Crowdsourced Digital Assistant Dialect Adaptation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional digital assistants face issues such as privacy concerns, misinterpretation of spoken commands, and limited dialect support, leading to user frustration due to their reliance on fixed algorithms and signal availability.
Innovation Solution
A crowd-sourced digital assistant system that creates and distributes action datasets for computing devices, allowing for intuitive command recognition and execution, with a server-driven selection mechanism that learns from user interactions and adapts to various dialects, reducing misinterpretation and enhancing privacy by utilizing existing device applications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional digital assistants use fixed algorithms for command interpretation, then device complexity is reduced and ease of manufacture is improved, but command interpretation accuracy deteriorates and dialect support is limited
Solution Approach 1:
The patent introduces a server as an intermediary between users and digital assistants. The server maintains and distributes dialect libraries and training data, allowing individual devices to benefit from improved command interpretation without storing all the complexity locally. Devices receive updated dialect models and training data from the server, resolving the contradiction between accuracy and device complexity.
Solution Approach 2:
The server provides universal dialect support across multiple devices. Instead of each device needing its own complete dialect library, the server maintains a universal collection that can be distributed to any device that needs it. This allows accurate multi-dialect support while keeping individual device complexity low.
2Adaptability or versatility
If digital assistants collect user data for training, then dialect support and command understanding improve, but privacy concerns increase
Solution Approach 1:
The server acts as a trusted intermediary that collects, processes, and aggregates training data from multiple users. Individual user data remains protected while the server creates generalized dialect models that improve all devices. This resolves the privacy concern by preventing individual devices from accessing raw user data while still enabling dialect adaptation.
Solution Approach 2:
Instead of sharing actual user data between devices, the system creates copies in the form of aggregated statistical models and training data representations. The server processes original data to create anonymized training sets and dialect models that can be distributed safely, maintaining privacy while enabling adaptation.
3Measurement precision
If digital assistants use centralized server processing, then command interpretation accuracy improves, but response time increases due to network dependency
Solution Approach 1:
The system performs preliminary processing by pre-training dialect models and storing them on the server. When a device needs dialect support, it receives pre-processed models and training data in advance, rather than processing raw data in real-time. This reduces response time while maintaining accuracy.
Solution Approach 2:
The system segments processing into offline batch processing (on the server) and online real-time processing (on devices). The server handles heavy computational tasks like data aggregation and model training offline, while devices perform lighter real-time inference tasks, reducing network dependency during actual command interpretation.
Data Source
AI summary
Embodiments described herein are generally directed towards systems and methods relating to a crowd-sourced digital assistant system and related methods. In particular, embodiments facilitate the intuitive creation, maintenance, and distribution of action datasets that include computing events or tasks that can be reproduced when an associated command is determined received by a digital assistant device. In various implementations, multiple action datasets may be determined associated with a received command and, as such, the digital assistant device or the digital assistant server can determine one or more action datasets that are most relevant to a particular user of the digital assistant device based on contextual data collected by the digital assistant device. In further implementations, the collected contextual data can be maintained by the digital assistant device, the digital assistant server, or a combination thereof.