LLM Communication Record Segmentation for Targeted Topic Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing communication analysis systems fail to efficiently provide targeted summaries of large discussions, requiring users to sift through extensive meeting transcripts or logs, and often lack sufficient detail on specific topics of interest.
Innovation Solution
Utilizing an AI assistant and a Large Language Model (LLM) to preprocess communication records, segment relevant portions based on user-defined constraints, and generate targeted analyses through prompts tailored to the user's needs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If users review entire meeting transcripts or logs to find specific information, then they can obtain complete information, but the time required increases significantly
Solution Approach 1:
The patent segments large meeting transcripts into smaller, topic-specific portions using NLP techniques to identify boundaries and structure. This allows the system to present only relevant segments to users rather than requiring review of entire transcripts, thus reducing time loss while maintaining information completeness through targeted segmentation.
Solution Approach 2:
The system extracts specific topics, action items, and key information from large communication records using NLP and LLMs. By taking out only the relevant portions related to user queries, the system eliminates the need for users to sift through extensive unrelated content, resolving the contradiction between information completeness and time efficiency.
2Adaptability or versatility
If communication analysis systems provide comprehensive summaries, then all topics are covered, but the level of detail on specific topics of interest is insufficient
Solution Approach 1:
The patent implements a dynamic analysis system that adapts to user needs by adjusting the level of detail and focus based on specific queries. The LLM generates different types of analyses (summaries, detailed topic analyses, action item extractions) depending on user requirements, allowing the system to be both comprehensive and detailed when needed.
Solution Approach 2:
The system applies local quality by providing different levels of analysis depth for different topics within the same communication record. Highly relevant topics receive detailed LLM-based analysis with specific insights, while less critical topics receive standard summarization, thus achieving both broad coverage and deep detail on specific topics.
3Quantity of substance
If existing systems analyze large volumes of communication data, then comprehensive coverage is achieved, but efficiency and targeted analysis are reduced
Solution Approach 1:
The patent applies preliminary action by pre-processing communication records to identify key structures, topics, and potential action items before user queries are submitted. The NLP system prepares segmented representations and topic tags in advance, so when users submit queries, the LLM can quickly generate targeted analyses without re-processing entire large volumes of data, thus improving efficiency while maintaining comprehensive analysis capability.
Data Source
AI summary
One example method includes receiving a set of communication records, the set of communication records representing one or more communications between a first person and a second person; receiving a user query, the user query comprising one or more constraints; generating a plurality of segments from the communication records, at least a subset of the plurality of segments based on the one or more constraints; selecting one or more segments based on the one or more constraints; generating one or more queries based on the one or more constraints; providing, to a trained large language model, the one or more selected segments and the one or more generated queries; and receiving, from the trained LLM, and outputting a multidimensional analysis of the set of communication records.