Abstract: The present invention relates to the field of intelligent recognition, and discloses an intent recognition method based on a deep learning network, resolving a technical problem that accuracy of intent recognition is not high.

Abstract: An audio signal encoding method is provided that comprises collecting audio signal samples, determining sinusoidal components in subsequent frames, estimating amplitudes and frequencies of the components for each frame, merging the obtained pairs into sinusoidal trajectories, splitting particular trajectories into segments, transforming particular trajectories to the frequency domain by way of a digital transform performed on segments longer than the frame duration, quantization and selection of transform coefficients in the segments, entropy encoding, outputting the quantized coefficients as output data, wherein segments of different trajectories starting within a particular time are grouped into Groups of Segments, and the partitioning of trajectories into segments is synchronized with the endpoints of a Group of Segments.

Abstract: Systems and processes for operating an electronic device to train a machine-learning translation system are described. In one process, a first set of training data is obtained. The first set of training data includes at least one payload in a first language and a translation of the at least one payload in a second language. The process further includes obtaining one or more templates for adapting the at least one payload; adapting the at least one payload using the one or more templates to generate at least one adapted payload formulated as a translation request; generating a second set of training data based on the at least one adapted payload; and training the machine-learning translation system using the second set of training data.

Abstract: A speech recognition device includes: an acoustic model based on an End-to-End neural network responsive to an observed sequence formed of prescribed acoustic features obtained from a speech signal by feature extracting unit, for calculating probability of the observed sequence being a certain symbol sequence; and a decoder responsive to a symbol sequence candidate, for decoding a speech signal by a WFST based on a posterior probability of each of word sequences corresponding to the symbol sequence candidate, probabilities calculated by the acoustic model for symbol sequences selected based on an observed sequence, and a posterior probability of each of the plurality of symbol sequences.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for hotword detection on multiple devices are disclosed. In one aspect, a method includes the actions of receiving, by a first computing device, audio data that corresponds to an utterance. The actions further include determining a first value corresponding to a likelihood that the utterance includes a hotword. The actions further include receiving a second value corresponding to a likelihood that the utterance includes the hotword, the second value being determined by a second computing device. The actions further include comparing the first value and the second value. The actions further include based on comparing the first value to the second value, initiating speech recognition processing on the audio data.

Abstract: An information processing apparatus accepts information corresponding to a text. The information processing apparatus refers to a storage unit that stores therein co-occurrence information on other texts with respect to the text and information corresponding to the other texts by associating both the information with the text. The information processing apparatus specifies, from among the pieces of information corresponding to the other texts, the text associated with the information corresponding to the other texts that is associated with the co-occurrence information that meets the standard.

Abstract: Implementations relate to determining a language for speech recognition of a spoken utterance, received via an automated assistant interface, for interacting with an automated assistant. Implementations can enable multilingual interaction with the automated assistant, without necessitating a user explicitly designate a language to be utilized for each interaction. Selection of a speech recognition model for a particular language can based on one or more interaction characteristics exhibited during a dialog session between a user and an automated assistant. Such interaction characteristics can include anticipated user input types, anticipated user input durations, a duration for monitoring for a user response, and/or an actual duration of a provided user response.

Abstract: System and method for controlling a home assistant device include: receiving an audio input; performing speech-to-text conversion on the audio input to obtain a textual string; searching a name list of multiple home devices to determine whether the textual string includes a respective alias of at least one home device in the name list; in accordance with a determination that the textual string includes a first alias corresponding to a first home device of a first device type, generating a first actionable intent to operate the first home device of the first device type; and in accordance with a determination that the textual string includes a second alias corresponding to a second home device of the first device type that is distinct from the first home device, generating a second actionable intent to operate the second home device of the first device type.

Abstract: A greeting card having an audio message recording and playback device permits recording of personalized audio messages to be played upon opening of the greeting card. The recording device is operable in either a trial mode or a use mode. In the trial mode, which would be applicable when the card is displayed in a store, a potential purchaser may experience the functionality of the card by recording their own test message. The test message is played back initially for the potential purchaser but is not subsequently played back to be later heard by other potential purchasers. In the use mode, which the card may be switched to after purchase by the giver of the greeting card, a user recorded message may be played back repeatedly upon subsequent openings of the card. The user recorded message may be followed by a prerecorded recording, such as a song. Additional prerecorded messages, such as voice prompts with instructions for recording a message, may also be included for activation in the trial mode.

Abstract: A natural language understanding (NLU) system that uses a reduced dimensionality of word embedding features to configure compressed NLU models that use reduced computing resources for NLU tasks. A modified NLU model may include a compressed vocabulary data structure of word embedding data vectors that include a set of values corresponding to a reduced dimensionality of the original word embedding features, resulting in a smaller sized vocabulary data structure and reduced size of the vocabulary data structure. Further components of the modified NLU model perform matrix operations to expand the dimensionality of the reduced word embedding data vectors up to the expected dimensionality of later layers of the NLU model. Additional training and reweighting can adjust for potential loses in performance resulting from reductions in the word embedding features. Thus the modified NLU model can achieve similar performance to an original NLU model with reductions in use of computing resources.

