Smart Device Utterance Generation Using Random Semantic Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current smart devices generate monotonous responses to user queries, leading to a lack of engagement in human-machine interactions, as they typically provide fixed answers to the same utterances, resulting in a repetitive conversation experience.
Innovation Solution
A computer-implemented method using a generative adversarial network to generate diverse answers by inputting random semantic vectors alongside user utterances, allowing the smart device to provide varied responses even to identical queries, enhancing interaction engagement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a fixed answer is provided for the same utterance, then the response consistency is maintained, but the human-computer interaction becomes monotonous and less engaging
Solution Approach 1:
The system dynamically selects between fixed answers and generated answers based on the utterance type. For common utterances, fixed answers provide consistency; for other utterances, the generator creates diverse responses. This dynamic adaptation resolves the contradiction between response consistency and interaction diversity.
Solution Approach 2:
The system changes the response generation parameter by introducing a randomness factor through the generator when fixed answers are not applicable. This parameter change allows the system to maintain consistency for standard queries while providing diversity for varied interactions, resolving the contradiction between reliability and adaptability.
2Adaptability or versatility
If a generator based on generative adversarial network is used, then the diversity of answers is improved, but the device complexity increases
Solution Approach 1:
The system segments the response generation process into two parts: a fixed answer module for common utterances and a generator module for other cases. This segmentation allows the complex generator to be used only when necessary, reducing overall system complexity while maintaining answer diversity where needed.
Solution Approach 2:
The system introduces an intermediary mechanism (the utterance analysis module) that determines whether to use fixed answers or invoke the generator. This intermediary layer manages the complexity by filtering when the sophisticated generative model needs to be activated, reducing the practical complexity burden while preserving answer diversity capabilities.
3Adaptability or versatility
If random semantic vectors are generated, then the variety of responses increases, but the computational resources required increase
Solution Approach 1:
The system dynamically activates random semantic vector generation only when fixed answers are not applicable. For common utterances, no generation occurs and energy is saved. For varied utterances, the generator is activated to provide response variety. This dynamic approach resolves the contradiction between response variety and computational energy consumption.
Data Source
AI summary
A computer-implemented method for utterance generation, a smart device, and a non-transitory computer readable storage medium are provided. The method includes: obtaining a first utterance to be answered, generating at least one random semantic vector, inputting the at least one random semantic vector and the first utterance into a trained generator, and obtaining at least one first answer outputted by the trained generator, wherein the trained generator is obtained based on a preset generative adversarial network. Due to the random semantic vector, even for the same utterance, the smart device can generate different answers corresponding to the different random semantic vectors, the possibility of generating too many identical answers during the human-machine conversation is reduced, and the fun during the human-machine conversation is enhanced.