Persona-Adaptive Assistant Outputs With Dynamic Visual Cues
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automated assistants lack visual richness in their interactions, providing static visual content that fails to replicate the dynamic and expressive nature of human communication.
Innovation Solution
Implementations enable automated assistants to dynamically adapt their outputs based on assigned personas, incorporating personalized textual content and visual cues, including animations and gestures, through the use of large language models (LLMs) to enhance dialog sessions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If automated assistants use static visual content to confirm audible information, then the system is simple and easy to implement, but the visual content lacks visual richness and cannot replicate human communication expressiveness
Solution Approach 1:
The patent applies dynamics by transitioning from static visual content to dynamic visual representations that can change based on context. The system now generates visual content that adapts in real-time to match the persona and conversational context, enabling expressions like facial emotions, body language, and animated gestures that mirror human communication patterns.
Solution Approach 2:
The system changes parameters by introducing persona-specific attributes (vocabulary, prosodic properties, visual cues) that modify the generated content. By varying these parameters based on the assigned persona, the automated assistant can produce diverse visual outputs ranging from formal to casual expressions, enhancing visual richness without requiring a complete system redesign.
2Adaptability or versatility
If automated assistants use hard-coded rules for visual representations, then the implementation is straightforward, but the range of visual expression is limited
Solution Approach 1:
The patent replaces the mechanical hard-coded rule system with an AI-driven generative model. Instead of predefined if-then rules, the system uses large language models to dynamically generate visual content based on contextual understanding, enabling a much broader range of expressions while maintaining implementation feasibility through standardized AI interfaces.
Solution Approach 2:
The system transitions from static hard-coded rules to dynamic generation where visual expressions are created in real-time based on the conversation context and persona attributes. This allows the automated assistant to adapt its visual expression range to match the specific interaction scenario.
3Adaptability or versatility
If automated assistants provide comprehensive visual content with animations and gestures, then visual richness is enhanced, but computational and network resources are consumed
Solution Approach 1:
The patent applies local quality by generating visual content selectively based on the specific interaction context. Rather than always providing comprehensive visual content, the system adjusts the level of visual detail and complexity to match the conversational needs, reducing resource consumption when full visual richness is not necessary while maintaining high quality when appropriate.
Solution Approach 2:
The system uses partial action by generating only the necessary visual content required for effective communication. The generative model can produce simplified visual representations when the context allows, reserving more computationally intensive detailed animations and gestures for situations where they add significant value to the interaction.
Data Source
AI summary
Implementations relate to dynamically adapting a given assistant output based on a given persona, from among a plurality of disparate personas, assigned to an automated assistant. In some implementations, the given assistant output can be generated and subsequently adapted based on the given persona assigned to the automated assistant. In other implementations, the given assistant output can be generated specific to the given persona and without having to subsequently adapt the given assistant output to the given persona. Notably, the given assistant output can include a stream of textual content to be synthesized for audible presentation to the user, and a stream of visual cues utilized in controlling a display of a client device and/or in controlling a visualized representation of the automated assistant. Various implementations utilize large language models (LLMs), or output previously generated utilizing LLMs, to reflect the given persona in the given assistant output.