AI Speech Recognition Using Usage Logs for Implicit Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current speech recognition technologies fail to accurately recognize user speech, leading to user dissatisfaction and reduced performance due to incorrect recognition, especially when explicit feedback is not provided.
Innovation Solution
An AI apparatus and method that utilizes a user's application usage log as implicit feedback to improve speech recognition performance by updating the language model, reflecting specific pronunciation, language habits, and dialects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If speech recognition is performed using a language model without considering application usage logs, then the system operation is simple, but the speech recognition accuracy deteriorates when explicit feedback is not provided
Solution Approach 1:
The patent implements feedback by utilizing application usage logs as implicit feedback signals. When speech recognition fails or intention determination is unsuccessful, the system retrieves application usage logs to infer the user's intended meaning and uses this information to update the language model, thereby improving future recognition accuracy without requiring explicit user feedback.
Solution Approach 2:
The system performs self-service by automatically updating its own language model using application usage logs. The AI apparatus independently identifies recognition failures, retrieves relevant usage log data, and triggers model updates without external intervention, enabling continuous self-improvement of speech recognition performance.
2Reliability
If the system requests users to utter again when speech is not recognized correctly, then the speech recognition opportunity is preserved, but the user satisfaction level deteriorates significantly
Solution Approach 1:
The system uses application usage logs as implicit feedback to understand user intent when speech recognition fails. By analyzing what applications the user is using or has recently used, the system can infer the intended command or query and provide appropriate responses, eliminating the need to ask users to repeat themselves and thereby maintaining high user satisfaction.
Solution Approach 2:
The application usage log acts as an intermediary that bridges the gap between failed speech recognition and user intent. Instead of directly confronting the recognition failure by asking the user to repeat, the system uses the usage log as a mediator to indirectly determine what the user intended to say, providing a smoother user experience.
3Measurement precision
If speech recognition fails once and the user does not provide explicit feedback, then the system operation remains simple, but the speech recognition performance deterioration occurs because it is difficult to improve
Solution Approach 1:
The patent transforms application usage logs into implicit feedback information. When explicit feedback is unavailable, the system analyzes application usage patterns to infer user intent, effectively converting passive usage data into active feedback signals that drive language model updates and improve speech recognition performance over time.
Solution Approach 2:
The application usage log serves as an intermediary information source that compensates for the absence of explicit user feedback. By using this intermediary data, the system can still identify recognition failures and trigger appropriate model updates, preventing performance deterioration even when users remain silent.
4Measurement precision
If the language model is updated using application usage logs, then the speech recognition accuracy is improved, but the data processing complexity increases
Solution Approach 1:
The system performs preliminary actions by continuously monitoring and storing application usage logs in advance. When speech recognition fails or intention determination is unsuccessful, the required feedback data is already available in the usage logs, eliminating the need for complex real-time data collection and processing during the update trigger moment.
Solution Approach 2:
The system automatically manages the complexity of data processing by implementing self-service mechanisms. The AI apparatus independently determines when updates are needed, selectively retrieves relevant usage log data, and triggers model updates without requiring complex external coordination or manual intervention, thereby managing data processing complexity through automated decision-making.
Data Source
AI summary
Provided is an artificial intelligence (AI) device for recognizing speech of user. The AI apparatus includes: a microphone; and a processor configured to: receive, via the microphone, a sound signal corresponding to speech of the user, recognize the speech from the sound signal using a language model, determine an intention of the user based on the recognition result, determine whether the determination of the intention is successful, obtain a user's application usage log if the determination of the intention is not successful, and update the language model using the obtained user's application usage log.