Abstract: Systems and methods for dynamically customizing a virtual assistant are disclosed. The systems and methods can receive information associated with a conversation involving the virtual assistant; determine whether a channel switching condition for switching the conversation from a first channel to a second channel is satisfied, based on the information associated with the conversation; determine whether a language switching condition for switching the conversation from a first language to a second language is satisfied, based on the information associated with the conversation; determine whether a configuration switching condition for switching a first configuration of the virtual assistant to a second configuration of the virtual assistant is satisfied, based on the information associated with the conversation; and perform an action based on at least one of the determinations.

Abstract: Systems and methods are configured to perform machine-learning-based predictive behavioral response. In various embodiments, one or more behavioral monitoring data objects are identified and processed using a behavioral pattern prediction machine learning model to generate a behavioral pattern prediction model. The behavioral pattern prediction model is processed using a risk generation machine learning model to generate a risk model, wherein: (i) the risk generation machine learning model is generated based at least in part by one or more risk factors, and (ii) the risk model comprises a per-risk factor score for each risk factor of the one or more risk factors. The risk model is processed using an adjustment generation machine learning model to generate an adjustment model and one or more prediction-based actions are performed based on the adjustment model.

Abstract: Systems and methods are disclosed for generating a personalized care path for a patient. The method includes receiving, by one or more processors, relevant data associated with the patient from a plurality of data sources. The relevant data includes demographic data and medical data associated with the patient. The one or more processors using a graph convolutional neural network-based model determine the personalized care path for the patient based on the relevant data associated with the patient. The graph convolutional neural network-based model is trained based on a plurality of care paths of a plurality of patients represented by a patient-bucket-procedure (PBP) graph. The one or more processors provide data associated with the determined personalized care path for the patient to a device associated with a user.

Abstract: A pill counting device includes a step feeder having moveable slats interleaved with fixed slats. The movable slats are vertically offset from one another and are configured to move vertically between top platforms of adjacent fixed slats to carry pills vertically up the step feeder. A top movable slat orients a subset of the plurality of pills into a row of pills to be fed into a count chamber. A camera configured to capture an image of a pill of the individual pills as it falls through the count chamber. The device further includes a control system configured to control movement of the movable slats and to control capturing of images by the camera. The control system is further configured to calculate a characteristic of the pill from the image, and to increment a pill counter in response to the characteristic of the pill meeting predefined criteria.

Abstract: Systems and methods are disclosed for processing medical claims to determine vector embeddings for predictive analytics and fraud detection. The method includes receiving claim data for a medical claim, the claim data comprising a plurality of medical codes. A plurality of embedding vectors is determined based on the plurality of medical codes, each embedding vector determined based on a corresponding medical code. A multi-layered matrix is determined based on the plurality of embedding vectors determined for the plurality of medical codes, wherein the multi-layered matrix is representative of the medical claim. One or more actions such as predictive analytics and fraud detection are performed based on the multi-layered matrix.

Abstract: Various embodiments of the present disclosure provide methods, apparatus, systems, computing devices, computing entities, and/or the like for classifying unstructured data by: (i) generating probability scores of natural language classification labels for classifying unstructured data elements using an NLP-based model, (ii) generating probability scores of structured data classification labels for classifying the unstructured data elements using a classification-based model, and (iii) assigning classifications labels based on: a) the probability scores of the natural language classification labels if a distance measure difference associated with the natural language classification labels is greater than a predetermined distance, or b) a determination using an ensemble model if the distance measure difference is less than a predetermined distance.

Abstract: A method includes predicting, before a first period of time, a plurality of predicted event metric values and value ranges corresponding to a first plurality of events to occur over the first period of time based on a plurality of prediction models, determining an actual event metric value of the first plurality of events, and determining a plurality of model range multi-partition error (MRME) values indicative of an accuracy, precision, and relative range of the forecasting models based on the actual event metric values, value ranges, and the predicted event metric values. The method includes selecting one of the forecasting models based on the MRME values, determining a cyber security control limit based on a predicted event metric value or predicted event metric range by the selected forecasting model, and outputting an alert based on an actual event metric value being outside of the cyber security control limit.

Abstract: A method can include receiving prescription information identifying a prescription for an account to be filled in a recurring manner at multiple fill time points; receiving payment option information; receiving a single selection indicating a standing instruction to fill the prescription for each of the multiple fill time points at a respective lowest payment amount for the account regardless of a plurality of payment options; based on the single selection and for each of the multiple fill time points: obtaining, from a plurality of external data sources via a network, a plurality of payment amounts, determining the respective lowest payment amount for the respective fill time point among the plurality of payment amounts, selecting a payment option among the plurality of payment options associated with the respective lowest payment amount for the respective fill time point, and causing the prescription to be filled using the selected payment option.

Abstract: There is a need for more effective and efficient seasonally-adjusted predictive data analysis solutions. This need can be addressed by, for example, solutions for performing seasonally-adjusted predictive data analysis that use autoregressive integrated moving average (ARIMA) machine learning models.

