Hybrid NLU System Combining Rule-Based and Statistical Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current natural language understanding (NLU) systems face challenges in balancing cost, speed of development, and control, as rule-based models offer control but are costly and time-consuming, while statistical models are cheaper but harder to modify and lack context control.
Innovation Solution
A hybrid NLU system combining a rule-based module and a statistical module, with a decider to determine the final intent based on outputs from both, allowing for efficient construction and adaptation without sacrificing accuracy or control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If rule-based models are used for NLU, then control and accuracy are improved, but development cost and time increase
Solution Approach 1:
The patent combines rule-based NLU models and statistical NLU models into a hybrid system. The rule-based component provides control and accuracy for specific intents, while the statistical component handles general patterns, achieving both reliability and efficiency simultaneously.
Solution Approach 2:
The NLU system is segmented into multiple specialized models organized in a tree structure, where different models handle different domains or tasks. This allows selective application of rule-based or statistical approaches for different intents, optimizing both control and development efficiency.
2Productivity
If statistical models are used for NLU, then development cost and time are reduced, but control and adaptability decrease
Solution Approach 1:
The system dynamically selects between rule-based and statistical models based on the input context and intent. The hierarchical structure allows the system to adaptively choose the appropriate modeling approach for each specific NLU task, maintaining both speed and control.
Solution Approach 2:
The patent introduces an intermediary layer that bridges rule-based and statistical approaches. This intermediary structure enables the system to leverage the speed of statistical models while maintaining the control capabilities of rule-based models through the hierarchical organization of specialized models.
3Ease of operation
If rule-based models are used, then context control and conversation flow management are improved, but maintenance cost increases
Solution Approach 1:
The rule-based language model is segmented into multiple specialized models organized in a tree structure, where each model handles specific domains or tasks. This segmentation reduces the complexity of maintaining rules, as changes can be made in isolated branches without affecting the entire system.
Solution Approach 2:
Different parts of the NLU system use different modeling approaches tailored to their specific needs. Rule-based models are applied locally where context control is critical, while statistical models handle areas where flexibility is more important, optimizing both operational ease and maintenance cost.
Data Source
AI summary
Hybrid natural language understanding (NLU) systems and methods are provided that capitalize on the strengths of the rule-based models and the statistical models, lowering the cost of development and increasing the speed of construction, without sacrificing control and accuracy. Two models are used for intent recognition, one statistical and one rule-based. Both models define the same set of intents, but the rule-based model is devoid of any grammars or patterns initially. Each model may or may not be hierarchical in that it may be composed of a set of specialized models that are in a tree form or it may be just a singular model.