Abstract: This disclosure relates to methods and apparatus for generating real-time alerts about a patient. In one arrangement, vital sign data representing vital sign information obtained from the patient at one or more input times within an assessment time window is received. A Gaussian process model of at least a portion of the vital sign information is used to generate a time series of synthetic vital sign data based on the received vital sign data, the synthetic vital sign data comprising at least a posterior mean for each of one or more components of the vital sign information at each of a plurality of regularly spaced time points in the assessment time window. The generated synthetic vital sign data is used as input to a trained recurrent neural network to generate an early warning score, the early warning score representing a probability of an adverse event occurring during a prediction time window of predetermined length after the assessment time window.

Abstract: Methods and apparatus for monitoring a patient are provided. In one arrangement, a multi- dimensional patient data set is received at each of a plurality of different reference times. Each dimension of the patient data set stores a value representing a different type of information about the patient. A plurality of predictions of a health trajectory of the patient are generated. Each prediction is generated using a trained machine learning model receiving as input a different one of the patient data sets. The trained machine learning model may be dimensionally adaptive, such that predictions of the patient trajectories are provided using patient data sets having different respective dimensionalities for at least a sub-set of the reference times. The trained machine learning model may use machine learned predictions of accuracy to select trained machine learning units from an ensemble of trained machine learning units.

Abstract: Methods and apparatus for classifying subjects based on time series phenotypic data are disclosed. In one arrangement, a data receiving unit receives a set of first subject-data-units, each first subject-data-unit in the set comprising time series data representing phenotypic information about a different respective one of a plurality of subjects to be classified. A data processing unit processes the set of first subject-data-units to reduce a dimensionality of each first subject-data-unit, thereby obtaining a corresponding set of second subject-data-units having lower dimensionality than the first subject-data-units. The set of second subject-data-units is processed to cluster the second subject-data-units into a plurality of clusters. Each of one or more of the subjects is classified by determining to which cluster a second subject-data-unit corresponding to the subject belongs.

Abstract: Methods and apparatus for subtyping subjects based on phenotypic information are disclosed. In one arrangement, a data receiving unit receives a subject data unit for each of a plurality of subjects. Each subject data unit represents a plurality of different phenotypic information items about the subject. A data processing unit uses a deep learning algorithm to derive a lower dimensional representation of each subject data unit and a clustering algorithm to detect clusters of the resulting lower dimensional representations. The deep learning algorithm and clustering algorithm are implemented by a single mathematical model in which the derivation of the lower dimensional representations and the detection of the clusters are performed jointly.

Abstract: A patient status monitor for providing an estimate of the risk of an adverse health event such as death or intensive care readmission based on a first risk estimate using patient data collected over a first time period, such as a stay in an intensive care unit, and a second risk estimate based on current vitals signs measurements. The first and second risk estimates are based on different models, the first being a static logistic regression on data selected from electronic patient records, and the second being a novelty detection algorithm based on a a training data set of vital signs measurements representing normality. The two risk estimates are combined in a weighted combination, with the first risk estimate having a weight which decreases with time from the end of the first time period.

Abstract: A structure for eliminating or reducing noise emission from a structure which vibrates at two or more discrete frequencies is described. Vibration damping tiles are secured to the surface of the structure with each tile have a high density layer linked to the structure by a layer of plastic material, such as polyurethane foam, with the mass of the high density layer, the dimensions of the layer of the plastic material and the properties of the plastic material selected so that the compression mode and shear mode resonant frequencies correspond to two of the discrete frequencies of the structure. The noise suppression device is particularly suited for use with wind turbines.