Abstract: There is a need for more effective and efficient anomaly detection. This need can be addressed by, for example, solutions for performing/executing graph convolutional anomaly detection. In one example, a method includes identifying related graph database input data associated with a predictive entity; generating related graph feature data for the predictive entity; generating, based on the related graph feature data and using a graph convolutional neural network model, an anomaly detection score for the predictive entity, wherein at least a portion of the graph convolutional neural network model is trained using confirmation feedback data; performing an anomaly confirmation to generate the confirmation feedback data object for the predictive entity, and integrating the confirmation feedback data object for the predictive entity into the confirmation feedback data associated with the graph convolutional anomaly detection.

Abstract: An example method includes obtaining first audio data representing one or more initial utterances during an interactive voice session with an interactive voice system (IVS); generating, based on the first audio data, a prediction regarding whether a subsequent utterance of the user in the interactive voice session will contain sensitive information; obtaining second audio data representing the subsequent utterance; determining, based on the prediction, whether to transmit the second audio data to the IVS via a first communication channel; based on a determination not to transmit the second audio data to the IVS via the first communication channel: transmitting third audio data to the IVS via the first communication channel in place of the second audio data; and transmitting the second audio data to a server via a second communication channel that bypasses the IVS.

Abstract: Various embodiments of the disclosure provide apparatuses, systems, and computer program products for predictive data labelling using a dual-model system. Embodiments provide various advantages in accuracy of predicted labels, for example in various contexts such as medical data analysis for difficult to diagnose diseases.

Abstract: Remote patient management involves receiving an enrollment request to enroll a patient, identification of vitals of the patient to be measured based on the enrollment request, providing an equipment request indicating the vitals of the patient to be measured, receiving equipment information corresponding to one or more health measurement devices for measuring the vitals of the patient, and associating the patient with the equipment information corresponding to the one or more health measurement devices. In addition, a user device establishes communication with a health measurement device and determines whether it is associated with the patient, and if so, receives measurements from the health measurement device and provides the measurements to a healthcare management server. Otherwise, if not associated with the patient, no measurements are received from the health measurement device.

Abstract: Various embodiments of the present invention provide methods, apparatus, systems, computing devices, computing entities, and/or the like for performing predictive prioritization. Certain embodiments utilize systems, methods, and computer program products that perform predictive prioritization using a combination of supervised machine learning models and unsupervised machine learning models that are in turn used to generate target features for a resultant prioritization machine learning model.

Abstract: Systems and methods are disclosed for processing medical claims to determine vector embeddings for predictive analytics and fraud detection. The method includes receiving claim data for a medical claim, the claim data comprising a plurality of medical codes. A plurality of embedding vectors is determined based on the plurality of medical codes, each embedding vector determined based on a corresponding medical code. A multi-layered matrix is determined based on the plurality of embedding vectors determined for the plurality of medical codes, wherein the multi-layered matrix is representative of the medical claim. One or more actions such as predictive analytics and fraud detection are performed based on the multi-layered matrix.

Abstract: Hierarchical data objects are generated via a computer-based system for applying a series of rules to establish episode-specific data objects reflecting a plurality of discrete claim records before further dissecting the generated episode-specific data objects prior to finalization of those episode-specific data objects to identify claim records within the episode-specific data objects that are eligible for generation of one or more sub-episodes within the episode-specific data objects. The identified sub-episodes are reflected within the episode-specific data object to designate complete episodes of care that additionally reflect interactions with the corresponding parent episode.

Abstract: Systems and methods are disclosed for scheduling appointments based on changing health conditions of users. The method includes inputting the initial health dataset of a user to a machine learning model configured to identify other users with similar health profiles. A frequency of monitoring, a frequency of appointments, a duration between the appointments, or a type and length of the appointments is determined to generate a schedule of appointments. A subsequent health dataset of the user is received, and health scores, rule scores, or medication scores for the user are determined based on the subsequent health dataset. A plurality of risk scores for the user is evaluated based on the health scores, rule scores, or medication scores. A recent risk score is determined based on a change in the plurality of risk scores. The schedule of appointments is adjusted based on the recent risk score and the user is notified.

Abstract: Various embodiments are directed to performing automated software testing configuration management operations. In one embodiment, a method is provided. The method includes generating a graph data structure including nodes representing test steps across test cases of a software testing data object for a software application platform and including segments each connecting two nodes and representing test cases having the two represented test steps. The method further includes receiving a query describing a testing change and generating extracted features from the query. The method further includes generating test outcome requirements and graph portions of the graph data structure relevant to the testing change using supervised machine learning models and the extracted features.

Abstract: There is a need for more effective and efficient predictive data analysis. This need can be addressed by, for example, solutions for performing/executing hybrid input predictive data analysis.

Abstract: A method comprises: storing, by a computing system, a relational data in a data lake; spinning-up, by an orchestration system of the computing system, an ephemeral computing node on a computing device of the computing system; importing, by the ephemeral computing node, a copy of the relational data from the data lake into a relational database management system (RDBMS) installed on the ephemeral computing node; after importing the copy of the relational data into the RDBMS, performing, by the ephemeral computing node, a computing process that runs within the RDBMS and that uses the copy of the relational data; and after completion of the computing process on the ephemeral computing node, spinning-down, by the orchestration system, the ephemeral computing node.