Abstract: The disclosure relates to systems and methods of translating event data to generate a scalable data structure to identify or predict an event of interest such as a clinical diagnosis. The scalable data structure is expandable to accommodate various types of event data each having different types of timing indications on a timeline. The system may translate the event data in a way that event data can inherit event data values from other event data in a single time series of events. The scalable data structure may be used to generate unified visualizations of all translated events as well as for forecasting and predicting events of interest. The scalable data structure may be implemented in various contexts such as for clinical diagnostics in which clinical trial data or medical health data from various sources are translated to generate the scalable data structure.

Abstract: A system may access a source image of the subject. A system may execute an image segmentation model that uses the source image to distinguish the target object from among other objects in the source image. A system may generate a binary image of the target object based on execution of the image segmentation model. A system may generate, based on the binary image, a size estimate of the target object. A system may execute a machine learning-based model that uses the size estimate and one or more covariates to predict a future size of the target object. A system may determine a risk classification for the subject based on the predicted growth rate, the risk classification being based on a probability that the target object of the subject will result in a disease state, the risk classification to be used to determine a treatment regimen to treat or prevent the disease state.

Abstract: The disclosure relates to systems and methods of generating physiological state classifications such as sleep classifications. Physiological state classification may refer to a machine-learning model's prediction of a subject's physiological state based on sensor data. In particular, the machine-learning model may generate a sleep classification that represents a prediction of a subject's sleep stage. A sleep stage may refer to whether the subject is awake or asleep (for a binary classification). In some examples, the sleep stage may refer to whether the subject is awake, N1, N2, N3, and Rapid Eye Movement (REM) (for a multi-class classification).

Abstract: Systems and methods are described for automatically inspecting and validating unstructured documents having natural language text, such as journal article describing clinical research. For example, a journal's prose may be parsed to identify domain-specific entities and values. Domain-specific rules may be evaluated against generated structured data storing the entities and their corresponding values. rule may relate to a domain-specific requirement for the document. Findings may be generated for each evaluated rule, indicating whether the document meets a corresponding requirement. Feedback indicating whether a given finding is incorrect or is to be updated, which may indicate that the corresponding rule should be updated or removed, may be obtained. Based on the feedback, the set of domain-specific rules may be updated to obtain an updated set of rules including an update to or deletion of the rule. Some embodiments include automatically validating documents using voice enabled features.

Abstract: Systems and methods are disclosed for data driven document creation and modification. The systems and methods include obtaining a first dataset having data records associated with entities, obtaining a list of entities associated with a first subset of data records in the first dataset, and obtaining configuration information, wherein the configuration information includes rules for identifying logical relationships in the data records and wherein the configuration information is specified using a vector-oriented language. The systems and methods further include extracting, for each entity in the list of entities, based on the rules, data records from the first subset of data records associated with the entity and generating a document for each entity in the list of entities using the extracted data records and the configuration information.

Abstract: Systems and methods are described for a scalable approach to build a knowledge database of clinical trial data by extracting, aligning, and synthesizing information from a variety of sources including clinical trial registries, abstracts of papers, and full-text medical journal articles, as well as external gazetteers, dictionaries, and lexicons. For examples, a system may implement a flexible and repeatable workflow that extracts both structured and semi-structured elements from unstructured data such as journal articles using a ‘back off strategy’ in which specialized rules are used to extract structured, clinical trial design parameters as well as information retrieval techniques that exploit regularities in language used in the medical literature to discover semi-structured trial outcomes.