Batch LLM Summarization for Aggregated Aspect Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data analytics systems using large language models struggle to interpret context, extract quantifiable metrics, and consider nuanced concepts or sentiments in unstructured textual data, especially when summarizing large amounts of such data.
Innovation Solution
A key-based or batch approach that assesses factors like the number of text entries and character length, sending batches to a large language model processor to develop numerical scores and summaries, leveraging LLM capabilities to interpret and quantify nuanced aspects, and using N-step processing with parallel execution and confidence-weighted averaging.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If large language models are used to process unstructured textual data, then contextual interpretation and nuanced concept extraction are improved, but processing efficiency and scalability deteriorate
Solution Approach 1:
The patent segments the unstructured textual data into multiple batches, which are then processed in parallel by multiple large language model instances. This segmentation approach maintains the contextual interpretation accuracy of LLMs while improving overall processing efficiency through parallel execution, directly resolving the contradiction between precision and productivity.
2Loss of information
If large language models process entire datasets at once, then comprehensive context analysis is improved, but processing time and resource consumption increase
Solution Approach 1:
The system performs preliminary actions by dividing the dataset into batches before processing, allowing parallel execution of multiple LLM instances. This preliminary segmentation enables comprehensive context analysis across all batches while reducing total processing time through concurrent operations, effectively addressing the trade-off between information completeness and time consumption.
3Productivity
If batch processing is used to improve efficiency, then processing speed increases, but handling of nuanced concepts and sentiments may deteriorate
Solution Approach 1:
The patent introduces an intermediary aggregation step that combines results from multiple batch processing operations. This intermediary layer ensures that nuanced concepts and sentiments are properly synthesized across batches, maintaining measurement precision while benefiting from the speed improvements of batch processing, thus resolving the contradiction between productivity and precision.
Data Source
AI summary
Embodiments described herein are generally related to data analytics environments, and to systems and methods for providing aggregated summaries and aspect scores associated with unstructured textual data. In accordance with an embodiment, the system uses a key-based or batch approach that assesses factors associated with an unstructured textual dataset, such as, for example, a total number of text entries per key, or the character length of each text entry. Based on a consideration of such factors, the system sends batches of text entries, and a prompt, to a large language model processor, to collect intermediate batch results. The intermediate batch results can be used first to develop a numerical score or summary for each key, directed to various aspects of interest within the data; and subsequently to generate aggregated summaries and/or aspect scores associated with the textual dataset, for use in displaying visualizations or returning additional analytical information.