Dialog State Tracking Using String Similarity and Coreference Rules
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional dialog state tracking systems face challenges in accurately identifying and selecting slot-value pairs from natural language inputs, especially when users do not provide exact words or phrases, and when there are many potential slot-value pairs to choose from, leading to incorrect or missing responses.
Innovation Solution
The implementation of rule-based dialog state tracking that employs string-matching, coreference resolution, and carryover rules to enhance the identification and selection of slot-value pairs, including the use of synonyms and templates to handle nuances in natural language.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional dialog state tracking systems use exact word matching to identify slot-value pairs, then the system structure remains simple, but the system fails to accurately identify slot-value pairs when users do not provide exact words or phrases
Solution Approach 1:
The system changes the matching parameter from exact word matching to fuzzy matching based on string similarity. This allows the system to identify slot-value pairs even when users do not provide exact words, thereby improving identification accuracy without requiring fundamentally new system architecture
Solution Approach 2:
The patent introduces string similarity scores as an intermediary mechanism between user input and slot-value pair identification. This intermediary allows for flexible matching by comparing the similarity between user utterances and predefined slot values, resolving the contradiction between simplicity and accuracy
2Adaptability or versatility
If the system expands the ontology to include many potential slot-value pairs to handle diverse user inputs, then the system becomes more versatile, but the difficulty of choosing the correct slot-value pairs increases
Solution Approach 1:
The system implements feedback through confidence scores and similarity thresholds. When multiple slot-value pairs match a user input, the system selects the one with the highest string similarity score, providing a mechanism to accurately choose among many potential matches and maintain precision despite increased versatility
Solution Approach 2:
The patent changes the selection parameter from simple presence/absence matching to similarity-based ranking. By calculating and comparing similarity scores across multiple potential slot-value pairs, the system can accurately select the correct match even when the ontology contains many similar options
3Reliability
If conventional systems use simple matching rules, then the system remains easy to operate, but the system produces incorrect or missing responses when users provide ambiguous or varied inputs
Solution Approach 1:
The patent segments the matching process into distinct rule-based components: exact match rules, fuzzy match rules, and coreference resolution rules. This segmentation allows the system to handle different types of user inputs with appropriate rules, improving reliability while keeping each individual rule component relatively simple and manageable
Data Source
AI summary
The present disclosure relates dialog states, which computers use to internally represent what users have in mind in dialog. A dialog state tracker employs various rules that enhance the ability of computers to correctly identify the presence of slot-value pairs, which make up dialog states, in utterances or conversational input of dialog. Some rules provide for identifying synonyms of values of slot-values pairs in utterances. Other rules provide for identifying slot-value pairs based on coreferences between utterances and previous utterances of dialog sessions. Rules are also provided for carrying over slot-value pairs from dialog states of previous utterances to a dialog state of a current utterance. Yet other rules provide for removing slot-value pairs from candidate dialog states, which are later used as dialog states of utterances.