Abstract: One or more non-transitory computer readable media contain program instructions that, when executed, cause one or more processors to: receive first raw data including one or more first data points generated by a first object of a plurality of objects associated with one or more buildings; generate first input timeseries according to the one or more data points; access a database of interconnected smart entities, the smart entities including object entities representing each of the plurality of objects and data entities representing stored data, the smart entities being interconnected by relational objects indicating relationships between the smart entities; identify a first object entity representing the first object from a first identifier in the first input timeseries; identify a first data entity from a first relational object indicating a relationship between the first object entity and the first data entity; and store the first input timeseries in the first data entity.

Abstract: Techniques for natural language visualizations using query formatting. In some implementations, a computing device receives data indicating a natural language query from a client device that indicates a topic of interest to a user. The computing device parses the data indicating the natural language query from the client device using a predefined delimiter. The computing device identifies one or more objects that represent each term parsed from the data indicating the natural language query. The computing device determines a visualization type based on the one or more objects to visualize a response to the natural language query provided by the mobile device. The computing device visualization data based on the visualization type, the one or more objects that represent each term parsed from the data, to a rendering component to transmit to the client device for display in response to the natural language query.

Abstract: Some embodiments are directed to a computation device configured to verify that an edit script is for transforming a first string to a second string. The edit script has match operations and difference operations as allowed edit operations. The computation device obtains a representation of the edit script and subsequently performs a validation computation. For each match operation, the computation device determines a character at a current position in the first string and a character at a current position in the second string, verifies that they match, increments the current position in the first string by one and increments the current position in the second string by one. For each difference operation, the computation device increments the current position in the first string and/or the current position in the second string by one.

Abstract: A configuration is implemented to receive, with a processor from a customer care platform, a request for spoken language interpretation of a user query from a first spoken language to a second spoken language. The first spoken language is spoken by a user situated at a display-based device that is remotely situated from the customer care platform. The user query is sent from the display-based device by the user to the customer care platform. The configuration performs, at a language interpretation platform, a first spoken language interpretation of the user query from the first spoken language to the second spoken language. Further, the configuration transmits, from the language interpretation platform to the customer care platform, the first spoken language interpretation so that a customer care representative speaking the second spoken language understands the first spoken language being spoken by the user.

Abstract: A method includes detecting, by sensors, a current context associated with an electronic device. The method includes dynamically loading a neural network and selected features into a phrase-spotting audio front-end (AFE) processor. The neural network is configured, based on the current context, with at least one domain having an associated set(s) of trigger words. The method includes detecting, audio content that matches a trigger word from among the sets of trigger words associated with the at least one selected domain. The method includes in response to detecting audio content that matches the trigger word, outputting a wake-up signal to an application processor (AP). The AFE processor utilizes less computational resources than the AP. The method includes, in response to receiving the wake-up signal, the AP waking up and performing additional computation based on the matching trigger word. The method includes outputting results of the additional computation to an output device.

Abstract: Example techniques involve a control hierarchy for a “smart” home having smart appliances and related devices, such as wireless illumination devices, home-automation devices (e.g., thermostats, door locks, etc.), and audio playback devices, among others. An example home includes various rooms in which smart devices might be located. Under the example control hierarchy described herein and referred to as “home graph,” a name of a room (e.g., “Kitchen”) may represent a smart device (or smart devices) within that room. In other words, from the perspective of a user, the smart devices within a room are that room. This hierarchy permits a user to refer to a smart device within a given room by way of the name of the room when controlling smart devices within the home using a voice user interface (VUI) or graphical user interface (GUI).

Abstract: A method for decoding an encoded audio bitstream is disclosed. The method includes receiving the encoded audio bitstream and decoding the audio data to generate a decoded lowband audio signal. The method further includes extracting high frequency reconstruction metadata and filtering the decoded lowband audio signal with an analysis filterbank to generate a filtered lowband audio signal. The method also includes extracting a flag indicating whether either spectral translation or harmonic transposition is to be performed on the audio data and regenerating a highband portion of the audio signal using the filtered lowband audio signal and the high frequency reconstruction metadata in accordance with the flag.

Abstract: A method of controlling a terminal is provided. The method includes analyzing a sensed voice when a voice is sensed, recognizing a context based a result of the analysis, and performing a predetermined control operation based on the recognized context.

Abstract: In some embodiments, a pitch filter for filtering a preliminary audio signal generated from an audio bitstream is disclosed. The pitch filter has an operating mode selected from one of either: (i) an active mode where the preliminary audio signal is filtered using filtering information to obtain a filtered audio signal, and (ii) an inactive mode where the pitch filter is disabled. The preliminary audio signal is generated in an audio encoder or audio decoder having a coding mode selected from at least two distinct coding modes, and the pitch filter is capable of being selectively operated in either the active mode or the inactive mode while operating in the coding mode based on control information.

Abstract: Methods and systems for providing message playback using a shared electronic device is described herein. In response to receiving a request to output messages, a speech-processing system may determine a group account associated with a requesting device, and may determine messages stored by a message data store for the group account. Speaker identification processing may also be performed to determine a speaker of the request. A user account associated with the speaker, and messages stored for the user account, may be determined. A summary response indicating the user account's messages and the group account's message may then be generated such that the user account messages are identified prior to the group account's messages. The messages may then be analyzed to determine an appropriate voice user interface for the requester such that the playback of the messages using a shared electronic device is more natural and conversational.

