4627results about "Kernel methods" patented technology

Systems and methods for massively parallel processing on an accelerator that includes a plurality of processing cores. Each processing core includes multiple processing chains configured to perform parallel computations, each of which includes a plurality of interconnected processing elements. The cores further include multiple of smart memory blocks configured to store and process data, each memory block accepting the output of one of the plurality of processing chains. The cores communicate with at least one off-chip memory bank.

A separation surface set storage part stores information defining a plurality of separation surfaces which separate a feature space into at least one known class region respectively corresponding to at least one known class and an unknown class region. Each of the at least one known class region is separated from outside region by more than one of the plurality of separation surfaces which do not intersect to each other. A data classification apparatus determine a classification of a classification target data whose inner product in the feature space is calculable by calculating to which region of the at least one known class region and the unknown class region determined by the information stored in the separation surface set storage part the classification target data belongs. A method and apparatus for data classification which can simultaneously perform identification and outlying value classification with high reliability in a same procedure are provided.

The present invention is directed to a computer automated method of selectively identifying a user specified behavior of a crowd. The method comprises receiving video data but can also include audio data and sensor data. The video data contains images a crowd. The video data is processed to extract hierarchical human and crowd features. The detected crowd features are processed to detect a selectable crowd behavior. The selected crowd behavior detected is specified by a configurable behavior rule. Human detection is provided by a hybrid human detector algorithm which can include Adaboost or convolutional neural network. Crowd features are detected using textual analysis techniques. The configurable crowd behavior for detection can be defined by crowd behavioral language.

Digitized image data are input into a processor where a detection component identifies the areas (objects) of particular interest in the image and, by segmentation, separates those objects from the background. A feature extraction component formulates numerical values relevant to the classification task from the segmented objects. Results of the preceding analysis steps are input into a trained learning machine classifier which produces an output which may consist of an index discriminating between two possible diagnoses, or some other output in the desired output format. In one embodiment, digitized image data are input into a plurality of subsystems, each subsystem having one or more support vector machines. Pre-processing may include the use of known transformations which facilitate extraction of the useful data. Each subsystem analyzes the data relevant to a different feature or characteristic found within the image. Once each subsystem completes its analysis and classification, the output for all subsystems is input into an overall support vector machine analyzer which combines the data to make a diagnosis, decision or other action which utilizes the knowledge obtained from the image.

In a pre-processing step prior to training a learning machine, pre-processing includes reducing the quantity of features to be processed using feature selection methods selected from the group consisting of recursive feature elimination (RFE), minimizing the number of non-zero parameters of the system (lo-norm minimization), evaluation of cost function to identify a subset of features that are compatible with constraints imposed by the learning set, unbalanced correlation score and transductive feature selection. The features remaining after feature selection are then used to train a learning machine for purposes of pattern classification, regression, clustering and / or novelty detection.

A predictive modeling method may include determining a time interval of time-series data; identifying one or more variables of the data as targets; determining a forecast range and a skip range associated with a prediction problem represented by the data; generating training data and testing data from the time-series data; fitting a predictive model to the training data; and testing the fitted model on the testing data. The forecast range may indicate a duration of a period for which values of the targets are to be predicted. The skip range may indicate a temporal lag between the time period corresponding to the data used to make predictions and the time period corresponding to the predictions. The skip range may separate input data subsets representing model inputs from subsets representing model outputs, and separate test data subsets representing model inputs from subsets representing validation data.

The present disclosure describes systems and methods of estimating sedation level of a patient using machine learning. For example, the integration of multiple QEEG features into a single sedation level estimation system using machine learning could result in a significant improvement in the predictability of the levels of sedation, independent of the sedative drug used. The present disclosure advantageously allows for the incorporation of large numbers of QEEG features and machine learning into the next-generation monitors of sedation level. Different QEEG features may be selected for different sedation drugs, such as propofol, sevoflurane and dexmedetomidine groups. The sedation level estimation system can maintain a high performance for detecting MOAA / S, independent of the drug used.

