Abstract: A system and a method including accessing a semantic model having a grammar syntax supporting an expression including a combination of at least one of a semantic model class and an instance of a semantic model class; specifying at least one missing property specification for at least one of a semantic model class and an instance of a semantic model class, including constraints for a use of the at least one missing property; adding the at least one missing property specification to the semantic model, the at least one missing property specification added to the semantic model to be referenced in a validation and translation of an expression including a combination of at least one of the class and the instance of a class having one of the at least one missing property specified therewith; and generating a record of the semantic model including the at least one missing property specification added thereto.

Abstract: A computer-implemented method for automatic item generation in an digital examination that includes obtaining a first string of text, the first string of text being organized based on a linguistic context, selecting a first textual term from the first string of text, identifying a first set of linguistic characteristics corresponding to the first textual term, obtaining a second textual term exhibiting at least one linguistic characteristic that is different from any linguistic characteristic of the first set of linguistic characteristics, and generating a second string of text by replacing the first textual term in the first string of text with the second textual term.

Abstract: A device receives a corpus of text documents, and utilizes feature extraction on a text document, of the corpus of text documents, to generate features from the text document, where the features include binary features, numeric features, and categorical features. The device performs feature engineering on one or more of the binary features, the numeric features, or the categorical features, to generate converted features, and performs feature encoding on the text document, based on the converted features, to represent the text document as a vector with a similarity score for a domain. The device provides the vector with the similarity score for the domain, as training data, to a machine learning model to generate a trained machine learning model, and performs an action using the trained machine learning model.

Abstract: A device includes one or more processor devices, an audio sensor, a speaker device, and one or more memory devices. The audio sensor is configured to generate an audio stream. The one or more memory devices store data including data identifying a wake phrase and instructions. The instructions are executable by the one or more processor devices to initiate, as a result of detecting the wake phrase in the audio stream and determining that the wake phrase is invalid, presentation via the speaker device of audio output identifying a valid wake input. The instructions are further executable by the one or more processors to initiate, as a result of detecting the valid wake input, communication of a portion of the audio stream to a natural language processing service.

Abstract: An automatic speech recognition (ASR) system includes a memory configured to store a filler model. The filler model includes one or more phonetic strings corresponding to one or more portions of a wake up phrase. The ASR system also includes one or more processors operatively coupled to the memory and configured to analyze a speech signal with the filler model to determine whether the speech signal includes the wake up phrase or any portion of the wake up phrase. The one or more processors are also configured to generate, based on the analysis, a hypothesis of underlying speech included in the speech signal. The hypothesis excludes the wake up phrase or any portion of the wake up phrase included in the speech signal.

Abstract: Disclosed is a method for generating an utterance of a chatbot based on a semantic graph database and executed by a computer, includes generating the semantic graph database by using a past utterance of a user, and generating the utterance of the chatbot for responding to a new utterance of the user, by using the semantic graph database.

Abstract: One embodiment provides a method that includes obtaining a default language corpus. A second language corpus is obtained based on a second language preference. A first transcription of an utterance is received using the default language corpus and natural language processing (NLP). At least one problem word in the first transcription is determined based on an associated grammatical relevance to neighboring words in the first transcription. Upon determining that a first probability score is below a first threshold, an acoustic lookup is performed for an audible match for the problem word in the first transcription based on an associated acoustical relevance. Upon determining that a second probability score is below a second threshold, it is determined whether a match for the problem word exists in the secondary language corpus. Upon determining that the match exists in the secondary language corpus, a second transcription for the utterance is provided.

Abstract: Techniques described herein relate to facilitating end-to-end multilingual communications with automated assistants. In various implementations, speech recognition output may be generated based on voice input in a first language. A first language intent may be identified based on the speech recognition output and fulfilled in order to generate a first natural language output candidate in the first language. At least part of the speech recognition output may be translated to a second language to generate an at least partial translation, which may then be used to identify a second language intent that is fulfilled to generate a second natural language output candidate in the second language. Scores may be determined for the first and second natural language output candidates, and based on the scores, a natural language output may be selected for presentation.

Abstract: Generating, updating, and using a knowledge graph. Concepts in a knowledge graph can have relations to one another. These relations may be expressed as confidence values. A training data set may be split into two portions, with the first portion used to update confidence values for existing relations between concept pairs, using the knowledge graph. These confidence values can be used, together with the second portion used to update confidence values for known phrases that express known relations. These confidence values, in turn, can be used, together with the first portion, to increase the accuracy of the original confidence scores with respect to existing relations. The process may be iteratively employed, with each iteration increasing the accuracy of confidence scores.

Abstract: An apparatus for decoding an encoded signal includes: an audio decoder for decoding an encoded representation of a first set of first spectral portions to obtain a decoded first set of first spectral portions; a parametric decoder for decoding an encoded parametric representation of a second set of second spectral portions to obtain a decoded representation of the parametric representation, wherein the parametric information includes, for each target frequency tile, a source region identification as a matching information; and a frequency regenerator for regenerating a target frequency tile using a source region from the first set of first spectral portions identified by the matching information.

