A computer system configured for autonomous learning of entity values is provided. The computer system includes a memory that stores associations between entities and fields of response data. The computer system also includes a processor configured to receive a request to process an intent; generate a request to fulfill the intent; transmit the request to a fulfillment service; receive, from the fulfillment service, response data specifying values of the fields; identify the values of the fields within the response data; identify the entities via the associations using the fields; store, within the memory, the values of the fields as values of the entities; and retrain a natural language processor using the values of the entities.

A method for dialog state tracking includes decoding, by a fertility decoder, encoded dialog information associated with a dialog to generate fertilities for generating dialog states of the dialog. Each dialog state includes one or more domains. Each domain includes one or more slots. Each slot includes one or more slot tokens. The method further includes generating an input sequence to a state decoder based on the fertilities. A total number of each slot token in the input sequence is based on a corresponding fertility. The method further includes encoding, by a state encoder, the input sequence to the state decoder, and decoding, by the state decoder, the encoded input sequence to generate a complete sequence of the dialog states.

A computer system includes memory storing computer-executable instructions and a processor configured to execute the computer-executable instructions. The computer-executable instructions include building a questions table including multiple first questions and multiple potential requirements to which the first questions correspond, respectively. The computer-executable instructions include adding, to the questions table, one or more second questions, each of which correspond to at least two requirements from among the potential requirements. The computer-executable instructions include adding, to the questions table, for each second question among the one or more second questions, the at least two requirements to which the second question corresponds. The computer-executable instructions include determining a question, from among the first questions and the second questions, that corresponds to a highest number of potential requirements, and displaying the determined question to a user.

A system and method are disclosed to generate, transmit, and automate communications with end user systems. Embodiments comprise an automation platform comprising a processor and memory. Embodiments generate a communication based, at least in part, on input from a rules engine and one or more communication templates. Embodiments modify the content of the generated communication and revise the one or more communication templates to include the modifications made to the communication content. Embodiments transmit, using one or more communication channels, the modified communication to one or more end user systems, and automate the generation and transmission of one or more subsequent communications to the one or more end user systems based, at least in part, on the revised one or more communication templates.

A system and method are disclosed to generate, transmit, and automate communications with end user systems. Embodiments comprise an automation platform comprising a processor and memory. Embodiments generate a communication based, at least in part, on input from a rules engine and one or more communication templates. Embodiments modify the content of the generated communication and revise the one or more communication templates to include the modifications made to the communication content. Embodiments transmit, using one or more communication channels, the modified communication to one or more end user systems, and automate the generation and transmission of one or more subsequent communications to the one or more end user systems based, at least in part, on the revised one or more communication templates.

Devices and techniques are generally described for application determination in speech processing. Input data corresponding to a spoken utterance may be received. Speech recognition processing may be performed on the input data to generate text data. A machine learning encoder may generate a vector representation of the input data. A first binary classifier may determine a first probability that the input data corresponds to a first speech-processing application. A second binary classifier may determine a second probability that the input data corresponds to a second speech-processing application. A selection between the first speech-processing application and the second speech-processing application may be made based at least in part on the first probability and the second probability.

A disconnection of a client device is disconnected during a multi-participant communication, such as a call or a conference. An indication of the disconnection is transmitted to the client device to cause an agent at the client device to record media locally at the client device. The media recorded by the agent at the client device based on the indication of the disconnection is later received and included within a recording of the communication. For example, a gap of the recording in which the disconnection occurred may be identified, such as by performing a comparison of media within the recording to identify a start time of the gap and an end time of the gap. The media is then inserted within a portion of the recording of the multi-participant communication corresponding to the gap.

A disconnection of a client device is disconnected during a multi-participant communication, such as a call or a conference. An indication of the disconnection is transmitted to the client device to cause an agent at the client device to record media locally at the client device. The media recorded by the agent at the client device based on the indication of the disconnection is later received and included within a recording of the communication. For example, a gap of the recording in which the disconnection occurred may be identified, such as by performing a comparison of media within the recording to identify a start time of the gap and an end time of the gap. The media is then inserted within a portion of the recording of the multi-participant communication corresponding to the gap.

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.

An embodiment includes tokenizing an input passage into an n-gram sequence of tokens. The embodiment also includes evaluating the input passage using a trained classification model that generates an output indicative of a classification of the input passage. The embodiment also includes generating a first token vector for a first token of the sequence of tokens and projecting the first token vector to a higher dimensional space, resulting in a first projected token vector. The embodiment also includes generating a first similarity score for the first projected token vector based on comparisons of the first projected token vector to each of a plurality of class representations. The embodiment also includes generating a ranked list of the tokens, wherein the generating of the ranked list includes ranking the first token among others of the tokens based on the first similarity score.