Intent Mining via Semantic Similarity for Bot Authoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current customer service systems face challenges in automating bot authoring and intent mining from natural language conversations, requiring manual processes that are time-consuming and costly, and lack efficient methods to analyze and understand customer intents.
Innovation Solution
A computer-implemented method for authoring conversational bots using intent seeding, which involves receiving conversation data, seed intent data, and employing an intent mining algorithm to automatically identify and associate new utterances with seed intents based on semantic similarity, thereby augmenting the bot's capabilities for automated conversations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual processes are used for bot authoring and intent mining, then accuracy and control over intent identification can be maintained, but time consumption and cost increase significantly
Solution Approach 1:
The system enables self-service by allowing the bot authoring process to automatically mine intents and utterances from conversation data without requiring manual intervention. The intent mining algorithm autonomously processes conversations, identifies intents, and populates the intent database, making the system serve itself rather than requiring human operators for each mining operation.
Solution Approach 2:
The patent replaces the mechanical manual process of intent mining with an automated computational system. The intent mining algorithm uses natural language processing and machine learning techniques to automatically analyze conversations, extract intents, and associate utterances with intents, substituting human manual work with automated computational mechanisms.
2Productivity
If manual intent mining is performed, then precise control over intent identification can be achieved, but development cost increases
Solution Approach 1:
The system achieves universality by creating a multi-functional platform that combines conversation data processing, intent mining, utterance extraction, semantic similarity calculation, and bot authoring support into a single integrated system. This universal platform can handle multiple tasks across different bot projects, reducing the need for separate tools and manual processes for each function.
Solution Approach 2:
The patent applies segmentation by breaking down the complex bot authoring process into distinct modular components: conversation data ingestion, intent identification, utterance extraction, semantic similarity computation, and intent database population. Each module operates independently but connects to form a complete workflow, allowing for easier maintenance, scaling, and customization while reducing overall system complexity through modular design.
3Loss of time
If automated intent mining is implemented, then development time is reduced, but the system requires sophisticated algorithms and processing capabilities
Solution Approach 1:
The system uses semantic similarity as an intermediary mechanism to bridge the gap between raw conversation data and structured intent identification. Instead of directly analyzing complex natural language to detect intents, the system calculates semantic similarity between conversation utterances and known intent examples, using this similarity measure as a mediator to automatically classify and identify intents with high accuracy without requiring complex direct detection algorithms.
4Measurement precision
If comprehensive conversation analysis is performed to understand customer intents, then intent accuracy improves, but processing time and computational resources increase
Solution Approach 1:
The system applies partial action by focusing the analysis on the most critical aspects of conversation data necessary for intent identification. Rather than performing exhaustive analysis of every linguistic feature, the system selectively processes key elements such as semantic similarity to known intents, essential conversation patterns, and high-value utterances. This partial processing approach achieves sufficient intent identification accuracy while maintaining high processing speed and avoiding unnecessary computational overhead.
Data Source
AI summary
A method for authoring a conversational bot including: receiving conversation data; receiving seed intent data that comprises seed intents having a seed intent label and sample intent-bearing utterances; using an intent mining algorithm to mine the conversation data to determine new utterances to associate with the seed intent; augmenting the seed intent data to include the mined new utterances associated with the seed intents; and uploading the augmented seed intent data into the conversation bot. The intent mining algorithm may include: identifying intent-bearing utterances; identifying candidate intents; for each of the seed intents, identifying seed intent alternatives from the sample intent-bearing utterances; associating the intent-bearing utterances from the conversation data with the seed intents via determining a degree of semantic similarity between the candidate intents of the intent-bearing utterances and the seed intent alternatives.