Deep Learning Reply Generation via External Components for Intent Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing intelligent systems have limited capabilities in processing user input data, resulting in poor quality of generated reply content.
Innovation Solution
A deep learning model is used to determine whether to call an external functional component, generate an intermediate inquiry recognizable by that component, and then call the component to obtain a result, enhancing the model's capability to generate replies that better meet user intentions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a deep learning model directly generates replies based on user input, then the system structure is simple, but the quality of generated replies is poor
Solution Approach 1:
The patent introduces an external functional component as an intermediary between the deep learning model and the final reply generation. The model generates intermediate inquiries that are passed to the external component, which processes them and returns results that are then used to generate the final reply. This intermediary structure improves reply quality without requiring complete rearchitecture of the system.
Solution Approach 2:
The reply generation process is segmented into multiple stages: (1) the deep learning model generates intermediate inquiries from user input, (2) the external functional component processes these inquiries, (3) the model generates final replies based on processed results. This segmentation allows each component to specialize, improving overall output quality while maintaining manageable system complexity.
2Manufacturing precision
If the deep learning model generates intermediate inquiries recognizable by external functional components, then the reply quality improves, but the system complexity increases
Solution Approach 1:
The deep learning model is designed with multi-functionality: it serves both as a traditional reply generator and as a query generator for external functional components. By making the model universal, the patent avoids adding separate specialized components, thus improving reply quality through enhanced functionality while minimizing increases in system complexity.
3Reliability
If the system calls external functional components to process intermediate inquiries, then the intelligence and responsiveness of the system improve, but the processing time increases
Solution Approach 1:
The deep learning model performs preliminary action by generating well-structured intermediate inquiries that are optimized for external functional components. This preliminary processing ensures that external components receive precisely formatted inputs, enabling faster and more efficient processing, thus reducing the time penalty associated with calling external components while maintaining high system intelligence.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present disclosure provides a data generation method based on a deep learning model, and a training method and apparatus, relates to the field of artificial intelligence technologies, and in particular, to the field of natural language processing and deep learning technologies, and can be used to improve the quality of reply data generated by the deep learning model based on input data of a user. The data generation method includes: determining an initial input of the deep learning model based on input data of a user; obtaining a first output of the model, where in response to the model determining that generating a reply based on the initial input requires calling a first functional component different from the deep learning model, the first output includes a first token for calling the first functional component and a first intermediate inquiry determined based on the initial input and recognizable by the first functional component; obtaining a first intermediate result determined by the first functional component based on the first intermediate inquiry; determining a second input for the model based on the initial input and the first intermediate result; and obtaining a second output of the model for generating a reply to the initial input.