Abstract: A method, system, and computer program product, including generating, using a linear probe, confidence scores through flattened intermediate representations and theoretically-justified weighting of samples during a training of the simple model using the confidence scores of the intermediate representations.

Abstract: Clouds in a satellite image are replaced with a prediction of what was occluded by those clouds. The cloudy portion of the image is interpolated from a series of satellite images taken over time, some of which are cloud-free in the target image's cloudy portion. In some configurations, clouds are removed taking into account each pixel's availability—a measure of certainty that a pixel is cloud-free. Furthermore, these images may have been taken under different amounts of illumination, making it difficult to determine whether a difference between two images is due to a change in illumination or a change to the location. The effect of illumination on each image is removed before interpolating the cloudy portion of the image. In some configurations, removing the effect of illumination also takes into account pixel availability.

Abstract: A computing system measures terrain coverage by: obtaining sample image data representing a multispectral image of a geographic region at a sample resolution; generating, based on the sample image data, an index array of pixels for a subject terrain in which each pixel has an index value that represents a predefined relationship between a first wavelength reflectance and a second wavelength reflectance; providing the index array to a trained calibration model to generate an estimated value based on the index array, the estimated value representing an estimated amount of terrain coverage within the geographic region for the subject terrain; and outputting the estimated value for the subject terrain. The trained calibration model may be trained based on training data representing one or more reference images of one or more training geographic regions containing the subject terrain at a higher resolution than the sample resolution.

Abstract: A computing device is provided, including a processor configured to receive imaging relevance data for a geographic area. The processor may be further configured to generate, based at least in part on the imaging relevance data, image mask instructions specifying a region of interest included in the geographic area. The processor may be further configured to transmit the image mask instructions to a satellite. The processor may be further configured to receive, from the satellite, filtered satellite image data of the region of interest. One or more deprioritized regions of the geographic area outside the region of interest may be excluded from the filtered satellite image data.

Abstract: Examples disclosed herein are related to using a machine learning model to generate image data. One example provides a system, comprising one or more processors, and storage comprising instructions executable by the one or more processors to obtain image data comprising an image with unoccluded features, apply a mask to the unoccluded features in the image to form partial observation training data comprising a masked region that obscures at least a portion of the unoccluded features, and train a machine learning model comprising a generator and a discriminator at least in part by generating image data for the masked region and comparing the image data generated for the masked region to the image with unoccluded features.

Abstract: Examples disclosed herein are related to using a machine learning model to generate image data. One example provides a system, comprising one or more processors, and storage comprising instructions executable by the one or more processors to obtain image data comprising an image with unoccluded features, apply a mask to the unoccluded features in the image to form partial observation training data comprising a masked region that obscures at least a portion of the unoccluded features, and train a machine learning model comprising a generator and a discriminator at least in part by generating image data for the masked region and comparing the image data generated for the masked region to the image with unoccluded features.

Abstract: A method, system, and computer program product, including generating, using a linear probe, confidence scores through flattened intermediate representations and theoretically-justified weighting of samples during a training of the simple model using the confidence scores of the intermediate representations.

Abstract: Techniques disclosed herein include systems and methods for privacy-sensitive training data collection for updating acoustic models of speech recognition systems. In one embodiment, the system locally creates adaptation data from raw audio data. Such adaptation can include derived statistics and/or acoustic model update parameters. The derived statistics and/or updated acoustic model data can then be sent to a speech recognition server or third-party entity. Since the audio data and transcriptions are already processed, the statistics or acoustic model data is devoid of any information that could be human-readable or machine readable such as to enable reconstruction of audio data. Thus, such converted data sent to a server does not include personal or confidential information. Third-party servers can then continually update speech models without storing personal and confidential utterances of users.

Abstract: Techniques disclosed herein include systems and methods for privacy-sensitive training data collection for updating acoustic models of speech recognition systems. In one embodiment, the system locally creates adaptation data from raw audio data. Such adaptation can include derived statistics and/or acoustic model update parameters. The derived statistics and/or updated acoustic model data can then be sent to a speech recognition server or third-party entity. Since the audio data and transcriptions are already processed, the statistics or acoustic model data is devoid of any information that could be human-readable or machine readable such as to enable reconstruction of audio data. Thus, such converted data sent to a server does not include personal or confidential information. Third-party servers can then continually update speech models without storing personal and confidential utterances of users.

Abstract: Techniques disclosed herein include systems and methods for privacy-sensitive training data collection for updating acoustic models of speech recognition systems. In one embodiment, the system locally creates adaptation data from raw audio data. Such adaptation can include derived statistics and/or acoustic model update parameters. The derived statistics and/or updated acoustic model data can then be sent to a speech recognition server or third-party entity. Since the audio data and transcriptions are already processed, the statistics or acoustic model data is devoid of any information that could be human-readable or machine readable such as to enable reconstruction of audio data. Thus, such converted data sent to a server does not include personal or confidential information. Third-party servers can then continually update speech models without storing personal and confidential utterances of users.

