7937 results about "Learning models" patented technology

A system and method for organizing knowledge in such a way that humans can find knowledge, learn from it, and add to it as needed is disclosed. The exemplary system has four components: a knowledge base (215), a learning model (21) and an associated tutor (13), a set or user tools (31, 32), and a backend system (33). The invention also preferably comprises a set of application programming interfaces (APIs) that allow these components to work together, so that other people can create their own versions of each of the components. In the knowledge web a community of people with knowledge to share put knowledge in the database using the user tools. The knowledge may be in the form of documents or other media, or it may be a descriptor of a book or other physical source. Each piece of knowledge is associated with various types of meta-knowledge about what the knowledge is for, what form it is in, and so on.

A system, computer readable storage medium, and computer-implemented method presents video search results responsive to a user keyword query. The video hosting system uses a machine learning process to learn a feature-keyword model associating features of media content from a labeled training dataset with keywords descriptive of their content. The system uses the learned model to provide video search results relevant to a keyword query based on features found in the videos. Furthermore, the system determines and presents one or more thumbnail images representative of the video using the learned model.

The invention provides a text emotion classification method based on the joint deep learning model which relates to the text emotion classification method. The method is designed with the object of solving the problems with the dimension disaster and sparse data incurred from the existing support vector machine and other shallow layer classification methods. The method comprises: 1) processing each word in the text data; using the word2vec tool to train each processed word in the text data so as to obtain a word vector dictionary; 2) obtaining the matrix M of each sentence; training the matrix M by the LSTM layer and converting it into vector with fixed dimensions; improving the input layer; generating d-dimensional h word vectors with context semantic relations; 3) using a CNN as a trainable characteristic detector to extract characteristics from the d-dimensional h word vectors with context semantic relations; and 4) connecting the extracted characteristics in order; outputting to obtain the probability of each classification wherein the classification with the maximal probability value is the predicated classification. The invention is applied to the natural language processing field.

A system and method for learning models from scarce and/or skewed training data includes partitioning a data stream into a sequence of time windows. A most likely current class distribution to classify portions of the data stream is determined based on observing training data in a current time window and based on concept drift probability patterns using historical information.

The present disclosure relates generally to an integrated machine learning platform. The machine learning platform can convert machine learning models with different schemas into machine learning models that share a common schema, organize the machine learning models into model groups based on certain criteria, and perform pre-deployment evaluation of the machine learning models. The machine learning models in a model group can be evaluated or used individually or as a group. The machine learning platform can be used to deploy a model group and a selector in a production environment, and the selector may learn to dynamically select the model(s) from the model group in the production environment in different contexts or for different input data, based on a score determined using certain scoring metrics, such as certain business goals.

A system and method of operating a monitoring and response system for an actor in a daily living environment that relies upon learned models of behavior for adapting system operation. The learned model of behavior preferably includes sequential patterns organized pursuant to assigned partition values that in turn are generated based upon an evaluation of accumulated data. Based upon reference to the learned model of behavior, the system can generate more appropriate response plans based upon expected or unexpected activities, more readily recognize intended activities, recognize abandoned tasks, formulate probabilities of method choice, build probabilities of action success, anticipate and respond to actor movement within the environment, optimize response plan effectiveness, and share learned models across two or more separate system installations.

The invention relates to an automatic selection method for machine learning in cloud computer environment. By using a cloud computing platform, a user can automatically and intelligently build a machine learning mathematic model which meets actual problems without building the operation environment of machine learning, selecting a machine to learn algorithm and even adjusting complicated machine learning functions and accompanying parameters but only using a Web method to upload sample data. Through the method, the use of machine learning is free from the environmental constraints and displays the advantages of the cloud computing platform, so that the machine learning model building is transparent to the user, so as to best reduce the use threshold of machine learning. The automatic selection method for machine learning in cloud computing environment solves the disadvantages of the unpredictability of model building selection, the manual experience of parameter adjustment, the difficulties of common users and the like when machine learning is applied in actual life.

A system, method, and computer program product for assigning an aesthetic score to an image. A method of the present invention includes receiving an image. The method further includes executing a neural network on the image to generate learned features. The method further includes applying a machine-learned model to assign an aesthetic score to the image, where a more aesthetically-pleasing image is given a higher aesthetic score and a less aesthetically-pleasing image is given a lower aesthetic score. The learned features are inputs to the machine-learned model.

A method for creating an interpretable model for healthcare predictions includes training, by a deep learning processor, a neural network to predict health information by providing training data, including multiple combinations of measured or observed health metrics and corresponding medical results, to the neural network. The method also includes determining, by the deep learning processor and using the neural network, prediction data including predicted results for the measured or observed health metrics for each of the multiple combinations of the measured or observed health metrics based on the training data. The method also includes training, by the deep learning processor or a learning processor, an interpretable machine learning model to make similar predictions as the neural network by providing mimic data, including combinations of the measured or observed health metrics and corresponding predicted results of the prediction data, to the interpretable machine learning model.

Mechanical systems wear or change over time. Data collected over a system's life can be input to statistical learning models to predict this wear/change. Previous work by the inventors trained a flexible empirical regression model at a fixed point of wear, and then applied it independently at time points over the life of an engine to predict wear. The embodiment disclosed herein relates those wear predictions over time using a time evolution model. The time evolution model is sequentially updated with new data, and effectively tunes the empirical model for each engine. The combined model predicts wear with dramatically reduced variability. The benefit of reduced variability is that engine wear is more evident, and it is possible to detect operational anomalies more quickly. In addition to tracking wear, the model is also used as the basis for a Bayesian approach to monitor for sudden changes and reject outliers, and adapt the model after these events.

