86 results about "Unlabelled data" patented technology

A method of modelling data, comprising: training an objective function of a linear classifier, based on a set of labeled data, to derive a set of classifier weights; defining a posterior probability distribution on the set of classifier weights of the linear classifier; approximating a marginalized loss function for an autoencoder as a Bregman divergence, based on the posterior probability distribution on the set of classifier weights learned from the linear classifier; and classifying unlabeled data using the autoencoder according to the marginalized loss function.

A system capable of dynamically reserving bandwidth and adjusting bandwidth reservations for active sessions of data communication in a data communications device is provided. The system generally separates the operation of bandwidth allocation and adjustment from the operation of data transport through the device, thereby allowing bandwidth reservations and adjustments to be made without disturbing sessions of data communication that are actively being transported through the device. The system can accept requests to allocate or reserve bandwidth in a data communications device using bandwidth reservation protocols such as RSVP. The reservation requests create sender state data that can be used to compute resource allocation data. The resource allocation data can be used to label data storage locations in a data storage mechanism according to the required bandwidth reservations. A data scheduling apparatus, which is ignorant of particular sessions and specific amounts of reserved bandwidth, examines data and deposits data into data storage locations having a label corresponding to a session identification specified in the data, if any. If an unknown or no session identification is specified in the data, the data scheduler deposits data into a data storage location that is unlabeled or that has an unreserved label. Thus session bandwidth is determined by the percentage of labeled data storage locations for the session. Changes in bandwidth reservations are reflected in the separate operation of alterations made in the data storage labeling scheme, and do not affect the data scheduler, or data dequeuing mechanisms, thus allowing data sessions to continue without interruption during bandwidth adjustments.

The spectral kernel machine combines kernel functions and spectral graph theory for solving problems of machine learning. The data points in the dataset are placed in the form of a matrix known as a kernel matrix, or Gram matrix, containing all pairwise kernels between the data points. The dataset is regarded as nodes of a fully connected graph. A weight equal to the kernel between the two nodes is assigned to each edge of the graph. The adjacency matrix of the graph is equivalent to the kernel matrix, also known as the Gram matrix. The eigenvectors and their corresponding eigenvalues provide information about the properties of the graph, and thus, the dataset. The second eigenvector can be thresholded to approximate the class assignment of graph nodes. Eigenvectors of the kernel matrix may be used to assign unlabeled data to clusters, merge information from labeled and unlabeled data by transduction, provide model selection information for other kernels, detect novelties or anomalies and / or clean data, and perform supervised learning tasks such as classification.

The subject invention relates to facilitating text detection. The invention employs a boosted classifier and a transductive classifier to provide accurate and efficient text detection systems and / or methods. The boosted classifier is trained through features generated from a set of training connected components and labels. The boosted classifier utilizes the features to classify the training connected components, wherein inferred labels are conveyed to a transductive classifier, which generates additional properties. The initial set of features and the properties are utilized to train the transductive classifier. Upon training, the system and / or methods can be utilized to detect text in data under text detection, wherein unlabeled data is received, and connected components are extracted therefrom and utilized to generate corresponding feature vectors, which are employed to classify the connected components using the initial boosted classifier. Inferred labels are utilized to generate properties, which are utilized along with the initial feature vectors to classify each connected component using the transductive classifier.

A method of modelling data, comprising: training an objective function of a linear classifier, based on a set of labeled data, to derive a set of classifier weights; defining a posterior probability distribution on the set of classifier weights of the linear classifier; approximating a marginalized loss function for an autoencoder as a Bregman divergence, based on the posterior probability distribution on the set of classifier weights learned from the linear classifier; and classifying unlabeled data using the autoencoder according to the marginalized loss function.