Entity Resolution in Conversational Applications
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current web search technologies face challenges in providing precise results due to ambiguous queries and the inability to handle multiple concepts, leading to irrelevant content and limitations in querying structured databases, which fail to respond to various surface forms of item descriptions.
Innovation Solution
The development of a system that uses entity normalization techniques, including lexical, syntactic, and contextual processing, to identify and classify entities within user input, and maps ambiguous representations to canonical values, enabling interaction with structured knowledge sources and applications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If query expansion techniques are used to address alternate word forms, then the search engine can match additional web pages, but the search results become less precise and include hyperlinks beyond the concept of interest
Solution Approach 1:
The query is segmented into multiple candidate concepts, each representing a possible interpretation of the ambiguous terms. The system then evaluates each concept separately using structured knowledge sources to determine which concept best matches the user's intent, thereby maintaining precision while handling alternate word forms.
Solution Approach 2:
The system changes the parameter of query interpretation by using multiple candidate concepts instead of a single expanded query. By evaluating each concept against structured knowledge sources and selecting the best match, the system achieves both adaptability to alternate word forms and precision in results.
2Productivity
If web services employ structured databases or catalogs to categorize items, then they can efficiently serve users for specific domains, but users are limited to submitting query terms that comply with the categorization schema
Solution Approach 1:
The system introduces an intermediary layer between the user's natural language query and the structured database. This intermediary uses entity normalization techniques to map ambiguous surface forms to canonical values in the structured knowledge source, thereby maintaining service efficiency while accepting flexible natural language queries.
Solution Approach 2:
The system performs preliminary entity normalization on the query before submitting it to the structured database. By pre-processing the query to resolve ambiguous terms and map them to canonical values, the system enables users to submit flexible natural language queries while maintaining efficient database querying.
3Measurement precision
If named entity normalization is used to link surface forms to canonical names, then ambiguous entity references can be resolved, but the system requires structured knowledge sources that may not cover all surface forms
Solution Approach 1:
The system uses multiple structured knowledge sources with different categorization schemes to resolve entity references. By querying multiple knowledge sources and aggregating results, the system achieves both high entity resolution accuracy and broad coverage of surface forms that may not be present in any single knowledge source.
Solution Approach 2:
The system creates a composite view by combining results from multiple structured knowledge sources. This composite approach allows the system to resolve ambiguous entity references with high accuracy while covering a broader range of surface forms than any individual knowledge source could provide alone.
4Ease of operation
If users query web services with natural language, then they can use various surface forms to describe items, but the query may not conform to the structured database's categorization schema
Solution Approach 1:
The system introduces an intermediary entity normalization process that translates natural language queries into structured database queries. This intermediary maps ambiguous surface forms to canonical values using structured knowledge sources, thereby maintaining query convenience while ensuring reliable results that conform to the database schema.
Solution Approach 2:
The system performs preliminary normalization of natural language queries before submitting them to the structured database. By pre-processing the query to resolve entities and map them to canonical values, the system enables users to query conveniently in natural language while ensuring reliable results through proper schema conformance.
Data Source
AI summary
User input expressed as text may be analyzed for determining a type of response, such as an application response, and/or determining a type of task that is requested by the user input. Entity representations may be identified, classified and/or or tagged based on a type of response, type of task and/or a set of entity types. A surface form of an entity, ambiguous entity representation and/or other type of expression within the user input may be resolved, normalized and/or mapped to a normalized value. Normalizing entities and/or entity attributes may involve using a set of normalization rules, a lookup table, one or more machined learned methods, and/or an entity normalization index that associates entities with alternate surface forms derived from web corpora. The normalized value may be used to construct a request to a structured knowledge source and/or an application.