Abstract: In one embodiment, a method for managing servicing of fleet vehicles includes: determining, based on data received from the fleet vehicles, current locations and statuses of the fleet vehicles in a region; determining service locations for servicing the fleet vehicles in the region, each service location being associated with a subset of the fleet vehicles; for each service location, determining a fleet vehicle demand and service task, the service task being determined at least based on the statuses of the subset of the fleet vehicles; for each service location, determining a priority score based on the service task, the demand, and the current vehicle locations; generating a service task list based on the service locations, service tasks, and priority scores; and providing for display the service task list to the user for instructing the user to service one or more of the fleet vehicles.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating data structures from graphs. The computer accesses a graph having patient nodes representing patients and patient health data nodes representing health data for patients, the nodes being connected by edges. The computer generates subgraphs by identifying patient nodes and patient health data nodes associated with a particular healthcare provider. The computer generates, by a subgraph neural network, a healthcare provider data structure for a respective subgraph. The computer generates, by a first graph neural network, patient data structures for a respective patient graph network and health data structures for a respective health data graph network. Each healthcare provider data structure, patient graph network, health data structure, has a lower dimension than the corresponding subgraph, patient graph network, and health data graph network, respectively.

Abstract: A method includes receiving data and integrating the data into a computing system. The method also includes applying a machine learning system to identify patients from the integrated data to place in one or more communities that include consumer-related data and social determinants of health data. The method also includes combining path projection, aggregation, and embedding to establish one or more paths to connect the patients to the communities based on the consumer-related data and/or the social determinants of health data in the one or more communities. The method also includes training a machine learning system to identify a correct path among the one or more established paths to place the patients on to be connected to the one or more communities.

Abstract: A deep learning model implements continuous, lifelong machine learning (LML) based on a Bayesian neural network using an inventive framework including wide, deep, and prior components that employ diverse algorithms to leverage available real-world healthcare data differently to improve prediction performance. The outputs from each component of the framework are fed into a wide and shallow neural network and the posterior structure of the final model output may be utilized as a prior structure when the deep learning model is refreshed with new data in a deep learning process. Lifelong learning is implemented by dynamically integrating present learning from the wide and deep learning components with past learning from traditional tree models in the prior component into future predictions. Thus, the present Bayesian deep neural network-based LML model increases accuracy in identifying patient profiles by continuously learning, as new data become available, without forgetting prior knowledge.

Abstract: Techniques are disclosed for fleet vehicle management. According to various embodiments, a status is determined for each fleet vehicle of a plurality of fleet vehicles. Based on the status, a service task is determined for at least a subset of the plurality of fleet vehicles. A location is determined for each service task. A service task list is generated based on the service tasks and the locations of the service tasks. A user to perform the service task list is determined based on the service tasks and the service task locations. The service task is communicated to the user and progress of the service task list is monitored.

Abstract: A computer-assisted method to timely provide notifications of treatments, the method including receiving de-identified longitudinal medical records, each de-identified longitudinal medical record representing a record of a different anonymized patient and encoding information identifying a treatment received by the anonymized patient and receiving notification data including notification records, each notification record encoding information identifying a channel through which the notification was provided. The method includes determining a first channel impact model representing an impact of a notification provided through a first channel on a treatment being received, a second channel impact model representing an impact of a notification provided through a second channel on a treatment being received, and determining a multi-channel impact model representing an impact of notifications being provided through both the first channel and the second channel on a treatment being received.

Abstract: Embodiments of the present disclosure provide a method for identifying those entities within a network that have the most influence on other entities within the network. A multi-relational network comprising links among a plurality of physicians is generated based on peer network data, wherein each link indicates a first physician that influences a second physician, and a weight of the influence. A decision by a treating physician of the plurality of physicians is decomposed, using a deep learning engine, into a magnitude of peer influence and a magnitude of control factor influence based on the multi-relational network and a plurality of control factors respectively. The magnitude of peer influence among one or more physicians in the multi-relational network is distributed among physicians in the multi-relational network based on the links each physician maintains with other physicians.

