An aspect of the present disclosure facilitates accelerating outlier prediction of performance metrics in performance managers deployed in new computing environments. In one embodiment, a digital processing system receives an input data specifying a business vertical to which a new computing environment is directed, a performance metric of interest, and a computing component of the new computing environment for which the performance metric is sought to be measured. In response, the system selects, from a set of prediction models, a prediction model for the performance metric, based on the input data. The selected prediction model is then used in a performance manager to predict outliers for the performance metric of interest during operation of the new computing environment.

A candidate generator generates a set of candidate three-dimensional image patches from an input volume. A candidate classifier classifies the set of candidate three-dimensional image patches as containing or not containing disease. Classifying the set of candidate three-dimensional image patches comprises generating an attention mask for each given candidate three-dimensional image patch within the set of candidate three-dimensional image patches to form a set of attention masks, applying the set of attention masks to the set of candidate three-dimensional image patches to form a set of masked image patches, and classifying the set of masked image patches as containing or not containing the disease. The candidate classifier applies soft attention and hard attention to the three-dimensional image patches such that distinctive image regions are highlighted proportionally to their contribution to classification while completely removing image regions that may cause confusion.

An object detection training method can include receiving a training sample set in a current iteration of an object detection training process over an object detection neural network. The training sample set can include first samples of a first class and second samples of a second class. A first center loss value of each of the first and second samples can be determined. The first center loss value can be a distance between a feature vector of the respective sample and a center feature vector of the first or second class which the respective sample belongs to. A second center loss value of the training sample set can be determined according to the first center loss values of the first and second samples. A first target loss value of the current iteration can be determined according to the second center loss value of the training sample set.

A device for and computer implemented method of image content recognition and of training a neural network for image content recognition. The method comprising collecting a first set of digital images from a database, the first set of digital images is sampled from digital images assigned to a many shot class; creating a first training set comprising the collected first set of digital images; training a first artificial neural network comprising a first feature extractor and a first classifier for classifying digital images using the first training set; collecting first parameters of the trained first feature extractor, collecting second parameters of the trained classifier, determining third parameters of a second feature extractor of a second artificial neural network depending on the first parameters, determining fourth parameters of a second classifier for classifying digital images of the second artificial neural network.

Aspects of the disclosure are directed to receiving numerical product data indicative of a product, the numerical product data comprising numerical values indicative of at least one of chemical composition, radius, tensile strength, a diameter, position and yield strength, and storing the numerical product data in a non-volatile memory device. The numerical product data is processed in a processor to create a plurality of explanatory variables indicative of the numerical product data. Multivariate data analysis is performed on the explanatory variables indicative of the numerical product data, where the multivariate data analysis includes an iterative cluster based outlier detection procedure. A confidence indicator value indicative of the likelihood that the numerical product data includes at least one fabricated or false data entry is generated.

Techniques are provided detecting diluted drugs using machine learning. Measurements and images corresponding to a product are obtained, wherein the product is formulated as a liquid, and wherein the measurements and images capture physical, spectral, optical, and/or chemical properties of the product. The measurements and images are provided to a machine learning model, wherein the machine learning model is trained using data generated from interactive learning modules (e.g., a generative adversarial network). The machine learning model detects whether the product or chemical is a real or counterfeit product. In addition, these techniques may be used by practitioners (e.g., medical personnel dispensing a prescribed dosage of a drug with a specific dilution level) to detect prescription errors at the point of administration.

A classifier network has at least two distinct sets of refined data, wherein the first two sets of refined data are sets of numbers representing the features values data received from sensors or a manufactured part. Performing, via at least two distinct types of support vector machines using an associated feature selection process for each classifier independently in a first layer, anomaly detection on the manufactured part. Then, using the stored data including refined data of at least two different types of data transforms and performing, via at least a two distinct types of support vector machines in a second layer, an associated feature selection process for each classifier independently. Forming at least four distinct compound classifier types for anomaly detection on the part using the stored data or coefficients. The ensemble of second layer support vector machine outputs compare the results to determine the presence of an anomaly.

Various embodiments include a method for product quality inspection on a group of products. The method may include: getting for each product in the group of products: image, value for each known fabrication parameter affecting quality of the group of products, and quality evaluation result; training a neural network. A layer of the neural network comprises at least one first neuron and at least one second neuron; each first neuron represents a known fabrication parameter affecting quality of the group of products and each second neuron represents an unknown fabrication parameter affecting quality of the group of products; and the images of the group of products are input to the neural network, the quality evaluation results are output of the neural network, and the value of each first neuron is set to the value for the known fabrication parameter the first neuron representing.

In general, one aspect disclosed features a system, comprising: a hardware processor; and a non-transitory machine-readable storage medium encoded with instructions executable by the hardware processor to perform operations comprising: receiving an electronic record, the electronic record representing a medical bill, the medical bill comprising a plurality of attributes; mapping each attribute in the medical bill to a single bucket of a predetermined second quantity of the buckets according to a predetermined correspondence between the attributes and the buckets, the first quantity exceeding the second quantity; and providing identifiers of the single buckets as input to a machine learning model, the machine learning model being trained according to historical correspondences between the buckets and decisions of whether human review was necessary, wherein responsive to the input, the machine learning model provides as output an indication of whether the medical bill should be reviewed by a human.

Systems, apparatuses, and methods include technology that identifies a first dataset that comprises a plurality of data values, and partitions the first dataset into a plurality of bins to generate a second dataset, where the second dataset is a compressed version of the first dataset. The technology randomly subsamples data associated with the first dataset to obtain groups of randomly subsampled data, and generates a plurality of decision tree models during an unsupervised learning process based on the groups of randomly subsampled data and the second dataset.