Abstract: Systems and methods for call detail record (CDR) analysis to determine a risk score for a call and identify fraudulent activity and for fraud detection in Interactive Voice Response (IVR) systems. An example method may store information extracted from received calls. Queries of the stored information may be performed to select data using keys, wherein each key relates to one of the received calls, and wherein the queries are parallelized. The selected data may be transformed into feature vectors, wherein each feature vector relates to one of the received calls and includes a velocity feature and at least one of a behavior feature or a reputation feature. A risk score for the call may be generated during the call based on the feature vectors.

Abstract: A speaker recognition algorithm is trained (one or more of its models are tuned) with samples of a microphone signal produced by an inside microphone of a headphone, while the headphone is worn by a speaker. The trained speaker recognition algorithm then tests other samples of the inside microphone signal and produces multiple speaker identification scores for its given models, or a single speaker verification likelihood score for a single given model. Other embodiments are also described and claimed.

Abstract: A cleaning robot and a shortest path planning method based on a cleaning robot are disclosed, a plurality of cleaning lines are formed by controlling the cleaning robot to perform cleaning in an area according to a zigzag-shaped path; association information of midpoints of at least a part of the cleaning lines is recorded to form a node skeleton tree in which midpoints are represented by nodes, the association information of each midpoint includes: position information of a node corresponding to the midpoint, position information of a parent node, and information of the number of child nodes of the parent node; in the process of traversing upwardly from a current node or traversing upwardly from both a current node and a target node in the node skeleton tree, the node skeleton tree is compressed, so as to determine the shortest planned path from the current node to the target node.

Abstract: The present disclosure relates to speech recognition systems and methods that enable personalized vocal user interfaces. More specifically, the present disclosure relates to combining a self-learning speech recognition system based on semantics with a speech-to-text system optionally integrated with a natural language processing system. The combined system has the advantage of automatically and continually training the semantics-based speech recognition system and increasing recognition accuracy.

Abstract: An acoustic environment identification system is disclosed that can use neural networks to accurately identify environments. The acoustic environment identification system can use one or more convolutional neural networks to generate audio feature data. A recursive neural network can process the audio feature data to generate characterization data. The characterization data can be modified using a weighting system that weights signature data items. Classification neural networks can be used to generate a classification of an environment.

Abstract: The present disclosure provides an electronic apparatus having an incremental enrollment unit and a method thereof. The electronic apparatus at least includes a microphone, a storage device, and a processor. The storage device stores a first screening rule, an enrollment database, and a first temporary storage library. The processor receives a command voice transmitted by the microphone, and compare the command voice with enrolled voices in the enrollment database. If determining that a similarity is larger than a threshold value, the processor stores the command voice as a first temporarily stored voice in the first temporary storage library. When a quantity of the first temporarily stored voices in the first temporary storage library is larger than a first predetermined value, the processor screens out a part of the first temporarily stored voices according to the first screening rule, so as to perform incremental enrollment.

Abstract: In non-limiting examples of the present disclosure, systems, methods and devices for providing personalized experiences to a computing device based on user input such as voice, text and gesture input are provided. Acoustic patterns associated with voice input, speech patterns, language patterns and natural language processing may be used to identify a specific user providing input from a plurality of users, identify user background characteristics and traits for the specific user, and topically categorize user input in a tiered hierarchical index. Topically categorized user input may be supplemented with user data and world knowledge and personalized responses and feedback for an identified specific user may be provided reactively and proactively.

Abstract: Systems and processes for operating a virtual assistant to provide natural assistant interaction are provided. In accordance with one or more examples, a method includes, at an electronic device with one or more processors and memory: receiving a first audio stream including one or more utterances; determining whether the first audio stream includes a lexical trigger; generating one or more candidate text representations of the one or more utterances; determining whether at least one candidate text representation of the one or more candidate text representations is to be disregarded by the virtual assistant. If at least one candidate text representation is to be disregarded, one or more candidate intents are generated based on candidate text representations of the one or more candidate text representations other than the to be disregarded at least one candidate text representation.