Abstract: Methods and systems to identify collaborative communities of individuals from graphs of multiple types of relationships amongst the individuals, including to mine data related to multiple types of relationships amongst individuals, construct graphs to represent the respective types of relationships amongst individuals, and perform a multiplex graph convolutional network (MGCN) artificial intelligence machine learning (AIML) analysis across the multiple graphs to identify the collaborative communities. A mathematical representation of the graphs may be learned/tuned to optimize clustering of the individuals. Multiple parameters (inter-graph weights, consensus regularization function) may be jointly tuned based on a joint optimization function. The collaborative communities may be displayed such that relative positions of the individuals represent measures of influence exerted by the respective individuals within the respective collaborative communities.

Abstract: Embodiments of the present disclosure provide a method for identifying those entities within a network that have the most influence on other entities within the network. A multi-relational network comprising links among a plurality of physicians is generated based on peer network data, wherein each link indicates a first physician that influences a second physician, and a weight of the influence. A decision by a treating physician of the plurality of physicians is decomposed, using a deep learning engine, into a magnitude of peer influence and a magnitude of control factor influence based on the multi-relational network and a plurality of control factors respectively. The magnitude of peer influence among one or more physicians in the multi-relational network is distributed among physicians in the multi-relational network based on the links each physician maintains with other physicians.

Abstract: Systems and techniques are disclosed for using machine-learning to identify potential opportunity patients that are more likely to adjust his/her preference for a healthcare provider or service. In some implementations, integrated patient data is obtained. A patient sequence feature vector, a provider sequence feature vector, and a set of entity-specific feature vectors are generated. A set of opportunity patients is identified. A notification is transmitted to the set of opportunity patients about a second treatment plan.

Abstract: Techniques are disclosed for fleet vehicle management. According to various embodiments, a status is determined for each fleet vehicle of a plurality of fleet vehicles. Based on the status, a service task is determined for at least a subset of the plurality of fleet vehicles. A location is determined for each service task. A service task list is generated based on the service tasks and the locations of the service tasks. A user to perform the service task list is determined based on the service tasks and the service task locations. The service task is communicated to the user and progress of the service task list is monitored.

Abstract: A computer-assisted method to provide timely multi-channel notification of treatments to healthcare providers and patients, the method including receiving de-identified longitudinal medical records, treatment prescription records of healthcare providers, and notification data. Relationships between the healthcare providers, the anonymized patients, and the notifications are identified using the de-identified longitudinal medical records, the treatment prescription records of the healthcare providers, and the notification data. An impact of notifications being received by both the healthcare provider for the anonymized patient and the anonymized patient on whether the anonymized patient received the treatment is determined.

Abstract: A computer-assisted method to timely provide notifications of treatments, the method including receiving de-identified longitudinal medical records, receiving notification data, identifying anonymized patients that received the treatment, identifying notifications for the treatment that were received by the recipients, determining, for each of the identified notifications, whether the recipient is an anonymized patient identified as having received the treatment, determining, for each of the identified notifications for the treatment determined to be received by a recipient that is an anonymized patient identified as having received the treatment, a time relationship between the time when the treatment was received by the anonymized patient and the time that the notification was received by the recipient that is the anonymized patient, and determining, for each of the anonymized patients that received the treatment, associations between one or more time relationships for notifications received by the anonymiz

Abstract: A computer-assisted method to timely provide notifications of treatments, the method including receiving de-identified longitudinal medical records, each de-identified longitudinal medical record representing a record of a different anonymized patient and encoding information identifying a treatment received by the anonymized patient and receiving notification data including notification records, each notification record encoding information identifying a channel through which the notification was provided. The method includes determining a first channel impact model representing an impact of a notification provided through a first channel on a treatment being received, a second channel impact model representing an impact of a notification provided through a second channel on a treatment being received, and determining a multi-channel impact model representing an impact of notifications being provided through both the first channel and the second channel on a treatment being received.

Abstract: A computer-assisted method to timely provide notifications of treatments, the method including receiving de-identified longitudinal medical records, receiving notification data, identifying anonymized patients that received the treatment, identifying notifications for the treatment that were received by the recipients, determining, for each of the identified notifications, whether the recipient is an anonymized patient identified as having received the treatment, determining, for each of the identified notifications for the treatment determined to be received by a recipient that is an anonymized patient identified as having received the treatment, a time relationship between the time when the treatment was received by the anonymized patient and the time that the notification was received by the recipient that is the anonymized patient, and determining, for each of the anonymized patients that received the treatment, associations between one or more time relationships for notifications received by the anonymiz