Abstract: Methods for compressing a transform associated with a feature space are presented. For example, a method for compressing a transform associated with a feature space includes obtaining the transform including a plurality of transform parameters, assigning each of a plurality of quantization levels for the plurality of transform parameters to one of a plurality of quantization values, and assigning each of the plurality of transform parameters to one of the plurality of quantization values to which one of the plurality of quantization levels is assigned. One or more of obtaining the transform, assigning of each of the plurality of quantization levels, and assigning of each of the transform parameters are implemented as instruction code executed on a processor device. Further, a Viterbi algorithm may be employed for use in non-uniform level/value assignments.

Abstract: Methods for compressing a transform associated with a feature space are presented. For example, a method for compressing a transform associated with a feature space includes obtaining the transform including a plurality of transform parameters, assigning each of a plurality of quantization levels for the plurality of transform parameters to one of a plurality of quantization values, and assigning each of the plurality of transform parameters to one of the plurality of quantization values to which one of the plurality of quantization levels is assigned. One or more of obtaining the transform, assigning of each of the plurality of quantization levels, and assigning of each of the transform parameters are implemented as instruction code executed on a processor device. Further, a Viterbi algorithm may be employed for use in non-uniform level/value assignments.

Abstract: A method, and a system to execute this method is being presented for the identification and separation of sources of an acoustic signal, which signal contains a mixture of multiple simultaneous component signals. The method represents the signal with multiple discrete state-variable sequences and combines acoustic and context level dynamics to achieve the source separation. The method identifies sources by discovering those frames of the signal whose features are dominated by single sources. The signal may be the simultaneous speech of multiple speakers.

Abstract: In an Automatic Speech Recognition (ASR) system having at least two language models, a method is provided for combining language model scores generated by at least two language models. A list of most likely words is generated for a current word in a word sequence uttered by a speaker, and acoustic scores corresponding to the most likely words are also generated. Language model scores are computed for each of the most likely words in the list, for each of the at least two language models. A set of coefficients to be used to combine the language model scores of each of the most likely words in the list is respectively and dynamically determined, based on a context of the current word. The language model scores of each of the most likely words in the list are respectively combined to obtain a composite score for each of the most likely words in the list, using the set of coefficients determined therefor.

Abstract: A method, and a system to execute this method is being presented for the identification and separation of sources of an acoustic signal, which signal contains a mixture of multiple simultaneous component signals. The method represents the signal with multiple discrete state-variable sequences and combines acoustic and context level dynamics to achieve the source separation. The method identifies sources by discovering those frames of the signal whose features are dominated by single sources. The signal may be the simultaneous speech of multiple speakers.

Abstract: Two statistics are disclosed for determining the quality of language models. These statistics are called acoustic perplexity and the synthetic acoustic word error rate (SAWER), and they depend upon methods for computing the acoustic confusability of words. It is possible to substitute models of acoustic data in place of real acoustic data in order to determine acoustic confusability. An evaluation model is created, a synthesizer model is created, and a matrix is determined from the evaluation and synthesizer models. Each of the evaluation and synthesizer models is a hidden Markov model. Once the matrix is determined, a confusability calculation may be performed. Different methods are used to determine synthetic likelihoods. The confusability may be normalized and smoothed and methods are disclosed that increase the speed of performing the matrix inversion and the confusability calculation. A method for caching and reusing computations for similar words is disclosed.

Abstract: A characteristic-specific digitization method and apparatus are disclosed that reduces the error rate in converting input information into a computer-readable format. The input information is analyzed and subsets of the input information are classified according to whether the input information exhibits a specific physical parameter affecting recognition accuracy. If the input information exhibits the specific physical parameter affecting recognition accuracy, the characteristic-specific digitization system recognizes the input information using a characteristic-specific recognizer that demonstrates improved performance for the given physical parameter. If the input information does not exhibit the specific physical parameter affecting recognition accuracy, the characteristic-specific digitization system recognizes the input information using a general recognizer that performs well for typical input information.

Abstract: There is provided a method for augmenting an alternate word list generated by a speech recognition system. The alternate word list includes at least one potentially correct word for replacing a wrongly decoded word. The method includes the step of identifying at least one acoustically confusable word with respect to the wrongly decoded word. The alternate word list is augmented with the at least one acoustically confusable word.

Abstract: A characteristic-specific digitization method and apparatus are disclosed that reduces the error rate in converting input information into a computer-readable format. The input information is analyzed and subsets of the input information are classified according to whether the input information exhibits a specific physical parameter affecting recognition accuracy. If the input information exhibits the specific physical parameter affecting recognition accuracy, the characteristic-specific digitization system recognizes the input information using a characteristic-specific recognizer that demonstrates improved performance for the given physical parameter. If the input information does not exhibit the specific physical parameter affecting recognition accuracy, the characteristic-specific digitization system recognizes the input information using a general recognizer that performs well for typical input information.

Abstract: There is provided a method for augmenting an alternate word list generated by a speech recognition system. The alternate word list includes at least one potentially correct word for replacing a wrongly decoded word. The method includes the step of identifying at least one acoustically confusable word with respect to the wrongly decoded word. The alternate word list is augmented with the at least one acoustically confusable word.