Abstract: Apparatus and methods for audio classifying and processing are disclosed. In one embodiment, an audio processing apparatus includes an audio classifier for classifying an audio signal into at least one audio type in real time; an audio improving device for improving experience of audience; and an adjusting unit for adjusting at least one parameter of the audio improving device in a continuous manner based on the confidence value of the at least one audio type.

Abstract: A system and method are presented for the correction of packet loss in audio in automatic speech recognition (ASR) systems. Packet loss correction, as presented herein, occurs at the recognition stage without modifying any of the acoustic models generated during training. The behavior of the ASR engine in the absence of packet loss is thus not altered. To accomplish this, the actual input signal may be rectified, the recognition scores may be normalized to account for signal errors, and a best-estimate method using information from previous frames and acoustic models may be used to replace the noisy signal.

Abstract: An approach is provided for decluttering a device desktop. Using a classification technique, a subject of a current task of a user using a device is determined. Based on a determination that the subject matches a category of first desktop object(s), the first desktop object(s) are identified as being related to the current task. Based on a determination that the subject does not match one or more categories of second desktop object(s), the second desktop object(s) are identified as being not related to the current task. Based on the second desktop object(s) being not related to the current task, the second desktop object(s) are hidden from being viewed on the desktop.

Abstract: Techniques for using a dynamic wakeword detection threshold are described. A device detects a wakeword in audio data using a first wakeword detection threshold value. Thereafter, the device receives audio including speech. If the device receives the audio within a predetermined duration of time after detecting the previous wakeword, the device attempts to detect a wakeword in second audio data, corresponding to the audio including the speech, using a second, lower wakeword detection threshold value.

Abstract: A method of detecting a replay attack on a voice biometrics system comprises: receiving an audio signal representing speech; detecting a magnetic field; determining if there is a correlation between the audio signal and the magnetic field; and if there is a correlation between the audio signal and the magnetic field, determining that the audio signal may result from a replay attack.

Abstract: A speech service control apparatus and a method thereof are provided. Speech data is obtained, and a keyword in the speech data is recognized to determine a confidence value corresponding to the keyword, which is a match level of the keyword relative to a wakeup keyword to request for speech services. When the confidence value is inferior to a recognized threshold, a number of cumulative failures is determined. The speech services are requested because the confidence value is greater than the recognized threshold, and the number of cumulative failure is a cumulative number accumulated when the speech data and previous speech data are inferior to the recognized threshold within a time period. The recognized threshold is modified according to the number of cumulative failure, a calculation relationship of confidence values of the speech data and the previous speech data, to enable the speech services successfully.

Abstract: The intelligent automated assistant system engages with the user in an integrated, conversational manner using natural language dialog, and invokes external services when appropriate to obtain information or perform various actions. The system can be implemented using any of a number of different platforms, such as the web, email, smartphone, and the like, or any combination thereof. In one embodiment, the system is based on sets of interrelated domains and tasks, and employs additional functionally powered by external services with which the system can interact.

Abstract: Systems and processes for operating an intelligent automated assistant are provided. An example process receives natural language input and determines a first and a second parsing result for the natural language input. The first and the second parsing results include respective mappings of one or more properties of a domain corresponding to the natural language input to one or more words of the natural language input. The process determines whether the second parsing result corresponds to a data item in a knowledge base, and in accordance with determining that the second parsing result corresponds to the data item in the knowledge base, the process ranks the second parsing result higher than the first parsing result. Based on the ranking, the process generates a task flow using the second parsing result and executes the task flow to provide an output based on the data item.

Abstract: In one embodiment, a method includes receiving a search query inputted by a first user, wherein the search query comprises one or more n-grams; calculating a needle-confidence score for the search query that is calculated by a needle-intent classifier based on at least the n-grams of the search query and a language model analysis of the n-grams, and wherein the needle-confidence score represents a probability that the search query was intended as a needle search; classifying the search query as a needle search if the calculated needle-confidence score is above a threshold confidence score; and generating a plurality of search-result modules, each search-result module comprising one or more search results matching the search query, wherein one of the search-result modules is a social module, and wherein the number of search results in the social module is based on the classification of the search query as a needle search.

