AI Intent Classification with GAN-Based Text Refinement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional interactive artificial intelligence systems fail to accurately distinguish between similar-text inputs with different intents, leading to irrelevant responses, causing inconvenience to users.
Innovation Solution
An electronic device equipped with a first artificial intelligence model that classifies user inputs based on similarity to pre-stored texts, providing error information when the intent differs, and a second AI model generates new texts with higher similarity, iteratively improving classification accuracy using generative adversarial networks (GAN).
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If conventional interactive artificial intelligence systems use text similarity matching to provide responses, then the system can automatically respond to user requests, but the system fails to distinguish between similar-text inputs with different intents, leading to irrelevant responses
Solution Approach 1:
The patent divides the text processing into separate components: a first AI model for initial classification and a second AI model for generating similar texts. This segmentation allows each model to specialize in specific tasks, improving overall intent recognition accuracy while maintaining automation.
Solution Approach 2:
The patent implements a feedback mechanism where the second AI model generates similar texts based on the first model's classification, and these generated texts are fed back into the system for re-evaluation. This iterative feedback loop continuously refines intent recognition accuracy.
2Adaptability or versatility
If the system stores vast amounts of text in a database for matching, then the system can handle diverse user requests, but the system cannot distinguish between texts that are similar in form but have different subjects or purposes
Solution Approach 1:
The patent introduces an intermediary layer in the form of a second AI model that generates similar texts as intermediaries between the user input and the database matching process. This intermediary layer enriches the text representation with contextual information about similar phrases, enabling the system to distinguish between texts that are similar in form but have different meanings.
3Speed
If the system provides response information based on similar text matching, then the system can respond quickly to user commands, but the system provides irrelevant responses when the user input has a different intent than stored texts
Solution Approach 1:
The patent performs preliminary action by using the first AI model to classify the user input and generate similar texts before actually querying the database. This preliminary classification step filters out irrelevant matches and prepares relevant text candidates in advance, ensuring that when the database is queried, only relevant responses are retrieved, thus maintaining both speed and reliability.
Data Source
AI summary
An electronic device is provided. The electronic device includes a memory configured to store a computer executable instructions; and a processor configured to execute the executable instructions to: determine a text corresponding to a received command, provide response information on the command based on a first artificial intelligence model classifying the text as a text corresponding to one of a plurality of pre-stored texts, and provide error information on the command based on the first artificial intelligence model classifying the text as an error, wherein the first artificial intelligence model is configured to classify the text as the error based on the text corresponding to the command being a similar text having one of an entity and an intent different from at least one of the plurality of pre-stored texts.