US 12,586,584 B1Granted

Selecting tones and text of voice browsing conversations

AssigneeAmazon Technologies, Inc.

InventorsEdward Bueche, Amaury Gutierrez Acosta, Francois Mairesse, Yun Suk Paik, Anmol Tiwari, Tao Ye

Life of this patent

Estimating the term…

Loading patent…

Abstract

Dialog acts (e.g., questions) are selected for voice browsing by a model trained to identify a dialog act that is most likely to lead to a desired outcome. Upon receiving an invocation to begin a conversation, a score indicative of a level of confidence that the conversation will have a successful outcome is determined, and a dialog act is selected based on the score. Subsequently, at each turn of the conversation, the score is updated or a new score is calculated, and a dialog act is selected based on the updated or new score. Confidence scores are calculated based on input features that are determined based on the user who uttered the invocation or responses to dialog acts, as well as a context of the conversation, and provided to a linear model or a machine learning model as inputs.

Claims20

Dependent claims

A method, comprising:

receiving, by a first computer system from a second computer system, first data corresponding to at least a first utterance, wherein the first data is derived at least in part from a first audio signal of a user captured by the second computer system;
determining, by the first computer system, that the first utterance comprises a first invocation to initiate a first conversation;
identifying, by the first computer system, a first set of input features, wherein each of the first set of input features relates to at least one of the user, the first utterance or the second computer system;
calculating, by the first computer system, a first confidence score based at least in part on the first set of input features;
in response to determining that the first data includes an affirmative response, increasing a pitch used in a default tone for use as a first tone;
in response to determining that the first data includes a negative response, decreasing the pitch used in the default tone for use as the first tone;
selecting, by the first computer system based at least in part on the first confidence score, at least the first tone of a first dialog act;
generating, by the first computer system, second data for audible presentation of the first dialog act to the user in response to the first data;
and
transmitting, by the first computer system, the second data using at least the first tone as an audible response to the second computer system.

7 more elements

A computer system comprising:

one or more processors;
and
memory storing one or more sets of instructions that, when executed by the one or more processors, cause the computer system to at least:
receive first data representing a request of a user;
analyze the request to determine features associated with at least one of the user or words included in the request;
generate second data as a response to the request, the second data based at least in part on the features;
select, for audio playback of the second data, a first tone from a plurality of available tones based at least in part on the features, wherein selection of the first tone includes at least one of:
selection of the first tone to have a lower pitch than a standard tone in response to the second data including a negative context;
and
selection of the first tone to have a higher pitch than the standard tone in response to the second data including a positive context;
generate first audio that includes the second data sonically produced at least in part using the first tone;
and
send the first audio to the user.

10 more elements

Dependencies are inferred from the claim text.