A request from a client is received to generate a differentially private random forest classifier trained using a set of restricted data. The differentially private random forest classifier is generated in response to the request. Generating the differentially private random forest classifier includes determining a number of decision trees and generating the determined number of decision trees. Generating a decision tree includes generating a set of splits based on the restricted data, determining an information gain for each split, selecting a split from the set using an exponential mechanism, and adding the split to the decision tree. The differentially private random forest classifier is provided to the client.

A 1D-CNN-based ((one-dimensional convolutional neural network)-based) distributed optical fiber sensing signal feature learning and classification method is provided, which solves a problem that an existing distributed optical fiber sensing system has poor adaptive ability to a complex and changing environment and consumes time and effort due to adoption of manually extracted distinguishable event features, The method includes steps of: segmenting time sequences of distributed optical fiber sensing acoustic and vibration signals acquired at all spatial points, and building a typical event signal dataset; constructing a 1D-CNN model, conducting iterative update training of the network through typical event signals in a training dataset to obtain optimal network parameters, and learning and extracting 1D-CNN distinguishable features of different types of events through an optimal network to obtain typical event signal feature sets; and after training different types of classifiers through the typical event signal feature sets, screening out an optimal classifier.

Methods for training machines to categorize data, and / or recognize patterns in data, and machines and systems so trained. More specifically, variations of the invention relates to methods for training machines that include providing one or more training data samples encompassing one or more data classes, identifying patterns in the one or more training data samples, providing one or more data samples representing one or more unknown classes of data, identifying patterns in the one or more of the data samples of unknown class(es), and predicting one or more classes to which the data samples of unknown class(es) belong by comparing patterns identified in said one or more data samples of unknown class with patterns identified in said one or more training data samples. Also provided are tools, systems, and devices, such as support vector machines (SVMs) and other methods and features, software implementing the methods and features, and computers or other processing devices incorporating and / or running the software, where the methods and features, software, and processors utilize specialized methods to analyze data.

In a pre-processing step prior to training a learning machine, pre-processing includes reducing the quantity of features to be processed using feature selection methods selected from the group consisting of recursive feature elimination (RFE), minimizing the number of non-zero parameters of the system (lo-norm minimization), evaluation of cost function to identify a subset of features that are compatible with constraints imposed by the learning set, unbalanced correlation score and transductive feature selection. The features remaining after feature selection are then used to train a learning machine for purposes of pattern classification, regression, clustering and / or novelty detection. (FIG. 3, 300, 301, 302, 304, 306, 308, 309, 310, 311, 312, 314)

Systems and methods for enhancing knowledge discovery from data using multiple learning machines in general and multiple support vector machines in particular. Training data for a learning machine is pre-processed in order to add meaning thereto. Multiple support vector machines, each comprising distinct kernels, are trained with the pre-processed training data and are tested with test data that is pre-processed in the same manner. The test outputs from multiple support vector machines are compared in order to determine which of the test outputs if any represents a optimal solution. Selection of one or more kernels may be adjusted and one or more support vector machines may be retrained and retested. Optimal solutions based on distinct input data sets may be combined to form a new input data set to be input into one or more additional support vector machine. The methods, systems and devices of the present invention comprise use of Support Vector Machines for the identification of patterns that are important for medical diagnosis, prognosis and treatment. Such patterns may be found in many different datasets. The present invention also comprises methods and compositions for the treatment and diagnosis of medical conditions.

A system and method for reviewing documents are provided. A collection of documents is portioned into sets of documents for review by a plurality of reviewers. For each set, documents in the set are displayed on a display device for review by a reviewer and temporarily organized through grouping and sorting. The reviewer's labels for the displayed documents are received. Based on the reviewer's labels, a class from a plurality of classes is assigned to each of the reviewed documents. A classifier model stored in computer memory is progressively trained, based on features extracted from the reviewed documents in the set and their assigned classes. Prior to review of all documents in the set, a calculated subset of documents for which the classifier model assigns a class different from the one assigned based on the reviewer's label is returned for a second review by a reviewer. Models generated from one or more other document sets can be used to assess the review of a first of the sets.

