Abstract: A voice command application allows a user to configure an infotainment system to respond to customized voice commands. The voice command application exposes a library of functions to the user which the infotainment system can execute via interaction with the vehicle. The voice command application receives a selection of one or more functions and then receives a speech sample of the voice command. The voice command application generates sample metadata that includes linguistic elements of the voice command, and then generates a command specification. The command specification indicates the selected functions and the sample metadata for storage in a database. Subsequently, the voice command application receives the voice command from the user and locates the associated command specification in the database. The voice command application then extracts the associated set of functions and causes the vehicle to execute those functions to perform vehicle operations.

Abstract: A system or method for manipulating audiovisual data using transcript information. The system or method performs the following actions. Creating a computer-generated transcript of audio data from the audiovisual data, the computer-generated transcript includes a plurality of words, at least some words of the plurality of words are associated with a respective timestamp and a confidence score. Receiving a traditional transcript of the audio data, the traditional transcript includes a plurality of words that are not associated with timestamps. Identifying one or more words from the plurality of words of the computer-generated transcript that match words from the plurality of words of the traditional transcript. Associating the timestamp of the one or more words of the computer-generated transcript with the matching word of the traditional transcript. Processing the audiovisual data using the traditional transcript and the associated timestamps.

Abstract: A computing system is provided. The computing system includes a processor configured to execute one or more programs and associated memory. The processor is configured to execute neural network system that includes a first neural network and a second neural network. The processor is configured to receive input text, and for each of a plurality of text spans within the input text: identify a vector of semantic entities and a vector of entity mentions; define an n-ary relation between entity mentions including subrelations; and determine mention-level representation vectors in the text spans that satisfy the n-ary relation or subrelations. The processor is configured to: aggregate the mention-level representation vectors over all of the text spans to produce entity-level representation vectors; input to the second neural network the entity-level representation vectors; and output a prediction of a presence of the n-ary relation for the semantic entities in the input text.

Abstract: Methods, computer program products, and systems are presented. The methods include, for instance: obtaining an input text for an output speech. The number of words and syllables are counted in each sentence, and a mean sentence length of the input text is calculated. Each sentence length is checked against the mean sentence length and a variation for each sentence is calculated. For the input text, the consumability-readability score is produced as an average of variations for all sentences in the input text. The consumability-readability score indicates the level of satisfaction for the listener of the output speech based on the input text.

Abstract: A computer-implemented method that may be carried out by a reader device or other computing device includes identifying boundaries of words and sentences of displayed content and receiving a language selection of a language for translating a selected unit of text in the displayed content; translating the unit of text to the selected language; retrieving from a picture dictionary, a pictogram associated with a tag applied to the word or at least one word from the sentence of the selected unit of text; generating pronunciation files for both an original language of the unit of text and the selected language; and displaying, in context with the displayed content, the unit of text in the original language and the selected language, the pictogram, and icons for requesting to play the pronunciation files. The selected unit of text can be a word or a sentence.

Abstract: A word vector processing method is provided. Word segmentation is performed on a corpus to obtain words, and n-gram strokes corresponding to the words are determined. Each n-gram stroke represents n successive strokes of a corresponding word. Word vectors of the words and stroke vectors of the n-gram strokes are initialized corresponding to the words. After performing the word segmentation, the n-gram strokes are determined, and the word vectors and stroke vectors are determined, training the word vectors and the stroke vectors.

Abstract: A multicore CPU of a text string data analyzing device: loads a plurality of blocks obtained by dividing a text string dictionary into a memory; executes, in parallel on block groups executable independently of each other, an entry registration process of registering, character by character, unregistered text strings of text string data as new entries in the blocks in order from last characters; and outputs, as BW transformed data of the text string dictionary in which the text string data is already registered, a text string obtained by coupling text strings registered in entries of the blocks in a state in which no unregistered text strings of the blocks exists.

Abstract: A semantic parsing method and a semantic parsing apparatus. The method includes: receiving semantic information; parsing the semantic information to obtain a primary text feature; calculating a similarity degree between the primary text feature and a linguistic data feature; using a linguistic data to which a linguistic data feature most similar to the primary text feature belongs, as a primary parsing text; extracting a feature of the primary parsing text to obtain a secondary text feature; obtaining a service feature matching the secondary text feature; obtaining a weight value associated with the obtained service feature; using a service field to which a service feature having a greatest weight value belongs, as the service field to which the primary parsing text belongs; and parsing the primary parsing text in the service field to which the primary parsing text belongs, to obtain a secondary parsing text.

Abstract: An electronic computing device for providing a response to an audio query where the response is determined to have a public safety impact. The electronic computing device includes a microphone and an electronic processor. The electronic processor is configured to receive audio data including a speech segment and determine a plurality of possible meanings of the speech segment. Each possible meaning is associated with a probability that the possible meaning is a correct meaning for the speech segment and a first possible meaning is associated with a first probability that is higher than a second probability associated with a second possible meaning. The electronic processor is also configured to determine public safety impact context information of the second possible meaning. The electronic processor is further configured to output a second response associated with the second possible meaning without first outputting a first response associated with the first possible meaning.