Abstract: A voice assistant system includes a server apparatus performing voice assistant and a plurality of devices, in which the server apparatus and the devices are communicatively connected to each other. The plurality of devices each records the same user's speech through a microphone, and then transmits recorded data of the same user's speech to the server apparatus. The server apparatus receives the recorded data transmitted from each of the plurality of devices, and then voice-recognizes two or more of the received recorded data in accordance with a predetermined standard to thereby interpret the contents of the user's speech to perform the voice assistant.

Abstract: One embodiment provides a method, including: receiving, at an information handling device, an audible command to perform a function; determining, using a processor, at least one aspect associated with the audible command that prevents performance of the function; and providing, based on the determining, a suggested modification to the audible command. Other aspects are described and claimed.

Abstract: A computing device includes a display configured to present a graphical user interface. The graphical user interface includes a transcript portion configured to display an unannotated transcript representing an ordered sequence of one or more dialogue events involving a client and a computerized assistant, at least one of the dialogue events taking the form of an example client utterance, and an annotation portion configured to display a hierarchical menu including a plurality of candidate utterance annotations. An utterance annotation machine is configured to receive one or more computer inputs selecting, for each of one or more response parameters in the example client utterance, utterance annotations from the hierarchical menu that collectively define a machine-readable interpretation of the example client utterance. An annotated utterance having a predetermined format usable to train the computerized assistant is output to a data store based on the example client utterance.

Abstract: Some aspects of the invention may include a computer-implemented method for enrolling voice prints generated from audio streams, in a database. The method may include receiving an audio stream of a communication session and creating a preliminary association between the audio stream and an identity of a customer that has engaged in the communication session based on identification information. The method may further include determining a confidence level of the preliminary association based on authentication information related to the customer and if the confidence level is higher than a threshold, sending a request to compare the audio stream to a database of voice prints of known fraudsters. If the audio stream does not match any known fraudsters, sending a request to generate from the audio stream a current voice print associated with the customer and enrolling the voice print in a customer voice print database.

Abstract: The present disclosure provides a method, an apparatus and a device for recognizing voice endpoints. In the method of the present disclosure, a start point recognition model and a finish point recognition model are obtained by training a cyclic neural network with a start point training set and a finish point training set, respectively, and a voice start point frame among audio frames is recognized according to each of acoustic features of the audio frames and the start point recognition model, thereby avoiding affecting a delay time of the finish point frame recognition while ensuring the accuracy of the start frame recognition as high as possible; and a voice finish point frame among the audio frames is recognized according to the acoustic features of the audio frames and the finish point recognition model.

Abstract: The present disclosure provides a method and an apparatus for correcting an input speech based on artificial intelligence. The method includes: receiving a speech input by a user; performing recognition on the speech to obtain a current recognition text; obtaining at least one candidate phrase of a first phrase to be corrected in the current recognition text and displaying the at least one candidate phrase to the user; detecting a select operation of the user, the select operation being configured to select one of the at least one candidate phrase as a target candidate phrase; and correcting the first phrase in the current recognition text by using the target candidate phrase, to obtain a target recognition text.

Abstract: A computer-implemented method, according to one embodiment, includes: receiving a complex audio signal which includes an intended audio signal and at least one interfering audio signal. Moreover, the intended audio signal is a voice-based command originating from a user. Information which corresponds to the at least one interfering audio signal is also received. The received information is used to identify portions of the complex audio signal as being the at least one interfering audio signal. Furthermore, the identified portion of the complex audio signal is removed from the complex audio signal, and a remaining portion of the complex audio signal is output.

Abstract: A computer-implemented method is provided. The computer-implemented method is performed by a speech recognition system having at least a processor. The method further includes performing a speech recognition operation on the audio signal data to decode the audio signal data into a textual representation based on the estimated sound identification information from a neural network having periodic indications and components of a frequency spectrum of the audio signal data inputted thereto. The neural network includes a plurality of fully-connected network layers having a first layer that includes a plurality of first nodes and a plurality of second nodes. The method further comprises training the neural network by initially isolating the periodic indications from the components of the frequency spectrum in the first layer by setting weights between the first nodes and a plurality of input nodes corresponding to the periodic indications to 0.

Abstract: An electronic device is provided. The electronic device includes a microphone, a display, a memory, and a processor electrically connected to the microphone, the display, and the memory. The processor is configured to execute a speech recognition function, to obtain a first utterance from a user through the microphone, to display a text generated based on the first utterance, in the display, and to display at least one item, which corresponds to a portion recognized as a proper noun in the text and which is determined based on the first utterance and a personalized database of the user, in the display.

Abstract: A system comprising at least one hardware processor; and a non-transitory computer-readable storage medium having stored thereon program instructions, the program instructions executable by the at least one hardware processor to: receive, as input, a plurality of electronic documents, apply a trained machine learning classifier to automatically classify at least some of said plurality of electronic documents, wherein said machine learning classifier comprises two or more attention layers, and wherein at least one of the attention layers comprises an adjustable parameter which controls a distribution of attention weights assigned by said attention layer.