Abstract: Methods and systems for generating an end-to-end neural text-to-speech (TTS) model to process an input text to generate speech representations. An annotated set of text documents including annotations inserted therein to indicate prosodic features are input into the TTS model. The TTS model is trained using the annotated dataset and a corresponding dataset of speech representations of the text documents that include prosody associated with the indicated prosodic features. The trained TTS model learns to associate the prosody with the annotations.

Abstract: Methods for training a neural sequence-to-sequence (seq2seq) model. A processor receives the model and training data comprising a plurality of training source sequences and corresponding training target sequences, and generates corresponding predicted target sequences. Model parameters are updated based on a comparison of predicted target sequences to training target sequences to reduce or minimize both a local loss in the predicted target sequences and an expected loss of one or more global or semantic features or constraints between the predicted target sequences and the training target sequences given the training source sequences. Expected loss is based on global or semantic features or constraints of general target sequences given general source sequences.

Abstract: In methods for training a natural language generation (NLG) model using a processor a document-level machine translation (MT) model is provided by training an MT model to receive as input, token sequences in a first language, and to generate as output, token sequences in a second language. An augmented document-level MT model is provided by training the document-level MT model to receive as input, paired language-independent structured data and token sequences in the first language, and to generate as output, token sequences in the second language. The augmented document-level MT model is trained to receive as input, language-independent structured data, and to generate as output, token sequences in the second language.

Abstract: There is disclosed a method for using a computer to enable correction of misclassified labels in a database. The computer initially applies a dataset (including labels pointing respectively to categories) to a first classifier, which includes a first loss function. Pursuant to the initial application, the computer determines that one or more labels have been misclassified. Responsive to such determination, the computer changes the first loss function to a second loss function to form a second classifier including the second loss function. The computer then applies the dataset to the second classifier for enabling correction of the one or more misclassified labels.

Abstract: In methods for training a natural language generation (NLG) model using a processor a document-level machine translation (MT) model is provided by training an MT model to receive as input, token sequences in a first language, and to generate as output, token sequences in a second language. An augmented document-level MT model is provided by training the document-level MT model to receive as input, paired language-independent structured data and token sequences in the first language, and to generate as output, token sequences in the second language. The augmented document-level MT model is trained to receive as input, language-independent structured data, and to generate as output, token sequences in the second language.

Abstract: Methods employing temporal and/or cultural data are provided for automatically assigning one or more semantic tags to a point-of-interest (POI) using a processor. The POI is represented by attribute data. A dataset including temporal attribute data and/or cultural data is provided to a multilabel classifier comprising a neural network model. One or more predicted semantic tags for the POI are received from an output of the multilabel classifier. The predicted semantic tags are stored in a database as additional attribute data of the POI.

Abstract: Methods for training a neural sequence-to-sequence (seq2seq) model. A processor receives the model and training data comprising a plurality of training source sequences and corresponding training target sequences, and generates corresponding predicted target sequences. Model parameters are updated based on a comparison of predicted target sequences to training target sequences to reduce or minimize both a local loss in the predicted target sequences and an expected loss of one or more global or semantic features or constraints between the predicted target sequences and the training target sequences given the training source sequences. Expected loss is based on global or semantic features or constraints of general target sequences given general source sequences.

Abstract: A method for extracting entities from a text document includes, for at least a section of a text document, providing a first set of entities extracted from the at least a section, clustering at least a subset of the extracted entities in the first set into clusters, based on locations of the entities in the document. Complete ones of the clusters of entities are identified. Patterns for extracting new entities are learned based on the complete clusters. New entities are extracted from incomplete clusters based on the learned patterns.

Abstract: A method for extracting entities from a text document includes, for at least a section of a text document, providing a first set of entities extracted from the at least a section, clustering at least a subset of the extracted entities in the first set into clusters, based on locations of the entities in the document. Complete ones of the clusters of entities are identified. Patterns for extracting new entities are learned based on the complete clusters. New entities are extracted from incomplete clusters based on the learned patterns.

Abstract: A method for data matching includes providing two sets of encrypted data elements by converting data elements to respective sets of vectors and encrypting each vector with a public key of a homomorphic encryption scheme. Each data element includes a sequence of characters drawn from an alphabet. For pairs of encrypted data elements, a comparison measure is computed between the sets of encrypted vectors. An obfuscated vector is generated for each encrypted data element in the first set, which renders the first encrypted data element indecipherable when the comparison measure does not meet a threshold for at least one of the pairs of data encrypted elements comprising that encrypted data element. The obfuscated vectors can be decrypted with a private key, allowing data elements in the first set to be deciphered if the comparison measure meets the threshold for at least one of the data elements in the second set.

Abstract: A Virtual Reality (VR) method and system for text manipulation. According to an exemplary embodiment, a VR method displays and provides a user with interaction with displayed documents in a Virtual Environment (VE), the VE including a VR head-mounted display and one or more gesture sensors. Text manipulation is performed using natural human body interactions with the VR system.

Abstract: A system and method for selection of a batch of objects are provided. Each object in a pool is assigned to a subset of a set of buckets. The assignment is based on signatures, generated, for example, by LSH hashing object representations of the objects in the pool. The signatures are then segmented into bands which are each assigned to a respective bucket in the set, based on the elements of the band. An entropy value is computed for each of a set of objects remaining in the pool using a current classifier model. A batch of objects for retraining the model is selected. This includes selecting objects from the set of objects based on their computed entropy values and respective assigned buckets.

Abstract: A method for data matching includes providing two sets of encrypted data elements by converting data elements to respective sets of vectors and encrypting each vector with a public key of a homomorphic encryption scheme. Each data element includes a sequence of characters drawn from an alphabet. For pairs of encrypted data elements, a comparison measure is computed between the sets of encrypted vectors. An obfuscated vector is generated for each encrypted data element in the first set, which renders the first encrypted data element indecipherable when the comparison measure does not meet a threshold for at least one of the pairs of data encrypted elements comprising that encrypted data element. The obfuscated vectors can be decrypted with a private key, allowing data elements in the first set to be deciphered if the comparison measure meets the threshold for at least one of the data elements in the second set.

Abstract: Data privacy is becoming increasingly important and, in some jurisdictions, required. Access to private data can be controlled by forcing all access to go through minimizations services that allow only authorized access to private data. These minimization services can become processing bottlenecks if the only way to modify private data is by way of requests to the minimization service. Certain homomorphic operations allow for encrypted data to be modified without being first decrypted although other operands must be encrypted. Augmenting a minimization service to provide a public encryption key provides for encryption of the other operands. Providing a records manager that can take advantage of homomorphic operations allows certain data operations to be performed without compromising security and without accessing the minimization service.

Abstract: Data privacy is becoming increasingly important and, in some jurisdictions, required. Access to private data can be controlled by forcing all access to go through minimizations services that allow only authorized access to private data. These minimization services can become processing bottlenecks if the only way to modify private data is by way of requests to the minimization service. Certain homomorphic operations allow for encrypted data to be modified without being first decrypted although other operands must be encrypted. Augmenting a minimization service to provide a public encryption key provides for encryption of the other operands. Providing a records manager that can take advantage of homomorphic operations allows certain data operations to be performed without compromising security and without accessing the minimization service.