Abstract: Audio data, corresponding to an utterance spoken by a person within a detection range of a voice communications device, can include an audio message portion. The audio data can be captured and analyzed to determine the intent to send a message. Based at least in part upon that intent, a remaining portion of the audio data can be analyzed to determine the intended message target or recipient, as well as the portion corresponding to the actual message payload. Once determined, the audio file can be trimmed to the message payload, and the message payload of the audio data can be delivered as an audio message to the target recipient.

Abstract: A cascaded finite-state-transducer array includes a plurality of finite-state-transducers, the finite-state-transducers being distributed in space. The finite-state-transducer array is configured with dedicated data transfer channels between the finite-state-transducers to transfer specific data types. Each data stream on a dedicated data transfer channel may transmit a particular data type, which may be sorted in increasing order of start offsets or token IDs.

Abstract: An Artificial Intelligence (AI) based document processing system receives a request including one or more documents related to a process to be automatically executed. The information including the fields and an intent required for the process are extracted from one or more of the request and the documents. The required documents and fields are selected based on the intent and a domain model. The required fields are validated using external knowledge and the discrepancies identified therein are resolved. An internal master document is built based on the required fields. The internal master document is employed for the automatic execution of the process.

Abstract: A speech-capture device can capture audio data during wakeword monitoring and use the audio data to determine if a user is present nearby the device, even if no wakeword is spoken. Audio such as speech, human originating sounds (e.g., coughing, sneezing), or other human related noises (e.g., footsteps, doors closing) can be used to detect audio. Audio frames are individually scored as to whether a human presence is detected in the particular audio frames. The scores are then smoothed relative to nearby frames to create a decision for a particular frame. Presence information can then be sent according to a periodic schedule to a remote device to create a presence “heartbeat” that regularly identifies whether a user is detected proximate to a speech-capture device.

Abstract: Methods, systems, and computer program product for automatically performing sentiment analysis on texts, such as telephone call transcripts and electronic written communications. Disclosed techniques include, inter alia, lexicon training, handling of negations and shifters, pruning of lexicons, confidence calculation for token orientation, supervised customization, lexicon mixing, and adaptive segmentation.

Abstract: This invention is an improvement of technology for automatically generating response voice to voice uttered by a speaker (user), and is characterized by controlling a pitch of the response voice in accordance with a pitch of the speaker's utterance. A voice signal of the speaker's utterance (e.g., question) is received, and a pitch (e.g., highest pitch) of a representative portion of the utterance is detected. Voice data of a responsive to the utterance is acquired, and a pitch (e.g., average pitch) based on the acquired response voice data is acquired. A pitch shift amount for shifting the acquired pitch to a target pitch having a particular relationship to the pitch of the representative portion is determined. When response voice is to be synthesized on the basis of the response voice data, the pitch of the response voice to be synthesized is shifted in accordance with the pitch shift amount.

Abstract: Apparatus, systems, devices, other articles of manufacture and associated methods are disclosed and described herein to process medical data to generate a classification of the medical data using artificial intelligence. An example apparatus includes a processor to execute instructions to implement a history of past illness (HPI) receiver to receive an HPI formatted as a string, the string including one or more words, the words organized in sentences, a natural language processor to tokenize the one or more words into tokens based on a context associated with at least one of the one or more words and a tensor generator to convert the tokens into hashes, each of the hashes forming a dimension of a tensor based on the context. The apparatus further includes a neural network to embed each of the hashes into vectors, process the vectors to classify the HPI as extended or brief based on a similarity to a set of classified HPIs and output a classification for the HPI.

Abstract: The present disclosure provides a synonym tag obtaining method and apparatus, a device, and a computer readable storage medium. By obtaining the central morpheme of the tag to be recognized and then obtaining the tag template of the tag to be recognized according to the tag to be recognized and the central morpheme of the tag to be recognized, to obtain the synonymy tag of the tag to be recognized according to the tag template, and the central morpheme of the tag to be recognized and at least one cluster of synonymy tag templates, it is possible to avoid the technical problem about reduction of the coverage rate of the synonymy tags caused by data sparsity of the mining resources and thereby improve the reliability of obtaining synonymy tags, without depending on mining resources such as a click log having a paraphrase relationship or data having bilingual parallel alignment corpus.

Abstract: Embodiments of the present invention disclose a method, a computer program product, and a computer system for proactive comment evaluation. In embodiments, the invention includes detecting a comment being added to a thread and extracting data relevant to the thread. In addition, the invention includes identifying an average sentiment range of the thread based on the extracted data and a model, as well as identifying a sentiment of the comment. Moreover, the invention involves determining whether the identified sentiment is outside the identified average sentiment range and, based on determining that the identified sentiment is outside the average sentiment range, recommending an alternative comment.

Abstract: Provided are a method by which a translation server collects translated content from at least one device, and a translation server therefor. The translation server may provide, to users using at least one device, a translation request category for participation in translation, and collect translated content from the users. The translation server may provide rewards to the users providing the translated content.