A classification system includes memory which stores, for each of a set of classes, a classifier model for assigning a class probability to a test sample from a target domain. The classifier model has been learned with training samples from the target domain and from at least one source domain. Each classifier model models the respective class as a mixture of components, the component mixture including a component for each source domain and a component for the target domain. Each component is a function of a distance between the test sample and a domain-specific class representation which is derived from the training samples of the respective domain that are labeled with the class, each of the components in the mixture being weighted by a respective mixture weight. Instructions, implemented by a processor, are provided for labeling the test sample based on the class probabilities assigned by the classifier models.

Providing enhanced video content includes processing at least one video feed through at least one spatiotemporal pattern recognition algorithm that uses machine learning to develop an understanding of a plurality of events and to determine at least one event type for each of the plurality of events. The event type includes an entry in a relationship library detailing a relationship between two visible features. Extracting and indexing a plurality of video cuts from the video feed is performed based on the at least one event type determined by the understanding that corresponds to an event in the plurality of events detectable in the video cuts. Lastly, automatically and under computer control, an enhanced video content data structure is generated using the extracted plurality of video cuts based on the indexing of the extracted plurality of video cuts.

The Collateral Management with Blockchain and Smart Contracts Apparatuses, Methods and Systems (“CMBSC”) transforms borrow transaction request inputs via CMBSC components into borrow transaction init notification, borrow transaction sync notification outputs. A borrow transaction request associated with a borrow transaction is obtained. Transaction attributes associated with the borrow transaction are stored in a database. The transaction process optimizer component is notified regarding the borrow transaction. A blockchain sync notification associated with the borrow transaction is obtained from the transaction process optimizer component. The stored transaction attributes associated with the borrow transaction are filtered. A smart contract associated with the borrow transaction is generated. The generated smart contract is sent to a blockchain node of a blockchain network. A smart contract notification associated with the smart contract is received. A push notification regarding the smart contract notification is provided to a user interface component of a user's client.

Systems and techniques are provided for linking subjects in an area of real space with user accounts. The user accounts are linked with client applications executable on mobile computing devices. A plurality of cameras are disposed above the area. The cameras in the plurality of cameras produce respective sequences of images in corresponding fields of view in the real space. A processing system is coupled to the plurality of cameras. The processing system includes logic to determine locations of subjects represented in the images. The processing system further includes logic to match the identified subjects with user accounts by identifying locations of the mobile computing devices executing client applications in the area of real space and matching locations of the mobile computing devices with locations of the subjects.

The present invention relates to a system and methodology to facilitate multiattribute adjustments and control associated with messages and other communications and informational items that are directed to a user via automated systems. An interface, specification language, and controls are provided for defining a plurality of variously configured groups that may attempt to communicate respective items. Controls include the specification of priorities and preferences as well as the modification of priorities and preferences that have been learned from training sets via machine learning methods. The system provides both a means for assessing parameters used in the control of messaging and communications and for the inspection and modification of parameters that have been learned autonomously.

A machine learning system for ranking a collection of filtered propensity to failure metrics of like components within an electrical grid that includes a raw data assembly to provide raw data representative of the like components within the electrical grid; (b) a data processor, operatively coupled to the raw data assembly, to convert the raw data to more uniform data via one or more data processing techniques; (c) a database, operatively coupled to the data processor, to store the more uniform data; (d) a machine learning engine, operatively coupled to the database, to provide a collection of propensity to failure metrics for the like components; (e) an evaluation engine, operatively coupled to the machine learning engine, to detect and remove non-complying metrics from the collection of propensity to failure metrics and to provide the collection of filtered propensity to failure metrics; and (f) a decision support application, operatively coupled to the evaluation engine, configured to display a ranking of the collection of filtered propensity to failure metrics of like components within the electrical grid.

