intended to be illustrative rather than limiting. Accordingly, reference should be made to the claims to determine the scope of the present disclosure.

Claims

1. A computing system, comprising: one or more processors; as well as one or more non-transitory computer-readable media that collectively store instructions that, when executed by the one or more processors, cause the computing system to perform operations comprising: training a triple-based machine learning model using a dataset and a modified triple-based objective function to create a trained triple-based machine learning model, the dataset comprising a healthcare dataset having at least one of a genetic characteristic, a phenotypic characteristic, or a lifestyle characteristic; Receiving a single time-point biospecimen from a patient; and generating one or more health-based predictions for the patient using the trained triplet-based machine learning model and the single time-point biological sample of the patient, Wherein the modified triple-based objective function includes a regularization component that operates to regularize distances between pairs of positive and negative data values ​​in a triplet tuple of the data set relative to an anchor value in the triplet tuple.

2. A computing system as described in claim 1, wherein the regularization component is capable of operating to regularize the distance between the pairs of positive data values ​​and negative data values ​​in the triplet tuple relative to the anchor value in the triplet tuple of the data set, so that a first distance between the positive data value and the negative data value in the triplet tuple is equal to or greater than a second distance between the negative data value in the triplet tuple and the anchor value.

3. The computing system of claim 1, wherein the triple-based machine learning model comprises a triple-based deep metric learning model, and the trained triple-based machine learning model comprises a trained triple-based deep metric learning model.

4. The computing system of claim 1, wherein the multidimensional space comprises a Euclidean space, and wherein each of the first distance, the second distance, and the distance between the pairs of positive and negative data values ​​in the triplet tuple comprises a Euclidean distance.

5. A computing system as described in claim 1, wherein the modified triple-based objective function is capable of operating to learn similarity-based embeddings of multiple data values ​​in the data set during training of the triple-based machine learning model using the data set and the modified triple-based objective function.

6. A computing system as described in claim 1, wherein the modified triple-based objective function is capable of operating to provide improved density of embeddings of data values ​​for each data value class in the data set to provide improved classification results across different classifier models that implement the modified triple-based objective function compared to classifier models that implement different objective functions.

7. A computing system as described in claim 1, wherein the modified triple-based objective function is capable of operating to provide improved separability between different data value classes in the data set to provide improved classification results across different classifier models that implement the modified triple-based objective function compared to classifier models that implement different objective functions.

8. A computing system as described in claim 1, wherein the modified triple-based objective function is capable of operating to provide improved separability between a first embedding of a first data value of a first data value class in the data set and a second embedding of a second data value of a second data value class in the data set to provide improved classification results across different classifier models that implement the modified triple-based objective function compared to a classifier model that implements another objective function.

9. The computing system of claim 1, the one or more health-based predictions comprising at least one of: classification of the patient in one or more predefined health-based classes, a health risk score prediction, a diagnosis prediction, or a prognosis prediction.

10. A computer-implemented method, the computer-implemented method comprising: integrating, by a computing system including one or more computing devices, a regularization component into a triple-based objective function associated with a triple-based machine learning model to create a modified triple-based objective function, the regularization component operable to regularize distances between pairs of positive and negative data values ​​in a triple-tuple of a dataset relative to an anchor value in the triple-tuple, the dataset comprising a healthcare dataset having at least one of a genetic characteristic, a phenotypic characteristic, or a lifestyle characteristic; training, by the computing system, a triple-based machine learning model using the data set and the modified triple-based objective function to create a trained triple-based machine learning model; Receive a single time-point biospecimen from a patient; as well as The trained triple-based machine learning model and the single time-point biological sample of the patient are used by the computing system to generate one or more health-based predictions for the patient.

11. A computer-implemented method as described in claim 10, wherein the regularization component is capable of operating to regularize the distance between the pairs of positive data values ​​and negative data values ​​in the triplet tuple of the data set relative to the anchor value in the triplet tuple of the data set, so that a first distance between the positive data values ​​and the negative data values ​​in the triplet tuple of the data set is equal to or greater than a second distance between the negative data values ​​in the triplet tuple and the anchor value.

12. A computer-implemented method as described in claim 10, wherein the triple-based machine learning model includes a triple-based deep metric learning model, and the trained triple-based machine learning model includes a trained triple-based deep metric learning model.

13. A computer-implemented method as described in claim 10, wherein the modified triple-based objective function is operable to learn similarity-based embeddings of multiple data values ​​in the data set during training of the triple-based machine learning model by the computing system using the data set and the modified triple-based objective function.

14. The computer-implemented method of claim 10, wherein the one or more health-based predictions include at least one of: classification of the patient in one or more predefined health-based classes, a health risk score prediction, a diagnostic prediction, or a prognostic prediction.

15. A computing device, comprising: one or more processors; as well as one or more non-transitory computer-readable media that collectively store instructions that, when executed by the one or more processors, cause the computing device to perform operations comprising: integrating a regularization component into a triple-based objective function associated with a triple-based machine learning model to create a modified triple-based objective function, the regularization component operable to regularize distances between pairs of positive and negative data values ​​in a triple-tuple of a dataset, the dataset comprising a healthcare dataset having at least one of a genetic characteristic, a phenotypic characteristic, or a lifestyle characteristic, relative to an anchor value in the triple-tuple; as well as A triplet-based machine learning model is trained using the dataset and the modified triplet-based objective function to create a trained triplet-based machine learning model that, when implemented, generates one or more health-based predictions for a patient based on a single time-point biological sample of the patient.

16. A computing device as described in claim 15, wherein the regularization component is capable of operating to regularize the distance between the pairs of positive data values ​​and negative data values ​​in the triplet tuple of the data set relative to the anchor value in the triplet tuple of the data set, so that the first distance between the positive data value and the negative data value in the triplet tuple of the data set is equal to or greater than the second distance between the negative data value in the triplet tuple and the anchor value.

17. A computing device as described in claim 15, the triple-based machine learning model includes a triple-based deep metric learning model, and the trained triple-based machine learning model includes a trained triple-based deep metric learning model.

18. A computing device as described in claim 15, wherein the modified triple-based objective function is capable of operating to learn similarity-based embeddings of multiple data values ​​in the data set during training of the triple-based machine learning model by the computing system using the data set and the modified triple-based objective function.

19. The computing device of claim 15, wherein the one or more health-based predictions include at least one of: classification of the patient in one or more predefined health-based classes, a health risk score prediction, a diagnosis prediction, or a prognosis prediction.

Images