Multi-Index Transcript Query System for Contextual Retrieval
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search engines for transcripts use a single indexing schema, making it difficult to answer queries without reviewing the entire transcript or associated audio data, as they fail to capture the full context of multi-party communications.
Innovation Solution
A computerized method that generates multiple section indexes based on different artifact types within a transcript, allowing for accurate and flexible query processing by analyzing natural language queries and selecting relevant section indexes for search.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single indexing schema is used for all documents in the transcript, then the search engine structure is simple and easy to implement, but the ability to answer context-specific queries is limited and requires reviewing the entire transcript
Solution Approach 1:
The patent segments the transcript into multiple document types (meeting summaries, action items, discussion points, etc.) and creates separate indexes for each document type. This segmentation allows the search engine to query specific contexts without reviewing the entire transcript, thereby preserving context information while maintaining a manageable indexing structure.
Solution Approach 2:
The patent adds a new dimension to the indexing system by creating multi-dimensional indexes based on different document types and contexts. Instead of a single flat index, the system organizes indexes hierarchically by document type, allowing queries to traverse specific dimensions (e.g., action items, decisions, discussions) independently, thus retrieving context-specific information efficiently.
2Loss of information
If the entire transcript is reviewed to understand full context, then complete context information is obtained, but the time and computational resources required increase significantly
Solution Approach 1:
The patent performs preliminary action by pre-processing the transcript during indexing, organizing content into structured document types with associated metadata and context tags. This preliminary structuring enables rapid query resolution by directly accessing relevant sections without requiring full transcript review at query time, thus reducing processing time while maintaining context accuracy.
Solution Approach 2:
The patent applies local quality by optimizing different parts of the transcript for different query types. Each document type index contains context-specific information tailored to its purpose (e.g., action item indexes contain task-related context, summary indexes contain overview context), allowing the system to retrieve only the necessary local context for each query rather than global context for all queries.
3Productivity
If multiple section indexes are generated based on different artifact types, then query accuracy and efficiency are improved, but the complexity of the indexing system increases
Solution Approach 1:
The patent implements universality by designing a unified indexing framework that handles multiple document types through a common architecture. The system uses a standardized index structure that can accommodate different artifact types (meetings, calls, webinars) and document categories, allowing the same indexing mechanism to serve multiple functions and reducing overall system complexity despite the multi-dimensional nature of the indexes.
Data Source
AI summary
The disclosure herein describes providing responses to natural language queries associated with transcripts at least by searching multiple indexes. A transcript associated with a communication among a plurality of speakers is obtained, wherein sets of artifact sections are identified in the transcript. A set of section indexes is generated from the transcript based on artifact type definitions. A natural language query associated with the transcript is analyzed using a natural language model and query metadata of the analyzed natural language query is obtained. At least one section index of the set of section indexes is selected based on the obtained query metadata and that selected at least one section index is searched. A response to the natural language query is provided including result data from the searched at least one search index, wherein the result data includes a reference to an artifact section referenced by the searched section index(es).