Abstract: The present disclosure provides systems and methods that enable distance metric learning using proxies. A machine-learned distance model can be trained in a proxy space in which a loss function compares an embedding provided for an anchor data point of a training dataset to a positive proxy and one or more negative proxies, where each of the positive proxy and the one or more negative proxies serve as a proxy for two or more data points included in the training dataset. Thus, each proxy can approximate a number of data points, enabling faster convergence. According to another aspect, the proxies of the proxy space can themselves be learned parameters, such that the proxies and the model are trained jointly. Thus, the present disclosure enables faster convergence (e.g., reduced training time). The present disclosure provides example experiments which demonstrate a new state of the art on several popular training datasets.

Abstract: The present disclosure provides systems and methods that enable distance metric learning using proxies. A machine-learned distance model can be trained in a proxy space in which a loss function compares an embedding provided for an anchor data point of a training dataset to a positive proxy and one or more negative proxies, where each of the positive proxy and the one or more negative proxies serve as a proxy for two or more data points included in the training dataset. Thus, each proxy can approximate a number of data points, enabling faster convergence. According to another aspect, the proxies of the proxy space can themselves be learned parameters, such that the proxies and the model are trained jointly. Thus, the present disclosure enables faster convergence (e.g., reduced training time). The present disclosure provides example experiments which demonstrate a new state of the art on several popular training datasets.

Abstract: The present disclosure provides systems and methods that enable distance metric learning using proxies. A machine-learned distance model can be trained in a proxy space in which a loss function compares an embedding provided for an anchor data point of a training dataset to a positive proxy and one or more negative proxies, where each of the positive proxy and the one or more negative proxies serve as a proxy for two or more data points included in the training dataset. Thus, each proxy can approximate a number of data points, enabling faster convergence. According to another aspect, the proxies of the proxy space can themselves be learned parameters, such that the proxies and the model are trained jointly. Thus, the present disclosure enables faster convergence (e.g., reduced training time). The present disclosure provides example experiments which demonstrate a new state of the art on several popular training datasets.

Abstract: In accordance with embodiments of the present disclosure, systems and methods for downsampling DAS data in a way that enables accurate interpretation of acoustic events occurring in the data are provided. Such methods may be particularly useful when interpreting large sets of data, such as DAS VSP data collected during hydrocarbon recovery operations. The methods generally involve identifying data channels affected by noise from a DAS data set, and then interpolating from the surrounding data. This may improve the quality of the resulting downsampled data, with respect to the signal to noise ratio, compared to what would have occurred by merely decimating unwanted data channels. In addition, a priori information about channel fading, the desired downsampling rate, and the slowest expected elastic waves may be used to filter the DAS data. This may achieve a higher signal-to-noise ratio in the downsampled data.

Abstract: In accordance with embodiments of the present disclosure, systems and methods for downsampling DAS data in a way that enables accurate interpretation of acoustic events occurring in the data are provided. Such methods may be particularly useful when interpreting large sets of data, such as DAS VSP data collected during hydrocarbon recovery operations. The methods generally involve identifying data channels affected by noise from a DAS data set, and then interpolating from the surrounding data. This may improve the quality of the resulting downsampled data, with respect to the signal to noise ratio, compared to what would have occurred by merely decimating unwanted data channels. In addition, a priori information about channel fading, the desired downsampling rate, and the slowest expected elastic waves may be used to filter the DAS data. This may achieve a higher signal-to-noise ratio in the downsampled data.

Abstract: Methods and an apparatuses for automatically recognizing and/or verifying objects in a digital image are presented. In one example, a method automatically recognizes objects in digital image data by detecting an object of interest in input digital image data, obtaining a normalized object of interest, assigning texton representations of the normalized object of interest to produce a first a texton array, and determining a similarity between the texton representations and previously determined texton representations of at least one other object. In another example, an apparatus for automatically recognizing objects in digital image data is presented, which includes an image processing control-operably coupled to memory and functional processing units for controlling recognition processing, where the functional processing units further include an object detection unit, a normalizing unit, a texton generation unit, and a similarity unit.

Abstract: A method of knowledge-based engineering design of an instrument panel for a vehicle includes the steps of defining a parameter of the instrument panel using a knowledge-based engineering library stored in a memory of a computer system, generating a model of the instrument panel based on the parameter and analyzing the model of the instrument panel. The method also includes the steps of comparing a result of the analysis of the model of the instrument panel to a predetermined criteria from the knowledge-based engineering library, and varying the parameter so that the model of the instrument panel meets the predetermined criteria.