Abstract: A navigation device includes a touch screen configured to receive a search term and to display at least one Point To Interest (POI) information corresponding to the search term, a computer-readable memory configured to store a POI data by region, and a controller configured to split the search term into at least two parts including a front portion and a second portion next to the front portion, determine a target region based on the second portion, find a matched POI data among POI data related to the target region based on the whole of the search term, and output a search result as the at least one POI information into the touch screen.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for performing dynamic, stroke-based alignment of touch displays. In one aspect, a method includes obtaining a candidate transcription that an automated speech recognizer generates for an utterance, determining a particular context associated with the utterance, determining that a particular n-gram that is included in the candidate transcription is included among a set of undesirable n-grams that is associated with the context, adjusting a speech recognition confidence score associated with the transcription based on determining that the particular n-gram that is included in the candidate transcription is included among the set of undesirable n-grams that is associated with the context, and determining whether to provide the candidate transcription for output based at least on the adjusted speech recognition confidence score.

Abstract: Provided is a machine learning-based object detection method performed by an object detection apparatus. the method comprises constructing an object detection model by performing machine learning on a training image set, wherein the object detection model is a model for detecting a target object in an input image based on the result of comparing a confidence score for the target object with a threshold value, obtaining an input image given a detection result for the target object, wherein the obtained input image is an image not included in the training image set, predicting one or more object regions, in which the target object exists, in the obtained input image by using the object detection model, classifying a region not matching the detection result among the predicted object regions as a false detection region; and adjusting the threshold value of the object detection model based on a confidence score of the false detection region.

Abstract: Systems and processes for operating a virtual assistant programmed to refer to shared domain concepts using concept nodes are provided. An example process includes receiving a user speech input, determining a primary domain corresponding to a textual representation of the user speech input, identifying, from the textual representation, a first substring that corresponds to a first concept of the primary domain, parsing the first substring to determine a secondary domain of the plurality of domains, and based on the secondary domain, obtaining a data item corresponding to the first substring. In accordance with determining that the data item is valid for resolving the first concept of the primary domain, extracting, from the data item, a parameter value for the first concept of the primary domain and invoking a service based on the primary domain to produce a result using the parameter value for the first concept.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for receiving audio data corresponding to an utterance, determining that the audio data corresponds to a hotword, generating a hotword audio fingerprint of the audio data that is determined to correspond to the hotword, comparing the hotword audio fingerprint to one or more stored audio fingerprints of audio data that was previously determined to correspond to the hotword, detecting whether the hotword audio fingerprint matches a stored audio fingerprint of audio data that was previously determined to correspond to the hotword based on whether the comparison indicates a similarity between the hotword audio fingerprint and one of the one or more stored audio fingerprints that satisfies a predetermined threshold, and in response to detecting that the hotword audio fingerprint matches a stored audio fingerprint, disabling access to a computing device into which the utterance was spoken.

Abstract: There is provided a speech processing apparatus to improve the flexibility of processing regarding speech recognition, the speech processing apparatus including: a determination unit configured to determine how to deal with a recognition result of speech data obtained by a first speech recognizer on a basis of a comparison between a certainty factor of the recognition result of the speech data obtained by the first speech recognizer and a threshold; and a threshold setting unit configured to set dynamically the threshold. The method further comprises using three operation modes where third, second and first modes comprise of three, two and one recognizers respectively, and the threshold in the third mode is lower than the second mode and higher than the first mode.

Abstract: Utterance-based user interfaces can include activation trigger processing techniques for detecting activation triggers and causing execution of certain commands associated with particular command pattern activation triggers without waiting for output from a separate speech processing engine. The activation trigger processing techniques can also detect speech analysis patterns and selectively activate a speech processing engine.

Abstract: The present disclosure provides a voice recognition system configured to generate a custom phoneme mapping for a user. The voice recognition system can analyze a user speech sample of a grammar training set in order to generate the custom phoneme mapping. The custom phoneme mapping can be used for subsequent recognition of the user's voice within an application.

