In an approach for optimizing an intelligent virtual assistant by using phonetic comparison to find a response stored in a local database, a processor receives an audio input on a computing device. A processor transcribes the audio input to text. A processor compares the text to a set of user queries and commands in a local database of the computing device using a phonetic algorithm. A processor determines whether a user query or command of the set of user queries and commands meets a pre-defined threshold of similarity. Responsive to determining that the user query or command meets the pre-defined threshold of similarity, a processor identifies an intention of a set of intentions stored in the local database corresponding to the user query or command. A processor identifies a response of a set of responses in the local database corresponding to the intention. A processor outputs the response audibly.

Systems, apparatuses, and methods are provided for identifying a corresponding string stored in memory based on an incomplete input string. A system can analyze and produce phonetic and distance metrics for a plurality of strings stored in memory by comparing the plurality of strings to an incomplete input string. These similarity metrics can be used as the input to a machine learning model, which can quickly and accurately provide a classification. This classification can be used to identify a string stored in memory that corresponds to the incomplete input string.

Systems and methods are presented herein for increasing user engagement with an interface by suggesting commands or queries for the user. A plurality of content items available for consumption are identified and metadata for each of the plurality of content items is retrieved. One or more candidate voice commands are generated based on a plurality of voice command templates based on a target verb and a subset of the metadata corresponding to the plurality of the content items available for consumption. A recall score is generated for each candidate voice command based at least in part on a detection of phonetic features that match between clauses of each candidate voice command. At least the candidate voice command with the highest recall score is selected and output using a suggestion system.

An artificial intelligence-based semantic recognition method, apparatus, and device. In the artificial intelligence-based semantic recognition method, a pre-trained semantic recognition model is trained by using a training corpus configured by a developer on a model training platform such as a Bot platform and a negative corpus provided on the model training platform, where the negative corpus is extracted by mapping an encoding value of the training corpus to a negative corpus set. Therefore, the negative corpus is extracted based on the encoding value of the training corpus, and a randomized method for generating the negative corpus is changed into a stable method.

Technical solution relates to the fields of voice technologies and knowledge graph technologies. A technical solution includes: acquiring script matched with a scenario from a speech package, and acquiring a broadcast template configured for the scenario in advance; and filling the broadcast template with the script to generate the broadcast speech.

The present disclosure in some embodiments provides a dialogue state tracking method and a dialogue state tracking apparatus in a goal-oriented dialogue system, which track a dialogue state through training based on an attention mechanism between utterances and contextual semantic vectors corresponding respectively to domain-slot types and based on a distance metric-based non-parametric method and thereby facilitate service extension to a multi-domain scenario in the dialogue system.

A data normalization system receives a first string and a second string that are ordered according to an initial string ordering. The data normalization system analyzes, the first string and the second string based on a list of known character sets included in surnames, yielding an analysis, and determines, based on the analysis, that a set of characters in the second string matches a known character set included in the list of known character sets included in surnames. In response to determining that the set of characters in the second string matches a known character set included in the list of known character sets included in surname, the data normalization system orders the first string and the second string according to an updated string ordering.

Described herein are techniques, devices, and systems for database management using sound-based identifiers. The sound-based identifiers can be encoded based on text-based identifiers input into one or more databases. The sound-based identifiers can be preprocessed and encoded by encoding the text-based identifiers with a double metaphone algorithm. First sound-based identifiers can be sorted in a cluster associated with a node of a hybrid prefix tree list, based on a longest common prefix of the group. The first sound-based identifiers can be re-encoded as second sound-based identifiers and organized into sub-clusters associated with nodes, based on characters of the second sound-based identifiers positioned after characters associated with the clusters. The re-encoded sound-based identifiers can be determined based on metadata. A query can be received and utilized to identify a re-encoded sound-based identifier. Data associated with the re-encoded sound-based identifier can be transmitted based on the query.

The present invention relates to an in-document search method and device for a query vector, and an object of the present invention is to improve the accuracy of a response by generating sentence data corresponding to data in a table form stored in database. The in-document search method for a query vector includes a step A of receiving a user query from a user terminal, a step B of generating a user query vector for the user query, a step C of extracting candidate table data based on the user query vector in a data storage module, a step D of searching for a response corresponding to the user query vector in the candidate table data, and a step E of providing the response to the user terminal.

Methods, apparatus, systems, and computer-readable media are provided for providing context specific schema files that allow an automated assistant to broker human-to-computer dialogs between a user and an application that is separate from the automated assistant. The context specific schema file can provide the automated assistant with sufficient data to be responsive to user queries without necessarily communicating with a remote device, such as a server. Multiple different context specific schema files can be made available to the automated assistant according to a context in which a user is interacting with the automated assistant. In this way, latency otherwise exhibited by the automated assistant can be mitigated by providing the automated assistant with the information needed to respond to a user without continually retrieving the information over a network.