Command Processing System Using Preliminary Speechlet Queries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speech processing systems face challenges in selecting the appropriate speechlet to respond to user inputs, often leading to inefficiencies due to latency and resource conservation issues, as they select a speechlet before obtaining potential responses, which may result in inappropriate or unable-to-execute selections.
Innovation Solution
The system improves by sending NLU results to multiple speechlets and obtaining partial output content before selecting, using post-result scores to ensure the most appropriate speechlet is chosen, allowing for potential results without affirmative execution, thus mitigating latency and resource waste.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If the system selects a speechlet before obtaining potential responses, then the selection process is faster, but the selected speechlet may be inappropriate or unable to execute the command
Solution Approach 1:
The system performs preliminary actions by sending NLU results to multiple speechlets simultaneously and obtaining potential responses before making the final selection. This allows the system to evaluate which speechlet can actually execute the command based on their responses, rather than guessing beforehand. The preliminary retrieval of potential results from multiple speechlets resolves the contradiction by enabling informed selection without significant time penalty.
2Reliability
If the system queries multiple speechlets for potential results, then the correct speechlet can be identified, but additional processing time and resources are consumed
Solution Approach 1:
The system applies partial action by querying multiple speechlets for potential results but not necessarily executing all of them. Instead, it uses the potential results information to select the most appropriate speechlet for actual execution. This partial querying approach provides enough information to make an accurate selection without the full overhead of engaging all speechlets, thus resolving the contradiction between reliability and productivity.
3Use of energy by moving object
If the system selects a speechlet without obtaining potential responses, then resource consumption is lower, but latency increases due to re-selection
Solution Approach 1:
The system performs a lightweight preliminary action by sending NLU results to multiple speechlets and obtaining potential responses before selection. This preliminary step consumes minimal additional resources compared to the cost of re-selection latency, as it only requires sending queries and receiving responses without full execution. The small initial resource investment prevents time loss from re-selection, resolving the contradiction between resource usage and latency.
Data Source
AI summary
Techniques for performing command processing are described. A system receives, from a device, input data corresponding to a command. The input data may originate as audio data, as text data, or as other data. The system determines NLU processing results corresponding to the input data. The NLU processing results may be associated with multiple speechlets. The system also determines NLU confidences for the NLU processing results for each speechlet. The system sends NLU processing results and an indication to provide potential results to a portion of the multiple speechlets, and receives potential results from the portion of the speechlets. The system also receives indications whether the speechlets need to be re-called if the speechlets are selected to execute with respect to the command. The system ranks the portion of the speechlets based at least in part on the NLU processing results as well as the potential results provided by the portion of the speechlets. The system may also rank the portion of the applications using other data. The system causes content provided by the highest ranked speechlet to be output to a user.