Abstract: Techniques are described for machine-trained analysis for multimodal machine learning. A computing device captures a plurality of information channels, wherein the plurality of information channels includes contemporaneous audio information and video information from an individual. A multilayered convolutional computing system learns trained weights using the audio information and the video information from the plurality of information channels, wherein the trained weights cover both the audio information and the video information and are trained simultaneously, and wherein the learning facilitates emotional analysis of the audio information and the video information. A second computing device captures further information and analyzes the further information using trained weights to provide an emotion metric based on the further information.

Abstract: An apparatus for generating an error concealment signal, includes: an LPC representation generator for generating a replacement LPC representation; an LPC synthesizer for filtering a codebook information using the replacement LPC representation; and a noise estimator for estimating a noise estimate during a reception of good audio frames, wherein the noise estimate depends on the good audio frames representation generator is configured to use the noise estimate estimated by the noise estimator in generating the replacement LPC representation.

Abstract: A speaker recognition system includes a previously-trained joint neural network. An enrollment machine of the speaker recognition system is configured to operate the previously-trained joint neural network to enroll a new speaker based on audiovisual data featuring the newly enrolled speaker. A recognition machine of the speaker recognition system is configured to operate the previously-trained joint neural network to recognize a previously-enrolled speaker based on audiovisual data featuring the previously-enrolled speaker.

Abstract: An apparatus for generating an error concealment signal includes an LPC (linear prediction coding) representation generator for generating a first replacement LPC representation and a different second replacement LPC representation; an LPC synthesizer for filtering a first codebook information using the first replacement representation to obtain a first replacement signal and for filtering a different second codebook information using the second replacement LPC representation to obtain a second replacement signal; and a replacement signal combiner for combining the first replacement signal and the second replacement signal to obtain the error concealment signal.

Abstract: Disclosed herein are system, method, and computer program product embodiments for recognizing a human emotion in a message. An embodiment operates by receiving a message from a user. The embodiment labels each word of the message with a part of speech (POS) thereby creating a POS set. The embodiment creates a bag of words (BOW) for the message. The embodiment determines an incongruity score for a combination of words in the POS set using a knowledgebase. The embodiment determines a preliminary emotion detection score for an emotion for the message based on the POS set and the BOW. Finally, the embodiment calculates a final emotion detection score for the emotion for the message based on the preliminary emotion detection score and the incongruity score.

Abstract: An apparatus for extracting selected information from a set of symbols includes said alignment module is configured to retrieve test patterns from a symbol input, and to attempt alignment of test patterns with a canonical pattern. Successful alignment between a particular test pattern and said canonical pattern indicates of existence of information of interest in a particular candidate pattern. Upon detection of a successful alignment, the alignment module passes information concerning the test pattern to a user. Additionally, in response to detecting an unsuccessful attempt to align the first test pattern and the canonical pattern, said alignment module passes, to said user, information concerning the first test pattern.

Abstract: Exemplary embodiments relate to the application of media effects, such as visual overlays, sound effects, etc. to a video conversation. A media effect may be applied as a reaction to an occurrence in the conversation, such as in response to an emotional reaction detected by emotion analysis of information associated with the video. Effect application may be controlled through gestures, such as applying different effects with different gestures, or canceling automatic effect application using a gesture. Effects may also be applied in group settings, and may affect multiple users. A real-time data channel may synchronize effect application across multiple participants. When broadcasting a video stream that includes effects, the three channels may be sent to an intermediate server, which stitches the three channels together into a single video stream; the single video stream may then be sent to a broadcast server for distribution to the broadcast recipients.

Abstract: A transmission controller monitors a sound pressure determination signal and a distance determination signal. The transmission controller controls a transmission voice processor to start an operation of generating a transmission voice signal, when the distance determination signal indicates that a distance is equal to or less than a first distance. The transmission controller controls to start an operation of determining a sound pressure of a voice signal, when the distance determination signal indicates that the distance is equal to or less than a second distance shorter than the first distance. The transmission controller supplies a transmission control signal to a transmission circuit so that the transmission circuit transmits the transmission voice signal as a radio wave, when the sound pressure determination signal indicates that the sound pressure is equal to or greater than a predetermined threshold value.