Disclosed herein are systems, methods, and devices for classifying ophthalmic images according to disease type, state, and stage. The disclosed invention details systems, methods, and devices to perform the aforementioned classification based on weighted-linkage of an ensemble of machine learning models. In some parts, each model is trained on a training data set and tested on a test dataset. In other parts, the models are ranked based on classification performance, and model weights are assigned based on model rank. To classify an ophthalmic image, that image is presented to each model of the ensemble for classification, yielding a probabilistic classification score—of each model. Using the model weights, a weighted-average of the individual model-generated probabilistic scores is computed and used for the classification.

Methods and systems are disclosed for learning a regression decision graph model using a Bayesian model selection approach. In a disclosed aspect, the model structure and/or model parameters can be learned using a greedy search algorithm applied to grow the model so long as the model improves. This approach enables construction of a decision graph having a model structure that includes a plurality of leaves, at least one of which includes a non-trivial linear regression. The resulting model thus can be employed for forecasting, such as for time series data, which can include single or multi-step forecasting.

The invention discloses a single-image super-resolution reconstruction method based on a symmetric depth network and belongs to the image processing technology field. The method mainly comprises the following steps of 1, making a high resolution image block and low resolution image block training set; 2, constructing a symmetric convolution-deconvolution depth network used for model training; 3, based on the constructed depth network and the made data set, carrying out network model training; and 4, based on a learned model parameter, inputting one low resolution image, wherein acquired output is a reconstructed high resolution image. In the invention, a convolution layer and a deconvolution layer are combined and simultaneously a network depth is increased; the network depth is used to increase network performance; a reconstruction capability of an image detail portion is enhanced and a good image super-resolution reconstruction effect is acquired. The method has a wide application prospect in fields of image high definition displaying, medical imaging, a remote sensing image and the like.

A computing system uses neural networks to translate natural language queries to database queries. The computing system uses a plurality of machine learning based models, each machine learning model for generating a portion of the database query. The machine learning models use an input representation generated based on terms of the input natural language query, a set of columns of the database schema, and the vocabulary of a database query language, for example, structured query language SQL. The plurality of machine learning based models may include an aggregation classifier model for determining an aggregation operator in the database query, a result column predictor model for determining the result columns of the database query, and a condition clause predictor model for determining the condition clause of the database query. The condition clause predictor is based on reinforcement learning.

An edge computing platform with machine learning capability is provided between a local network with a plurality of sensors and a remote network. A machine learning model is created and trained in the remote network using aggregated sensor data and deployed to the edge platform. Before being deployed, the model is edge-converted (“edge-ified”) to run optimally with the constrained resources of the edge device and with the same or better level of accuracy. The “edge-ified” model is adapted to operate on continuous streams of sensor data in real-time and produce inferences. The inferences can be used to determine actions to take in the local network without communication to the remote network. A closed-loop arrangement between the edge platform and remote network provides for periodically evaluating and iteratively updating the edge-based model.

A method of using data-driven predictive modeling to predict and classify battery cells by lifetime is provided that includes collecting a training dataset by cycling battery cells between a voltage V1 and a voltage V2, continuously measuring battery cell voltage, current, can temperature, and internal resistance during cycling, generating a discharge voltage curve for each cell that is dependent on a discharge capacity for a given cycle, calculating, using data from the discharge voltage curve, a cycle-to-cycle evolution of cell charge to output a cell voltage versus charge curve Q(V), generating transformations of ΔQ(V), generating transformations of data streams that include capacity, temperature and internal resistance, applying a machine learning model to determine a combination of a subset of the transformations to predict cell operation characteristics, and applying the machine learning model to output the predicted battery operation characteristics.

A method and an apparatus of allocating available resources in a cluster system with learning models and tuning methods are provided. The learning model may be trained from historic performance data of previously executed jobs and used to project a suggested amount of resources for execution of a job. The tuning process may suggest a configuration for the projected amount of resources in the cluster system for an optimal operating point. An optimization may be performed with respect to a set of objective functions to improve resource utilization and system performance while suggesting the configuration. Through many executions and job characterization, the learning/tuning process for suggesting the configuration for the projected amount of resources may be improved by understanding correlations of historic data and the objective functions.

The invention relates to a method which extracts the abstracts and key words of a file at the same time, belonging to language words processing technique. The existing method takes the extraction of abstracts of the file and the extraction of the key words of the file as two irrelative tasks and respectively processes the two tasks which have the same nature; the method can utilize the same nature of the extraction and completes the extractions of the abstracts and key words at the same time. The method utilizes a figure learning model and comprehensively utilizes the relationships between sentences in the file, between the sentence and the words in the file, and between the words in the file, exactly evaluates the importance of the sentences and the words, and finally adopts the important sentences and words as the abstracts and key words of the file. The method can extract the abstracts and key words of the file at the same time on the one hand, and can gain a better effect of the extraction of abstracts and key words on the other hand; the method can be widely applied to the fields such as text information processing and digging and the like.

The invention provides a method of constructing a text classification model. The method comprises the steps of: constructing a training sample set according to structure features of characters, wordsand sentences of text information, wherein each piece of sample data in the training sample set corresponds to a feature matrix A of a piece of text information about the words and a feature matrix Babout the characters and a category vector O corresponding to the piece of text information, and the dimension number of O is the same as a category number; and using the feature matrix A about the words and the feature matrix B about the characters in the training matrix set as input and the corresponding category vector O as output to train a deep learning model to obtain the text classificationmodel. Classification is carried out according to the classification model constructed by the method, an accuracy rate of text classification can be improved, and the method is particularly suitablefor use in short-text classification.