Abstract: Implementations described herein relate to an automated assistant that iteratively renders various GUI elements as a user iteratively provides a spoken utterance, or sequence of spoken utterances, corresponding to a request directed to the automated assistant. These various GUI elements can be dynamically adapted as the user iteratively provides the spoken utterance to assist the user with efficiently completing the request. In some implementations, a generic container graphical element associated with candidate intent(s) can be initially rendered at a display interface of a computing device and dynamically adapted with tailored container graphical elements as a particular intent is determined while the user iteratively provides the spoken utterance.

Abstract: Variational Autoencoders (VAEs) have been shown to be effective in modeling complex data distributions. Conventional VAEs operate with fully-observed data during training. However, learning a VAE model from partially-observed data is still a problem. A modified VAE framework is proposed that can learn from partially-observed data conditioned on the fully-observed mask. A model described in various embodiments is capable of learning a proper proposal distribution based on the missing data. The framework is evaluated for both high-dimensional multimodal data and low dimensional tabular data.

Abstract: A method of providing an assistant service, performed by an electronic device, includes: determining content identification information for identifying content displayed on the electronic device; determining a user context for identifying a use situation of a user of the electronic device by using the determined content identification information; generating an utterance list based on the determined content identification information and the determined user context; and, in response to an occurrence of a predefined utterance providing event, outputting the generated utterance list.

Abstract: A soft decision audio decoding system for preserving audio continuity in a digital wireless audio receiver is provided that deduces the likelihood of errors in a received digital signal, based on generated hard bits and soft bits. The soft bits may be utilized by a soft audio decoder to determine whether the digital signal should be decoded or muted. The soft bits may be generated based on the detected point and a detected noise power, or by using a soft-output Viterbi algorithm. The value of the soft bits may indicate confidence in the strength of the hard bit generation. The soft decision audio decoding system may infer errors and decode perceptually acceptable audio without requiring error detection, as in conventional systems, as well as have low latency and improved granularity.

Abstract: Methods and systems are described herein for generating dynamic conversational responses. For example, dynamic conversational responses may facilitate an interactive exchange with users. Therefore, the methods and systems used specialized methods to enriched data that may be indicative of a user's intent prior to processing that data through the machine learning model, as well as a specialized architecture for the machine learning models that take advantage of the user interface format.

Abstract: An extraction apparatus includes: a pre-processing unit configured to perform, on training data that is data described in natural language and in which a tag has been provided to an important description portion in advance, pre-processing for calculating an information gain that indicates a degree of relevance to the tag for each word and deleting a description portion with low relevance to the tag from the training data based on the information gain of each word; and a learning unit configured to learn the pre-processed training data and generate a list of conditional probabilities relating to the tagged description portion.

Abstract: A device for signal processing includes a memory and a processor. The memory is configured to store a parameter associated with a bandwidth-extended audio stream. The processor is configured to select a plurality of non-linear processing functions based at least in part on a value of the parameter. The processor is also configured to generate a high-band excitation signal based on the plurality of non-linear processing functions.

Abstract: A method and an apparatus for automatic speech recognition are provided. The method includes: generating a weight matrix for a layer of a plurality of layers in a neural network; dividing the weight matrix into a plurality of blocks, each block including a plurality of weights; selecting a pre-determined percentage of weights from at least one block for block-wise pruning; and generating a block-wise pruned weight matrix by setting the pre-determined percentage of weights selected from the at least one block to zero. The weight matrix includes a set of weights associated with the layer, the plurality of layers includes a first layer receiving a first input associated with one or more audio feature sequences, and the plurality of layers are executed on one or more processors. The method efficiently accelerates model inference using irregular pruning.

Abstract: An apparatus decodes an encoded audio signal. The apparatus includes a spectral domain audio decoder that generates a decoded representation of a set of spectral portions, the decoded representation being spectral prediction residual values. A frequency regenerator generates a reconstructed spectral portion using a portion of the same set spectral portions. The reconstructed spectral portion also includes spectral prediction residual values. An inverse prediction filter is configured using prediction filter information included in the encoded audio signal and performs an inverse prediction over frequency using the spectral prediction residual values.

Abstract: Systems and methods for similarity search are described. Embodiments identify a document and a query corresponding to a matching phrase in the document, encode the query and a candidate phrase, score the candidate phrase using at least one learning-based score and at least one surface form score, wherein the at least one learning based score is based on the encoding, and the at least one surface form score is based on a surface form of the query and a surface form of the candidate phrase, and select the matching phrase based on the scoring.

Abstract: A signal processing device according to an embodiment of the present invention includes: a conversion unit configured to convert an input mixed acoustic signal into a plurality of first internal states, a weighting unit configured to generate a second internal state which is a weighted sum of the plurality of first internal states based on auxiliary information regarding an acoustic signal of a target sound source when the auxiliary information is input, and generate the second internal state by selecting one of the plurality of first internal states when the auxiliary information is not input, and a mask estimation unit configured to estimate a mask based on the second internal state.