Abstract: Disclosed are apparatuses, methods, and computer readable media for improved intelligent personal assistant (IPA) software agents that are configured to interact with various people, service providers, and/or smart devices across multiple connection protocols, communications formats, and communication protocols in a seamless and more accurate fashion. More particularly, but not by way of limitation, this disclosure relates to apparatuses, methods, and computer readable media for an improved Message Understanding Service (MUS) that is able to match generic user commands and queries (i.e., commands and queries that are not explicitly directed to a particular service endpoint or smart device) with the service endpoint(s) that have the greatest confidence level of being able to handle the generic command or query.

Abstract: Systems and processes for providing user-specific acoustic models are provided. In accordance with one example, a method includes, at an electronic device having one or more processors, receiving a plurality of speech inputs, each of the speech inputs associated with a same user of the electronic device; providing each of the plurality of speech inputs to a user-independent acoustic model, the user-independent acoustic model providing a plurality of speech results based on the plurality of speech inputs; initiating a user-specific acoustic model on the electronic device; and adjusting the user-specific acoustic model based on the plurality of speech inputs and the plurality of speech results.

Abstract: A system and method is provided for providing searchable customer call indexes. Consistent with disclosed embodiments, a system may receive call information associated with telephone conversations between callers and a vendor, the call information including an audio recording or transcript for each telephone conversation. The system may also identify one or more keywords from the audio recordings or transcripts and index the call information into one or more indexes based on the identified keywords. Finally, the system may determine search results responsive to a search query based on the indexing. In some embodiments, changes to customer service may be identified based on the search results.

Abstract: Embodiments of the invention are directed to systems and methods for biometrics transaction processing. A location of a device associated with a user may be determined. A reference to a biometric data model associated with the user stored within a database may be retrieved, based at least in part on the location. Biometric data may be received from the user. Using the reference, the biometric data may be compared to the biometric data model stored within the database. A determination may be made whether the user is authenticated for the transaction based on the comparing step.

Abstract: A call handling platform receives a call placed by a caller to a calling number. The call handling platform computes an experience score for the caller using measurements of a subset of data points based on an interaction of the caller with an interactive voice response (IVR) module during the call. The experience score reflects a numerical measure of a level of satisfaction of the caller in interacting with the IVR module. The call handling platform compares the experience score to a predetermined threshold that indicates a minimum level of caller satisfaction, and determines that the experience score indicates that the caller has a lower level of satisfaction than the minimum level of satisfaction. Conditioned on this determination, the call handling platform routes the call to a human agent at a call center, along with enabling the agent to perceive a representation of the experience score.

Abstract: A voice control type bath system and an operating method thereof are disclosed. The voice control type bath system is utilized for a massage bath equipment and includes at least one attached device for actuating the massage bath equipment, a voice receiving unit for receiving at least one voice signal, a voice analyzing module for analyzing the at least one voice signal to generate at least one controlling command; and a main control device for controlling the at least one attached device to actuate the massage bath equipment according to the at least one controlling command and/or for controlling an actuation of the at least one attached device according to the at least one controlling command. The voice control type bath system and the operating method thereof can directly control the at least one attached device via the at least one voice signal.

Abstract: Systems and processes for efficient word encoding are provided. In accordance with one example, a method includes, at an electronic device with one or more processors and memory, receiving a user input including a word sequence, and providing a representation of a current word of the word sequence. The representation of the current word may be indicative of a class of a plurality of classes and a word associated with the class. The method further includes determining a current word context based on the representation of the current word and a previous word context, and providing a representation of a next word of the word sequence. The representation of the next word of the word sequence may be based on the current word context. The method further includes displaying, proximate to the user input, the next word of the word sequence.

Abstract: Systems, methods, and computer-readable storage devices for crowd-sourced data labeling. The system requests a respective response from each of a set of entities. The set of entities includes crowd workers. Next, the system incrementally receives a number of responses from the set of entities until one of an accuracy threshold is reached and m responses are received, wherein the accuracy threshold is based on characteristics of the number of responses. Finally, the system generates an output response based on the number of responses.