Embodiments are directed towards a configurable accelerator framework device that includes a stream switch and a plurality of convolution accelerators. The stream switch has a plurality of input ports and a plurality of output ports. Each of the input ports is configurable at run time to unidirectionally pass data to any one or more of the output ports via a stream link. Each one of the plurality of convolution accelerators is configurable at run time to unidirectionally receive input data via at least two of the plurality of stream switch output ports, and each one of the plurality of convolution accelerators is further configurable at run time to unidirectionally communicate output data via an input port of the stream switch.

The present invention relates to a method for closed-loop controlling a processing operation of a workpiece, comprising the steps of: (a) recording a pixel image at an initial time point of an interaction zone by means of a camera, wherein the workpiece is processed using an actuator having an initial actuator value; (b) converting the pixel image into a pixel vector; (c) representing the pixel vector by a sum of predetermined pixel mappings each multiplied by a corresponding feature value; (d) classifying the set of feature values on the basis of learned feature values into at least two classes of a group of classes comprising a first class of a too high actuator value, a second class of a sufficient actuator value and a third class of a too low actuator value at the initial time point; (e) performing a control step for adapting the actuator value by minimizing the error et between a quality indicator ye and a desired value; and (f) repeating the steps (a) to (e) for further time points to perform a closed-loop controlled processing operation.

A traffic classifier has a plurality of binary classifiers, each associated with one of a plurality of calibrators. Each calibrator trained to translate an output score of the associated binary classifier into an estimated class probability value using a fitted logistic curve, each estimated class probability value indicating a probability that the packet flow on which the output score is based belongs to the traffic class associated with the binary classifier associated with the calibrator. The classifier training system configured to generate a training data based on network information gained using flow and packet sampling methods. In some embodiments, the classifier training system configured to generate reduced training data sets, one for each traffic class, reducing the training data related to traffic not associated with the traffic class.

Systems and methods implemented by a computer to detect abnormal behavior in a network include obtaining Performance Monitoring (PM) data including one or more of production PM data, lab PM data, and simulated PM data; determining a model based on machine learning training with the PM data; receiving live PM data from the network; utilizing the live PM data with the model to detect an anomaly in the network; and causing an action to address the anomaly.

A system is disclosed for chatbot conversation construction, and includes user interface, coupled to a conversation construction and management platform, configured to provide entry of intents, associated responses and search terms, and to construct and store conversation controllers, each based at least in part on one or more of the intents, and build conversation flows, each based at least in part on one or more of the conversation controllers. An artificial intelligence (AI) connector platform couples to the conversation construction and management platform, and is configured to receive natural language (NL) search data from a chatbot channel, format the NL search data into an input protocol of a selectable NL processing engine and receive a response from the NLP engine and, based at least in part on one or more of the conversation flows, convert the NLP engine response to a personalized response, and send the personalized response to the chatbot channel.

The present disclosure describes a computer-implemented method of transforming a low-resolution MR image to a high-resolution MR image using a deep CNN-based MRI SR network and a computer-implemented method of transforming an MR image to a pseudo-CT (sCT) image using a deep CNN-based sCT network. The present disclosure further describes a MR image-guided radiation treatment system that includes a computing device to implement the MRI SR and CT networks and to produce a radiation plan based in the resulting high resolution MR images and sCT images.

The present invention is directed to a method for detecting unknown malicious code, such as a virus, a worm, a Trojan Horse or any combination thereof. Accordingly, a Data Set is created, which is a collection of files that includes a first subset with malicious code and a second subset with benign code files and malicious and benign files are identified by an antivirus program. All files are parsed using n-gram moving windows of several lengths and the TF representation is computed for each n-gram in each file. An initial set of top features (e.g., up to 5500) of all n-grams IS selected, based on the DF measure and the number of the top features is reduced to comply with the computation resources required for classifier training, by using features selection methods. The optimal number of features is then determined based on the evaluation of the detection accuracy of several sets of reduced top features and different data sets with different distributions of benign and malicious files are prepared, based on the optimal number, which will be used as training and test sets. For each classifier, the detection accuracy is iteratively evaluated for all combinations of training and test sets distributions, while in each iteration, training a classifier using a specific distribution and testing the trained classifier on all distributions. The optimal distribution that results with the highest detection accuracy is selected for that classifier.