Hierarchical Command Graph for Faster Chatbot Intent Recommendations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing chatbot systems face computational inefficiencies and inaccuracies in interpreting user commands due to imprecise natural language inputs, requiring extensive processing to determine the intended action.
Innovation Solution
A system that maps hierarchical intents to canonical actions, reformats this into a graph representation, and uses machine learning to recommend contextually or semantically similar actions based on a ranked count of executed actions, reducing computational load while improving recommendation quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the system performs extensive natural language processing and intent interpretation to handle imprecise user commands, then the accuracy of command interpretation is improved, but the computational cost and processing time increase significantly
Solution Approach 1:
The system pre-processes and stores hierarchical intent relationships in a graph structure before user input arrives. By establishing the intent hierarchy and canonical action mappings in advance, the system avoids performing extensive real-time NLP processing for each user command, instead relying on pre-built structural knowledge to quickly determine intended actions.
Solution Approach 2:
The patent introduces an intermediary graph representation that mediates between raw user input and final action execution. This graph structure serves as a middle layer that translates imprecise natural language queries into standardized canonical actions by leveraging pre-established hierarchical relationships, reducing the need for computationally intensive real-time interpretation.
2Measurement precision
If the system performs extensive natural language processing to determine user intent, then the accuracy of intent recognition is improved, but the processing time increases
Solution Approach 1:
The system pre-establishes the hierarchical intent graph and action mappings before user interaction. This preliminary structuring allows the system to quickly navigate through pre-defined relationships rather than performing time-consuming real-time analysis, significantly reducing processing time while maintaining accurate intent recognition through the pre-built hierarchical knowledge.
Solution Approach 2:
The patent creates a simplified graph representation that copies and preserves the essential hierarchical relationships between intents and actions. This copied structure enables fast lookup and matching operations during user interaction, avoiding the need for complex real-time computation while maintaining the semantic relationships necessary for accurate intent recognition.
3Productivity
If the system uses a graph representation of hierarchical intents, then the efficiency of command recommendation is improved, but the system complexity increases
Solution Approach 1:
The patent segments the intent system into hierarchical levels, where higher-level intents represent broader categories and lower-level intents represent specific actions. This segmentation organizes the complexity into manageable layers, allowing the graph structure to efficiently represent relationships without becoming unmanageably complex. The hierarchical breakdown enables systematic navigation and recommendation generation.
Solution Approach 2:
The graph representation serves multiple functions simultaneously: it stores hierarchical intent relationships, maps intents to canonical actions, enables efficient recommendation generation, and provides a structure for handling imprecise user input. By making the graph multi-functional, the patent reduces overall system complexity despite the added structural element, as one component performs multiple critical roles.
Data Source
AI summary
Systems, methods, and computer-readable storage media for recommending computational commands based on a weighted, ranked hierarchical graph. The system converts a map of intended user actions to canonical actions (commands) into a graph representation. Then, via a chatbot algorithm, the system receives an initial intent as a request by a user of the chatbot algorithm. Based on the popularity of alternative canonical actions, the system makes recommendations of alternative actions which are similar to the initial intent provided by the user.