Abstract: Disclosed herein is a method for analyzing an utterance meaning. The method includes collecting a voice signal of an utterer; converting the collected voice signal into information in a text form, extracting a keyword of the text information from the text information, and deriving at least one utterance topic on the basis of the extracted keywords of the text information.

Abstract: Techniques are provided for training of a text independent (TI) speaker recognition (SR) model. A methodology implementing the techniques according to an embodiment includes measuring context data associated with collected TI speech utterances from a user and identifying the user based on received identity measurements. The method further includes performing a speech quality analysis and a speaker state analysis based on the utterances, and evaluating a training merit value of the utterances, based on the speech quality analysis and the speaker state analysis. If the training merit value exceeds a threshold value, the utterances are stored as training data in a training database. The database is indexed by the user identity and the context data. The method further includes determining whether the stored training data has achieved a sufficiency level for enrollment of a TI SR model, and training the TI SR model for the identified user and context.

Abstract: A method of recognizing speech. The method includes the steps of obtaining an auditory signal; processing the auditory signal into a plurality of frequency components; processing the plurality of frequency components using a plurality of feature detectors, each feature detector producing a feature detector response; generating a spike for each instance in which a feature detector response identifies a characteristic auditory feature to produce a test spike pattern for the auditory signal; and comparing the test spike pattern to a plurality of predetermined spike patterns corresponding to a plurality of speech elements to determine whether the auditory signal includes one of the plurality of speech elements.

Abstract: A system and method are disclosed for scoring an interaction over one or more channels by an end user and an entity by monitoring the communications over the channels and assigning penalties and scores based on topics associated with the communications.

Abstract: A system and method for detecting phonetically similar imposter phrases may include using automatic speech recognition (ASR) to search for a first phrase in a set of objects; producing a list of references by searching for the first phrase in the set of objects using phonetic search; using output produced by the ASR to determine whether or not a reference in the list points to a phrase that is the same as the first phrase; and if it is determined that the reference points to a second phrase that is different from the first phrase then marking the second phrase as a potential cause for a phrase search false positive.

Abstract: The present disclosure provides a voice recognition method, when receiving a voice signal, an image acquisition device is controlled to acquire images, and when the voice signal being over, the image acquisition device is controlled to stop acquiring the image; the received voice signal is recognized to obtain a voice signal recognition result; accuracies of the voice signal recognition result and the lip-reading recognition result is calculated, the recognition result with a higher accuracy is taken as a current voice recognition result. The present disclosure also provides a voice recognition system. The present disclosure improves the accuracy of the voice recognition.

Abstract: A signal processing method for cochlear implant is performed by a speech processor and comprises a noise reduction stage and a signal compression stage. The noise reduction stage can efficiently reduce noise in a electrical speech signal of a normal speech. The signal compression stage can perform good signal compression to enhance signals to stimulate cochlear nerves of a patient with hearing loss. The patient who uses a cochlear implant performing the signal processing method of the present disclosure can accurately hear normal speech.

Abstract: A server receives sensor information from a wireless device and provides control signals based on the sensor information to the wireless device for enabling the device to perform a control task. The server receives an estimate of imminent connection state of the wireless device. Upon the state indicating a connection quality parameter to have a quality below a first threshold, the server provides an information package to be sent to the wireless device for enabling the device to perform an increased part of the control task locally at the device. Upon the state indicating the connection quality parameter to have a quality above a second threshold, the server provides an information package to be sent to the wireless device for enabling the device to be relieved from performing at least a part of the control task by the server providing increased aid in performing the control task.

Abstract: Applications and services are disclosed for transforming spoken thoughts to a visual representation of related topics found in the spoken thoughts. Transforming the spoken thoughts can include receiving audio input of a spoken thought of a user and converting the audio input to text of the spoken thought. The system can identify a plurality of topics from the text using word frequency, and can create a topic graph for the plurality of topics by determining relationships between topics of the plurality of topics, and assigning a weight to each of the plurality of topics based at least on the determined relationships. The system can generate a visual display of the topic graph for the plurality of topics by representing relevant topics of the plurality of topics using topic shapes, which can be connected by the determined relationships and sized by the assigned weight.