Hybrid Search Syntax for Backend-Agnostic Content Retrieval
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing hybrid search solutions require users to navigate complex pipelines and have limited control over search parameters, necessitating knowledge of backend search services and specific syntaxes for each service.
Innovation Solution
A hybrid search syntax that is agnostic to underlying search services, allowing users to define and execute hybrid queries with both lexical and semantic components using a unified syntax, which automatically invokes and combines results from different search services.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If users directly interact with backend search services (Elastic Search, Vector Database), then they can control search parameters in depth, but they must become acquainted with backend implementation details and various nuances of specific search services
Solution Approach 1:
The patent introduces an intermediary layer (hybrid search interface) between the user and the backend search services. This interface translates user-friendly search queries into appropriate calls to underlying services like Elastic Search and Vector Database, shielding users from backend complexity while maintaining control capabilities.
Solution Approach 2:
The patent creates a universal search interface that can work with multiple different backend search services through a single unified API. This multi-functional interface handles both lexical search (Elastic Search) and semantic search (Vector Database) through the same user-facing syntax, eliminating the need for users to learn service-specific nuances.
2Ease of operation
If a fixed user interface is provided for hybrid search, then the system is easier to use, but users have limited access to parameters of lexical and hybrid search engines and few options to control how searches are combined
Solution Approach 1:
The patent implements a dynamic search interface where the level of control adapts to user needs. The unified syntax provides simple operation by default, but allows advanced parameter control when needed, with the interface dynamically adjusting between simplicity and versatility based on the specific search requirements.
Solution Approach 2:
The patent merges multiple search control capabilities (lexical search parameters, semantic search parameters, result combination logic) into a single unified syntax. This combining allows users to control all aspects of hybrid search through one interface rather than separate interfaces for each search type.
3Adaptability or versatility
If users need to swap backend search services, then the system becomes more flexible, but users must change the search query syntax to accommodate different services
Solution Approach 1:
The patent creates a universal search syntax that serves as a stable interface layer above the backend services. This universal syntax can query any combination of search services (Elastic Search, Vector Database, or future services) without requiring syntax changes, enabling service swapping while maintaining operational consistency.
Solution Approach 2:
The patent segments the search system into two independent layers: a stable user-facing syntax layer and a flexible backend service layer. This segmentation allows the backend services to be swapped or modified without affecting the user syntax, as the interface layer handles the translation and adaptation.
Data Source
AI summary
One example method includes enabling a user to define a query that comprises a hybrid search syntax (HSS), and the HSS is in the form of a single string that comprises a lexical component and a semantic component, receiving the query from the user, parsing the lexical component and the semantic component into respective groups of one or more sub-components, performing a lexical search based on the sub-components of the lexical component, and performing a semantic search based on the sub-components of the semantic component, ranking, separately from each other, results of the lexical search and results of the semantic search, combining the results of the lexical search and the results of the semantic search to create aggregate results, and re-ranking the aggregate results.