Amazon Technologies, Inc.

Kevin Boehm

Mark Conrad Kockerbeck

Rohit Sauhta

Muhammad Yahia

Jordan Michael Hughes

Ian Christopher Suttle

Srikanth Nori

Tina Ram Orooji

Robert Pulciani

John Peach

The present disclosure describes techniques for dynamically determining when information is to be output to a user, as well as what information is to be output to a user. A natural language processing system may receive, from a first device, first data representing information to be output at a first point during a skill session. The natural language processing system may also receive, from a second device, second data representing a natural language input. The natural language processing system may determine a skill component is to execute with respect to the natural language input. The natural language processing system may send, to the skill component, second data representing the natural language input. The natural language processing system may receive, from the skill component, an indication that an ongoing first skill session with the second device has reached the first point. After receiving the indication and based at least in part on system usage data associated with at least one user, the natural language processing system may determine third data representing a prompt corresponding to the information and send, to the second device, the third data for output.

Contextual content for voice user interfaces

Techniques for selecting a skill, to respond to a user input, using skill rankings are described. A skill's ranking may be determined in different manners. In one example, a skill's ranking may be determined based on a number of different users inputting commands that invoke the skill over a period of time. In another example, a skill's ranking may be determined based on a number of different devices that capture user inputs that invoke the skill over a period of time. A system may determine whether to automatically enable a skill (e.g., without user input received after the original user input), or ask the user whether the skill should be enabled, based on the skill's ranking. Moreover, a system may use a skill's ranking to determine whether to interact with a user, to enable the skill, using a graphical user interface or a voice user interface.

Skill enablement

Techniques for recommending speechlets to users via voice user interfaces (VUIs) are described. A system may include predetermined rankings of speechlets. The predetermined rankings may be determined offline and may represent reliabilities of the speechlets. During runtime, when the system receives a user input that can be handled by multiple speechlets, the system may determine a type of speechlet configured to handle the user input. The system may also identify information specific to the user's previous interactions with the system. The system may rerank the predetermined rankings using the user specific information, as well as other information. The system may recommend speechlets to the user based on the speechlets' rerankings. The system may first recommend the top reranked speechlet. If the user accepts the recommendation, the system may cause the top reranked speechlet to handle the original user input. If the user declines the recommendation, the system may recommend the next top reranked speechlet to the user. The system may work its way down the rerankings of the speechlets until the user accepts a recommendation, until the system recommends all of the reranked speechlets, or until the user indicates the system should stop recommending speechlets.

Speechlet recommendation based on updating a confidence value

Growing community of inventors

Seattle, WA, United States of America

Kevin Boehm