Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for using embedded function with a deep network. One of the methods includes receiving an input comprising a plurality of features, wherein each of the features is of a different feature type; processing each of the features using a respective embedding function to generate one or more numeric values, wherein each of the embedding functions operates independently of each other embedding function, and wherein each of the embedding functions is used for features of a respective feature type; processing the numeric values using a deep network to generate a first alternative representation of the input, wherein the deep network is a machine learning model composed of a plurality of levels of non-linear operations; and processing the first alternative representation of the input using a logistic regression classifier to predict a label for the input.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for using embedded function with a deep network. One of the methods includes receiving an input comprising a plurality of features, wherein each of the features is of a different feature type; processing each of the features using a respective embedding function to generate one or more numeric values, wherein each of the embedding functions operates independently of each other embedding function, and wherein each of the embedding functions is used for features of a respective feature type; processing the numeric values using a deep network to generate a first alternative representation of the input, wherein the deep network is a machine learning model composed of a plurality of levels of non-linear operations; and processing the first alternative representation of the input using a logistic regression classifier to predict a label for the input.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for using embedded function with a deep network. One of the methods includes receiving an input comprising a plurality of features, wherein each of the features is of a different feature type; processing each of the features using a respective embedding function to generate one or more numeric values, wherein each of the embedding functions operates independently of each other embedding function, and wherein each of the embedding functions is used for features of a respective feature type; processing the numeric values using a deep network to generate a first alternative representation of the input, wherein the deep network is a machine learning model composed of a plurality of levels of non-linear operations; and processing the first alternative representation of the input using a logistic regression classifier to predict a label for the input.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for using embedded function with a deep network. One of the methods includes receiving an input comprising a plurality of features, wherein each of the features is of a different feature type; processing each of the features using a respective embedding function to generate one or more numeric values, wherein each of the embedding functions operates independently of each other embedding function, and wherein each of the embedding functions is used for features of a respective feature type; processing the numeric values using a deep network to generate a first alternative representation of the input, wherein the deep network is a machine learning model composed of a plurality of levels of non-linear operations; and processing the first alternative representation of the input using a logistic regression classifier to predict a label for the input.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for using embedded function with a deep network. One of the methods includes receiving an input comprising a plurality of features, wherein each of the features is of a different feature type; processing each of the features using a respective embedding function to generate one or more numeric values, wherein each of the embedding functions operates independently of each other embedding function, and wherein each of the embedding functions is used for features of a respective feature type; processing the numeric values using a deep network to generate a first alternative representation of the input, wherein the deep network is a machine learning model composed of a plurality of levels of non-linear operations; and processing the first alternative representation of the input using a logistic regression classifier to predict a label for the input.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for using embedded function with a deep network. One of the methods includes receiving an input comprising a plurality of features, wherein each of the features is of a different feature type; processing each of the features using a respective embedding function to generate one or more numeric values, wherein each of the embedding functions operates independently of each other embedding function, and wherein each of the embedding functions is used for features of a respective feature type; processing the numeric values using a deep network to generate a first alternative representation of the input, wherein the deep network is a machine learning model composed of a plurality of levels of non-linear operations; and processing the first alternative representation of the input using a logistic regression classifier to predict a label for the input.

Abstract: Systems, methods, and computer-readable storage media that may be used to provide advertisements to multilingual users are provided. One method includes receiving a search query in the first language entered by the user into a search engine. The method further includes determining a plurality of languages understood by a user based on one or more inputs received from the user. The plurality of languages includes a first language and a second language. The method further includes identifying, using at least one processing circuit, first advertising content in the first language based on the search query. The method further includes identifying, using the at least one processing circuit, second advertising content in the second language based on the search query. The method further includes providing the first advertising content and the second advertising content to the user.

Abstract: A method and apparatus for obtaining, hashing, storing and using fingerprint data related to fingerprint minutia including the steps of: a) determining minutia points within a fingerprint, b) determining a plurality of sets of proximate determined minutia points, c) subjecting a plurality of representations of the determined sets of minutia points to a hashing function, and d) storing or comparing resulting hashed values for fingerprint matching.

Abstract: Method and apparatus for securing biometric data using multiple biometrics. The method includes the steps of: a) converting a first biometric to an irreversibly altered biometric using a parameterized transform function and b) prior to said conversion parameterizing a non-invertible transform function using at least one additional biometric. The apparatus is an apparatus for converting a first biometric to an irreversibly altered biometric using a parameterized non-invertible transform function comprising a computer containing a program for calculating the irreversibly altered biometric based upon input of a first biometric into the parameterized non-invertible transform function.

Abstract: A method and apparatus for obtaining, hashing, storing and using fingerprint data related to fingerprint minutia including the steps of: a) determining minutia points within a fingerprint, b) determining a plurality of sets of proximate determined minutia points, c) subjecting a plurality of representations of the determined sets of minutia points to a hashing function, and d) storing or comparing resulting hashed values for fingerprint matching.

Abstract: The invention provides a method, system, and program product for transforming a multi-dimensional biometric feature point set. More particularly, the invention provides a method for transforming a biometric image using surface folding of the image. In one embodiment, the invention provides a method for transforming a multi-dimensional biometric feature point set, the method comprising: converting the multi-dimensional biometric feature point set to a canonical position and orientation; applying a non-invertible transform function to each of a plurality of points of the biometric feature point set; and providing a transformed biometric feature point set comprising a plurality of transformed points.

Abstract: Method and apparatus for securing biometric data using multiple biometrics. The method includes the steps of: a) converting a first biometric to an irreversibly altered biometric using a parameterized non-convertible transform function and b) prior to said conversion parameterizing a non-convertible transform function using at least one additional biometric. The apparatus is an apparatus for converting a first biometric to an irreversibly altered biometric using a parameterized non-convertible transform function comprising a computer containing a program for calculating the irreversibly altered biometric based upon input of a first biometric into the parameterized non-convertible transform function.