Abstract: A calculation unit (15a) calculates a degree of association between words included in compound word candidates that are each constituted by consecutive words in a document. The calculation unit (15a) calculates the degree of association by using a conditional probability that is a probability of a word that precedes a last word being present in the document under the condition that the last word follows. A selection unit (15b) selects, as a compound word, a compound word candidate whose degree of association thus calculated is higher than a predetermined threshold value. A learning unit (15c) learns a positional relationship between words that include the selected compound word in the document and predetermined tags added to the document. An addition unit (15d) adds the predetermined tags to the document based on the positional relationship between words that include the learned compound word in the document and the predetermined tags added to the document.

Abstract: Provided methods and systems allow dynamic rendering of a reflexive questionnaire based on a modifiable spreadsheet for users with little to no programming experience and knowledge. Some methods comprise receiving a modifiable spreadsheet with multiple rows, each row comprising rendering instructions for a reflexive questionnaire from a first computer, such as a data type cell, statement cell, logic cell, and a field identifier; rendering a graphical user interface, on a second computer, comprising a label and an input element corresponding to the rendering instructions of a first row of the spreadsheet; receiving an input from the second compute; evaluating the input against the logic cell of the spreadsheet; in response to the input complying with the logic cell of the spreadsheet, dynamically rendering a second label and a second input element to the displayed on the graphical user interface based on the logic of the first row.

Abstract: Artificial intelligence (AI) technology can be used in combination with composable communication goal statements to facilitate a user's ability to quickly structure story outlines using “explanation” communication goals in a manner usable by an NLG narrative generation system without any need for the user to directly author computer code. This AI technology permits NLG systems to determine the appropriate content for inclusion in a narrative story about a data set in a manner that will satisfy a desired explanation communication goal such that the narratives will express various ideas that are deemed relevant to a given explanation communication goal.

Abstract: This disclosure relates to a method of generating a symbol string based on an input sentence represented by a sequence of symbols. In particular, the method involves receiving an input symbol string representing a sentence, generating, using a neural network based on a sequence of dependency structure of elements in the input symbol string, an output symbol string corresponding to the input sentence. The neural network includes an encoder that converts elements of the input symbol string to a first hidden state in a form of a multi-dimensional vector, an attention mechanism that applies a weight to the first hidden state and generates the weighted first hidden state as a second hidden state, a decoder that generates a third hidden state based on at least one element of the input symbol string, an element of the output symbol string, and the second hidden state, and an output generator that generates an element of the output symbol string based on the second hidden state and the third hidden state.

Abstract: The present disclosure provides systems and methods that perform machine-learned natural language processing. A computing system can include a machine-learned natural language processing model that includes an encoder model trained to receive a natural language text body and output a knowledge graph and a programmer model trained to receive a natural language question and output a program. The computing system can include a computer-readable medium storing instructions that, when executed, cause the processor to perform operations. The operations can include obtaining the natural language text body, inputting the natural language text body into the encoder model, receiving, as an output of the encoder model, the knowledge graph, obtaining the natural language question, inputting the natural language question into the programmer model, receiving the program as an output of the programmer model, and executing the program on the knowledge graph to produce an answer to the natural language question.

Abstract: The simultaneous transmission and reproduction of a compressed audio signal and a linear PCM signal is satisfactorily achieved. An audio signal of a predetermined unit is sequentially transmitted via a predetermined transmission line to a reception side. The audio signal of the predetermined unit is a mixed signal of a compressed audio signal and a linear PCM signal. For example, the audio signal of the predetermined unit is an audio signal of a sub-frame unit. In this case, for example, in the audio signal of the sub-frame unit, the compressed audio signal is arranged on an upper-order bit side, and the linear PCM signal is arranged on a lower-order bit side.

Abstract: The present disclosure relates generally to systems and methods for analyzing intent. Intents may be analyzed to determine to which device or agent to route a communication. The analyzed intent information can also be used to formulate reports and analyze the accuracy of the identified intents with respect to the received communication.

Abstract: A method for deep reinforcement learning using a neural network model includes receiving a distribution including a plurality of related tasks. Parameters for the reinforcement learning neural network model is trained based on gradient estimation associated with the parameters using samples associated with the plurality of related tasks. Control variates are incorporated into the gradient estimation by automatic differentiation.

Abstract: Computer-readable media, systems and methods may improve automatic summarization of transcripts of financial earnings calls. For example, a system may generate segments, such as by disambiguating sentences, from a transcript to be summarized. The system may use an estimator that assesses whether or not the segment should be included in the summary. Different types of estimators may be used. For example, the estimator may be rule-based, trained based on machine-learning techniques, or trained on based on machine-learning with language modeling using natural language processing to fine-tune language models specific to financial earnings calls.