Voice Assistant Personalization via User-Specific Sub-Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Voice assistants shared among multiple users often fail to provide personalized recommendations due to combined user preferences, leading to suboptimal content suggestions, and do not consider varying environments or circumstances.
Innovation Solution
An information processing system with an audio processing device and a response system that identifies users through voiceprint analysis, manages playback history, and recommends content based on similar user preferences, while also accounting for environmental circumstances.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a single AI model is used to serve multiple users in a shared environment, then the device complexity is reduced and ease of operation is improved, but the recommendation precision deteriorates because the model cannot distinguish between individual user preferences
Solution Approach 1:
The patent segments the unified AI model into user-specific sub-models by extracting and storing individual user preference features from playback history. Each user is assigned a unique identification information, and the system maintains separate preference profiles for each user, enabling personalized recommendations while using a single overall system architecture.
Solution Approach 2:
The patent applies local quality by customizing the AI model's behavior for each user based on their specific preference features extracted from playback history. The system adjusts recommendation parameters and content selection locally for each user while maintaining the same overall system framework, thus achieving personalized service without requiring separate complete systems.
2Measurement precision
If the AI model stores and processes detailed playback history information for each user, then the recommendation precision is improved, but the device complexity and information management burden increase
Solution Approach 1:
The patent extracts key preference features from the comprehensive playback history data and stores only these extracted features in association with user identification information. Instead of storing complete playback histories, the system extracts and retains essential preference patterns, reducing data storage requirements and processing complexity while maintaining recommendation accuracy.
Solution Approach 2:
The patent applies partial action by processing and storing only the necessary portion of playback history information - specifically the preference features - rather than all detailed playback data. This selective processing approach reduces the information management burden while preserving the essential data needed for accurate recommendations.
3Adaptability or versatility
If the voice assistant uses a unified AI model for all users, then the adaptability to different environments is reduced, but the ease of operation and device simplicity are improved
Solution Approach 1:
The patent implements dynamics by making the AI model adaptable to different users and environments through dynamic adjustment of preference features. The system automatically adjusts its behavior based on identified user characteristics and playback history, enabling the same model to adapt to various users' needs without requiring separate static models for each user or environment.
Data Source
AI summary
A voice assistant is provided whereby, in response to receipt of a voice instruction for requesting a recommended piece of content from a user, the voice assistant receives, from a system, a response including information on a recommended piece of content determined using another playback history having a feature similar to a feature of playback histories stored in association with a user ID of the user, and outputs the piece of content.