Natural Language Model Dictionary Limiting for Summary Generation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Training natural language models to generate summaries is challenging due to the vast number of words and variations, leading to the inclusion of irrelevant words and increased processing resources needed.

Innovation Solution

A method that limits the natural language model to use only words from the document being summarized, including variations, synonyms, and aphorisms, and trains the model by identifying longest substrings within human-generated summaries to create relevant and efficient summaries.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a natural language model is trained to generate summaries using a vast vocabulary, then the model can produce diverse language variations, but it includes irrelevant words and increases processing resources

Engineering Contradiction:

Improvelanguage variationVSAvoidprocessing resources

Core Design Contradiction:

Adaptability or versatilityVSLoss of energy

Solution Approach 1:

The patent extracts only the necessary words from the source document to form a restricted vocabulary for summary generation. By taking out only the relevant terms present in the original document and excluding unrelated vocabulary, the system maintains language diversity while reducing processing resources and eliminating irrelevant words from summaries.

Inventive Principle:

Principle #2Taking out (Extraction)

2Adaptability or versatility

If a natural language model uses a large vocabulary, then it can generate diverse sentence structures, but it increases the likelihood of including irrelevant words

Engineering Contradiction:

Improvesentence structure diversityVSAvoidsummary relevance

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The patent applies local quality by allowing diverse sentence structures and language variations only within the context of words present in the source document. Each summary is generated with locally restricted vocabulary specific to that document, ensuring relevance while maintaining natural language diversity through grammatical variations and different phrasing options.

Inventive Principle:

Principle #3Local quality

3Ease of operation

If the natural language model is trained with more words and variations, then it can produce more natural summaries, but it requires more processing resources

Engineering Contradiction:

Improvesummary naturalnessVSAvoidprocessing efficiency

Core Design Contradiction:

Ease of operationVSProductivity

Solution Approach 1:

The patent applies partial action by using only the portion of vocabulary necessary for summary generation - specifically, words extracted from the source document. This partial vocabulary approach maintains naturalness in summaries by allowing all grammatical variations and sentence structures using only the document's words, while improving processing efficiency by excluding unnecessary vocabulary.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS11281854B2Limiting a dictionary used by a natural language model to summarize a document

Publication Date: 2022.03.22 PRIMER TECHNOLOGIES INC

AI summary

The technology disclosed herein summarizes a document using a dictionary derived from tokens within the document itself. In a particular implementation, a method provides identifying a first document for summarization and inputting the first document into a natural language model. The natural language model is configured to summarize the first document using words from a first dictionary compiled based on tokens from the first document. The method further provides receiving a first summary output by the natural language model after the natural language model summarizes the first document.