Abstract: Systems, methods, and non-transitory computer readable media can align face images, classify face images, and verify face images by employing a deep neural network (DNN). A 3D-aligned face image can be generated from a 2D face image. An identity of the 2D face image can be classified based on provision of the 3D-aligned face image to the DNN. The identity of the 2D face image can comprise a feature vector.

Abstract: A system for training a model to predict a sequence (e.g. a sequence of words) given a context is disclosed. A model can be trained to make these predictions using a combination of individual predictions compared to base truth and sequences of predictions based on previous predictions, where the resulting sequence is compared to the base truth sequence. In particular, the model can initially use the individual predictions to train the model. The model can then be further trained over the training data in multiple iterations, where each iteration includes two processes for each training element. In the first process, an initial part of the sequence is predicted, and the model and model parameters are updated after each prediction. In the second process, the entire remaining amount of the sequence is predicted and compared to the corresponding training sequence to adjust model parameters to encourage or discourage each prediction.

Abstract: Systems, methods, and non-transitory computer readable media can align face images, classify face images, and verify face images by employing a deep neural network (DNN). A 3D-aligned face image can be generated from a 2D face image. An identity of the 2D face image can be classified based on provision of the 3D-aligned face image to the DNN. The identity of the 2D face image can comprise a feature vector.

Abstract: Systems, methods, and non-transitory computer readable media can align face images, classify face images, and verify face images by employing a deep neural network (DNN). A 3D-aligned face image can be generated from a 2D face image. An identity of the 2D face image can be classified based on provision of the 3D-aligned face image to the DNN. The identity of the 2D face image can comprise a feature vector.

Abstract: A system for training a model to predict a sequence (e.g. a sequence of words) given a context is disclosed. A model can be trained to make these predictions using a combination of individual predictions compared to base truth and sequences of predictions based on previous predictions, where the resulting sequence is compared to the base truth sequence. In particular, the model can initially use the individual predictions to train the model. The model can then be further trained over the training data in multiple iterations, where each iteration includes two processes for each training element. In the first process, an initial part of the sequence is predicted, and the model and model parameters are updated after each prediction. In the second process, the entire remaining amount of the sequence is predicted and compared to the corresponding training sequence to adjust model parameters to encourage or discourage each prediction.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating labeled images. One of the methods includes selecting a plurality of candidate videos from videos identified in a response to a search query derived from a label for an object category; selecting one or more initial frames from each of the candidate videos; detecting one or more initial images of objects in the object category in the initial frames; for each initial frame including an initial image of an object in the object category, tracking the object through surrounding frames to identify additional images of the object; and selecting one or more images from the one or more initial images and one or more additional images as database images of objects belonging to the object category.

Abstract: A system and method for improved classification. A first classifier is trained using a first process running on at least one computing device using a first set of training images relating to a class of images. A set of additional images are selected using the first classifier from a source of additional images accessible to the computing device. The first set of training images and the set of additional images are merged using the computing device to create a second set of training images. A second classifier is trained using a second process running on the computing device using the second set of training images. A set of unclassified images are classified using the second classifier thereby creating a set of classified images. The first classifier and the second classifier employ different classification methods.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for identifying objects in images. One of the methods includes receiving an input image; down-sampling the input image to generate a second image; generating a respective first score for each of the plurality of object categories; selecting an initial patch of the input image; generating a respective second score for each of the plurality of object categories; and generating a respective third score for each of the plurality of object categories from the first scores and the second scores, wherein the respective third score for each of the plurality of object categories represents a likelihood that the input image contains an image of an object belonging to the object category.

Abstract: Methods and systems for processing multilingual DNN acoustic models are described. An example method may include receiving training data that includes a respective training data set for each of two or more or languages. A multilingual deep neural network (DNN) acoustic model may be processed based on the training data. The multilingual DNN acoustic model may include a feedforward neural network having multiple layers of one or more nodes. Each node of a given layer may connect with a respective weight to each node of a subsequent layer, and the multiple layers of one or more nodes may include one or more shared hidden layers of nodes and a language-specific output layer of nodes corresponding to each of the two or more languages. Additionally, weights associated with the multiple layers of one or more nodes of the processed multilingual DNN acoustic model may be stored in a database.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating labeled images. One of the methods includes selecting a plurality of candidate videos from videos identified in a response to a search query derived from a label for an object category; selecting one or more initial frames from each of the candidate videos; detecting one or more initial images of objects in the object category in the initial frames; for each initial frame including an initial image of an object in the object category, tracking the object through surrounding frames to identify additional images of the object; and selecting one or more images from the one or more initial images and one or more additional images as database images of objects belonging to the object category.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for identifying objects in images. One of the methods includes receiving an input image; down-sampling the input image to generate a second image; generating a respective first score for each of the plurality of object categories; selecting an initial patch of the input image; generating a respective second score for each of the plurality of object categories; and generating a respective third score for each of the plurality of object categories from the first scores and the second scores, wherein the respective third score for each of the plurality of object categories represents a likelihood that the input image contains an image of an object belonging to the object category.

Abstract: Methods and apparatus related to training speech recognition devices are presented. A computing device receives training samples for training a neural network to learn an acoustic speech model. A curriculum function for speech modeling can be determined. For each training sample of the training samples, a corresponding curriculum function value for the training sample can be determined using the curriculum function. The training samples can be ordered based on the corresponding curriculum function values. In some embodiments, the neural network can be trained utilizing the ordered training samples. The trained neural network can receive an input of a second plurality of samples corresponding to human speech, where the second plurality of samples differs from the training samples. In response to receiving the second plurality of samples, the trained neural network can generate a plurality of phones corresponding to the captured human speech.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for identifying objects in images. One of the methods includes obtaining a first training image; down-sampling the first training image to generate a low-resolution first training image; processing the low-resolution first training image using a first neural network to generate a plurality of features of the low-resolution first training image and first scores for the low-resolution first training image; processing the first scores and the features of the low-resolution first training image using an initial patch locator neural network to generate an initial location of an initial patch of the first training image; locally perturbing the initial location to select an adjusted location for the initial patch of the first training image; and updating the current values of the parameters of the initial patch locator neural network to generate updated values using the adjusted location.

Abstract: Systems, methods, and non-transitory computer readable media can align face images, classify face images, and verify face images by employing a deep neural network (DNN). A 3D-aligned face image can be generated from a 2D face image. An identity of the 2D face image can be classified based on provision of the 3D-aligned face image to the DNN. The identity of the 2D face image can comprise a feature vector.

Abstract: A system and method for improved classification. A first classifier is trained using a first process running on at least one computing device using a first set of training images relating to a class of images. A set of additional images are selected using the first classifier from a source of additional images accessible to the computing device. The first set of training images and the set of additional images are merged using the computing device to create a second set of training images. A second classifier is trained using a second process running on the computing device using the second set of training images. A set of unclassified images are classified using the second classifier thereby creating a set of classified images. The first classifier and the second classifier employ different classification methods.

Abstract: A method and system provide the ability to automatically recognize biological particles. An image of biological particles (e.g., airborne pollen or urine) is obtained. One or more parts of the image are detected as containing one or more particles of interest. Feature vector(s) are extracted from each detected part of the image. Non-linearities are applied to each feature vector. Each part of the image is then classified into a category of biological particle based on the one or more feature vectors for each part of the image.