Abstract: Techniques for determining supplemental content of potential interest to a user are described. A system may receive system usage data of a user and supplemental content presentation data including supplemental content and indicating how it is to be presented. The system may determine first embedding data representing the system usage data and second embedding data representing the supplemental content presentation data. The system may receive dialog data including a user input and a system response to the user input. The system may determine, using a language model, third embedding data representing the user input, fourth embedding data representing the system response, and fifth embedding data representing the supplemental content. The system may determine a pairwise dot product using the first through fifth embedding data, and may process, using a neural network, the pairwise dot product to determine the supplemental content is to be presented.

Abstract: Methods and systems for determining an intent of an utterance using contextual information associated with a requesting device are described herein. Voice activated electronic devices may, in some embodiments, be capable of displaying content using a display screen. Entity data representing the content rendered by the display screen may describe entities having similar attributes as an identified intent from natural language understanding processing. Natural language understanding processing may attempt to resolve one or more declared slots for a particular intent and may generate an initial list of intent hypotheses ranked to indicate which are most likely to correspond to the utterance. The entity data may be compared with the declared slots for the intent hypotheses, and the list of intent hypothesis may be re-ranked to account for matching slots from the contextual metadata. The top ranked intent hypothesis after re-ranking may then be selected as the utterance's intent.

Abstract: Methods and systems for resolving entities using multi-modal functionality are described herein. Voice activated electronic devices may, in some embodiments, be capable of displaying content using a display screen. Contextual metadata representing the content rendered by the display screen may describe entities having similar attributes as an identified intent from natural language understanding processing. When natural language understanding processing attempts to resolve one or more declared slots for a particular intent, matching slots from the contextual metadata may be determined, and the matching entities may be placed in an intent selected context file to be included with the natural language understanding's output data. The output data may be provided to a corresponding application for causing one or more actions to be performed.

Abstract: Methods and systems for determining an intent of an utterance using contextual information associated with a requesting device are described herein. Voice activated electronic devices may, in some embodiments, be capable of displaying content using a display screen. Entity data representing the content rendered by the display screen may describe entities having similar attributes as an identified intent from natural language understanding processing. Natural language understanding processing may attempt to resolve one or more declared slots for a particular intent and may generate an initial list of intent hypotheses ranked to indicate which are most likely to correspond to the utterance. The entity data may be compared with the declared slots for the intent hypotheses, and the list of intent hypothesis may be re-ranked to account for matching slots from the contextual metadata. The top ranked intent hypothesis after re-ranking may then be selected as the utterance's intent.

Abstract: Methods and systems for determining an intent of an utterance using contextual information associated with a requesting device are described herein. Voice activated electronic devices may, in some embodiments, be capable of displaying content using a display screen. Entity data representing the content rendered by the display screen may describe entities having similar attributes as an identified intent from natural language understanding processing. Natural language understanding processing may attempt to resolve one or more declared slots for a particular intent and may generate an initial list of intent hypotheses ranked to indicate which are most likely to correspond to the utterance. The entity data may be compared with the declared slots for the intent hypotheses, and the list of intent hypothesis may be re-ranked to account for matching slots from the contextual metadata. The top ranked intent hypothesis after re-ranking may then be selected as the utterance's intent.

Abstract: Methods and systems for resolving entities using multi-modal functionality are described herein. Voice activated electronic devices may, in some embodiments, be capable of displaying content using a display screen. Contextual metadata representing the content rendered by the display screen may describe entities having similar attributes as an identified intent from natural language understanding processing. When natural language understanding processing attempts to resolve one or more declared slots for a particular intent, matching slots from the contextual metadata may be determined, and the matching entities may be placed in an intent selected context file to be included with the natural language understanding's output data. The output data may be provided to a corresponding application for causing one or more actions to be performed.

Abstract: Methods and systems for resolving entities using multi-modal functionality are described herein. Voice activated electronic devices may, in some embodiments, be capable of displaying content using a display screen. Contextual metadata representing the content rendered by the display screen may describe entities having similar attributes as an identified intent from natural language understanding processing. When natural language understanding processing attempts to resolve one or more declared slots for a particular intent, matching slots from the contextual metadata may be determined, and the matching entities may be placed in an intent selected context file to be included with the natural language understanding's output data. The output data may be provided to a corresponding application for causing one or more actions to be performed.