Speech Interaction with Context for Incomplete Commands
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Voice assistants struggle to accurately identify user intents and slots in non-speech instructions, leading to reduced user experience when interacting with electronic devices.
Innovation Solution
A server determines intents and slots from non-speech instructions, using pre-stored information about audio content to enhance intent and slot extraction, even when speech instructions are incomplete.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the voice assistant uses only speech instructions for interaction, then the system can accurately extract intents and slots from complete speech commands, but the system cannot respond when users provide incomplete speech instructions or use non-speech control methods
Solution Approach 1:
The system performs preliminary action by storing extracted intents and slots from complete speech instructions in a context buffer before users provide subsequent incomplete instructions. This allows the system to have contextual information ready in advance, enabling accurate interpretation of later incomplete or ambiguous speech inputs without requiring users to repeat complete commands.
Solution Approach 2:
The system implements feedback by using previously extracted contextual information to supplement and complete current speech instructions. When the system receives an incomplete instruction, it feeds back the stored context (intents and slots from prior interactions) to fill in missing information, thereby maintaining high extraction accuracy even with incomplete user input.
2Measurement precision
If the system stores context from previous speech instructions, then the system can accurately interpret incomplete current instructions, but the system complexity increases due to context management requirements
Solution Approach 1:
The system applies segmentation by dividing context management into distinct functional modules: one module extracts intents and slots from speech instructions, another module stores them in a structured context buffer, and a third module retrieves and applies relevant context to current instructions. This modular segmentation reduces overall system complexity by making each component's responsibility clear and manageable.
Solution Approach 2:
The system uses parameter changes by dynamically adjusting the weight and relevance of stored contextual information based on current interaction state. The context buffer tracks parameters such as instruction completeness, temporal recency, and semantic relevance, automatically adjusting which contextual elements are applied to supplement current instructions, thereby simplifying complex context management through parameter-based control.
Data Source
AI summary
A speech interaction method includes receiving, by a server, a first play message, where the first play message includes an identifier of first audio content corresponding to a first non-speech instruction. The server determines a first intent and first slot information that correspond to the first non-speech instruction. In response to the first play message, the server instructs a playback device to play the first audio content. The server receives a first speech instruction input by a user into the playback device, where a second intent or second slot information or both in the first speech instruction are incomplete. The server determines, based on the first intent and the first slot information, the second intent and the second slot information that correspond to the first speech instruction, and the server, based on the second intent and the second slot information, instructs the playback device to play second audio content.