Abstract: Systems and methods are provided for detecting inaccuracy in a product title, comprising identifying, by running a string algorithm on a title associated with a product, at least one product type associated with the product, predicting, using a machine learning algorithm, at least one product type associated with the product based on the title, detecting an inaccuracy in the title, based on at least one of the identification or the prediction, and outputting, to a remote device, a message indicating that the title comprises the inaccuracy. Running the string algorithm may comprise receiving a set of strings, generating a tree based on the received set of strings, receiving the title, and traversing the generated tree using the title to find a match. Using the machine learning algorithm may comprise identifying words in the title, learning a vector representation for each character n-gram of each word, and summing each character n-gram.

Abstract: A method for term-dependent output of information based on a voice input to a specific group includes the steps of: capturing the voice input; analyzing the captured voice input for the presence of a group-specific key term, associated with the specific group; and on detection of the group-specific key term in the analyzed voice input, outputting the information based on the voice input to the specific group.

Abstract: A voice signal decoding device includes a first decoder, a second decoder, a signal switch, and a noise adder. The first decoder decodes first encoded data encoded by a first encoding method. The second decoder decodes second encoded data encoded by a second encoding method. The second encoded data has a narrower band than a band of the first encoded data. The signal switch switches an output signal of the first decoder and an output signal of the second decoder. The noise adder adds a noise signal to a high-frequency band in the output signal of the second decoder when the signal switch switches an output signal from the output signal of the first decoder to the output signal of the second decoder. The high-frequency band is a band where a signal component is lacking as compared with the output signal of the first decoder.

Abstract: A system for categorizing words into clusters includes a receiver to receive a set of sentences formed by a plurality of words. The set of sentences is indicative of interaction of a user with a virtual assistant. A categorizer categorizes the plurality of words into a first set of clusters by using a first clustering technique, and categorizes the plurality of words into a second set of clusters by using a second clustering technique. A detector detects words that appear in similar clusters after categorization by the first clustering technique and the second clustering technique. Similarity of clusters is based on a nature of words forming the clusters. A generator generates a confidence score for each of the plurality of words based on the detection. The confidence score of a word is indicative of accuracy of the categorization of the word.

Abstract: The present disclosure relates to a method, device and computer readable storage medium for presenting an emotion. The method for presenting the emotion includes obtaining a first emotion presentation instruction, wherein the first emotion presentation instruction includes at least one first emotion presentation modality and at least one emotional style, and the at least one first emotion presentation modality includes a text emotion presentation modality; and presenting an emotion corresponding to one or more of the at least one emotional style according to each of the at least one first emotion presentation modality. The present disclosure can realize text-based multi-modal emotion presentation modes, and thus user experience can be improved.

Abstract: A smartphone-based telephone translation system includes a translation machine that starts a call translation process when the call translation is set to ON. The system detects an incoming call or dials, translates a voice signal to obtain a translation text for the ON call state. The voice signal of the caller is received and recognized. A voice signal is translated to obtain a translation text for the ON call state. The localized voice signal is recognized, using a TTS machine to obtain a reply voice signal and sent to the caller. The system performs voice recognition and text translation on the call voice, translates the language into localized language, and performs the function of calling between different languages in real time, so as to realize a telephone communication between two persons speaking different languages. The system supports communication with a hearing-impaired person or deaf-mute person.

Abstract: The present disclosure discloses a method and apparatus for verifying information. The method includes: acquiring a target combination of question and answer, the target combination of question and answer comprising a question and an answer to the question; sending the question in the target combination of question and answer to a terminal device; receiving to-be-confirmed voice information sent by the terminal device, and recognizing the to-be-confirmed voice information to generate a to-be-confirmed answer; and determining that the to-be-confirmed voice information is verified, if the to-be-confirmed answer matches the answer in the target combination of question and answer.

Abstract: The present disclosure provides a semantic disambiguation method, a semantic disambiguation device, a server and a storage medium. The method includes the following. Semantic analysis is performed on a current dialog message input by a user. It is determined, based on an analysis result, whether a current service field corresponding to the current dialog message is an ambiguous field. When the current service field is the ambiguous field, disambiguation is performed on the current service field based on one or more predetermined historical service fields corresponding to one or more historical dialog messages input by the user.

Abstract: An artificial neural network architecture is provided for processing raw audio waveforms to create speaker representations that are used for text-independent speaker verification and recognition. The artificial neural network architecture includes a strided convolution layer, first and second sequentially connected residual blocks, a transformer layer, and a final fully connected (FC) layer. The strided convolution layer is configured to receive raw audio waveforms from a speaker. The first and the second residual blocks both include multiple convolutional and max pooling layers. The transformer layer is configured to aggregate frame level embeddings to an utterance level embedding. The output of the FC layer creates a speaker representation for the speaker whose raw audio waveforms were inputted into the strided convolution layer.

Abstract: The present invention makes it possible to see an original text string which (i) is contained in a captured image and (ii) has been translated, even after a translation of the original text string is displayed. The text string decoration display control section (12) decorates a part indicating a text string contained in a captured image and causes the decorated part to be displayed. The translation image generating section (13) generates a translation image showing a result of translating the text string into another language. The translation display control section (16) switches between display and non-display of the translation image in accordance with an input carried out by a user.

Abstract: A voice to text model used by a voice-enabled electronic device is dynamically and in a context-sensitive manner updated to facilitate recognition of entities that potentially may be spoken by a user in a voice input directed to the voice-enabled electronic device. The dynamic update to the voice to text model may be performed, for example, based upon processing of a first portion of a voice input, e.g., based upon detection of a particular type of voice action, and may be targeted to facilitate the recognition of entities that may occur in a later portion of the same voice input, e.g., entities that are particularly relevant to one or more parameters associated with a detected type of voice action.