Abstract: A system, computer-readable medium, and a method including accessing a semantic model, the semantic model having a grammar syntax supporting a comparison of at least one of a semantic model class and an instance of a semantic model class to an incompatible value; specifying an implied property for at least one of a semantic model class and an instance of a semantic model class in the semantic model, the implied property having an associated range; adding the implied property to the semantic model, the implied property to be referenced in a validation and translation of an expression including at least one of the class and the instance of a class having the implied property specified therewith compared to an incompatible value; and generating a record of the semantic model including the implied property added thereto.

Abstract: Various embodiments may be generally directed to the use of an adversarial learning framework for persona-based dialogue modeling. In some embodiments, automated multi-turn dialogue response generation may be performed using a persona-based hierarchical recurrent encoder-decoder-based generative adversarial network (phredGAN). Such a phredGAN may feature a persona-based hierarchical recurrent encoder-decoder (PHRED) generator and a conditional discriminator. In some embodiments, the conditional discriminator may include an adversarial discriminator that is provided with attribute representations as inputs. In some other embodiments, the conditional discriminator may include an attribute discriminator, and attribute representations may be handled as targets of the attribute discriminator. The embodiments are not limited in this context.

Abstract: A method, an apparatus, and a system for speech recognition are provided. A third-party application corresponding to a speech signal of a user can be determined according to the speech signal and by means of semantic analysis; and third-party application registry information is searched for and a third-party program is started, so that the user does not need to tap the third-party application to start the corresponding program, thereby providing more intelligent service for the user and facilitating use for the user.

Abstract: Techniques for performing command processing are described. A system receives, from a device, input data corresponding to a command. The input data may originate as audio data, as text data, or as other data. The system determines NLU processing results corresponding to the input data. The NLU processing results may be associated with multiple speechlets. The system also determines NLU confidences for the NLU processing results for each speechlet. The system sends NLU processing results and an indication to provide potential results to a portion of the multiple speechlets, and receives potential results from the portion of the speechlets. The system also receives indications whether the speechlets need to be re-called if the speechlets are selected to execute with respect to the command. The system ranks the portion of the speechlets based at least in part on the NLU processing results as well as the potential results provided by the portion of the speechlets.

Abstract: The invention relates to a control device with voice and/or gestural recognition for the interior lighting of a motor vehicle comprising a plurality of lighting modules (Mn) each capable of emitting at least one light beam in an interior of the vehicle. The device includes at least distinct means (Cn) for detecting a plurality of voice and/or gestural commands (Sn), executed by at least one occupant of the vehicle in specific zones (Zn) of the vehicle interior targeted by the detection means, the detection means (Cn) being capable of transmitting data representative of the commands (Sn) to at least processing and analysis means (ECU) configured to merge and prioritize the data and generate, after comparison with reference data, at least one instruction (An) to at least one interior lighting module (Mn).

Abstract: Systems and methods for building a response for a machine learning-based dialogue agent includes implementing machine learning classifiers that predict slot segments of the utterance data based on an input of the utterance data; predict a slot classification label for each of the slot segments of the utterance data; computing a semantic vector value for each of the slot segments of the utterance data; assessing the semantic vector value of the slot segments of the utterance data against a multi-dimensional vector space of structured categories of dialogue, wherein the assessment includes: for each of a distinct structured categories of dialogue computing a similarity metric value; selecting one structured category of dialogue from the distinct structured categories of dialogue based on the computed similarity metric value for each of distinct structured categories; and producing a response to the utterance data.

Abstract: Systems, methods and apparatus for authoring an electronic message. In one example, the system, method, and apparatus include receiving an electronic message; categorizing the electronic message based on a message layout type associated with the electronic message; determining, with a sentiment analyzer, a sentiment type and a sentiment score associated with the electronic message; identifying a first plurality of words in the electronic message associated with the sentiment type; and outputting one or more selected from the group consisting of the message layout type, the sentiment type, and the sentiment score associated with the electronic message.

Abstract: Techniques for mapping policy documents to regulatory documents to check for compliance between the policies and documents are provided. In one example, a computer-implemented method determining, by a system operatively coupled to a processor, an information input, a control framework, and a document from a first group consisting of a regulatory document and a policy document, wherein the information input is a corpora from a second group consisting of a domain corpora and a global corpora. The computer-implemented method can also comprise mapping, by the system, the received regulatory document or the received policy document to the control framework using a supervised machine learning technique.

Abstract: Techniques are provided for training, by a system operatively coupled to a processor, an attention weighted recurrent neural network encoder-decoder (AWRNNED) using an iterative process based on one or more paragraphs of agent sentences from respective transcripts of one or more conversations between one or more agents and one or more customers, and based on one or more customer response sentences from the respective transcripts, and generating, by the system, one or more groups respectively comprising one or more agent sentences and one or more customer response sentences selected based on attention weights of the AWRNNED.

Abstract: Generating, updating, and using a knowledge graph. Concepts in a knowledge graph can have relations to one another. These relations may be expressed as confidence values. A training data set may be split into two portions, with the first portion used to update confidence values for existing relations between concept pairs, using the knowledge graph. These confidence values can be used, together with the second portion used to update confidence values for known phrases that express known relations. These confidence values, in turn, can be used, together with the first portion, to increase the accuracy of the original confidence scores with respect to existing relations. The process may be iteratively employed, with each iteration increasing the accuracy of confidence scores.