Dynamic Narrative Generation via Topic Segmentation and Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing natural language document generation technologies, such as robot journalism, face challenges in creating a narrative flow that is not predetermined, resulting in formal and inflexible document structures.
Innovation Solution
A method and apparatus for learning a narrative by generating a topic database, splitting documents into segments using topic vectors, grouping segments into clusters based on similarity, and assigning cluster labels, allowing for the generation of a label sequence that represents the narrative of a document in a computing device.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a predetermined narrative flow is designed and sentences are arranged according to the designed flow, then the document generation process is systematic and controllable, but the generated document structure becomes formal and inflexible
Solution Approach 1:
The patent applies dynamics by transitioning from a static, predetermined narrative flow to a dynamic, data-driven narrative structure. The system automatically determines narrative flows based on real-time analysis of document data, allowing the narrative structure to adapt and change according to the actual content rather than following a fixed template. This resolves the contradiction by making the narrative flow flexible while maintaining systematic control through automated algorithms.
Solution Approach 2:
The patent changes the parameters of narrative flow from being predetermined and fixed to being dynamically determined based on document characteristics. By analyzing features such as topic distribution, sentence structure, and semantic relationships, the system adjusts narrative parameters automatically, enabling flexible document structures that adapt to different content types while maintaining generation control through computational methods.
2Productivity
If automatic sentence generation using algorithms is applied, then document generation efficiency is improved, but the narrative structure becomes formal and lacks diversity
Solution Approach 1:
The patent implements feedback mechanisms where the system continuously analyzes generated documents and adjusts the narrative flow accordingly. By monitoring document characteristics and comparing them against desired outcomes, the system refines its automatic generation process, maintaining high efficiency while producing diverse and adaptable narrative structures that reflect the actual content rather than following rigid templates.
Solution Approach 2:
The system performs self-service by automatically analyzing and reorganizing document content to determine optimal narrative flows without human intervention. The algorithm processes document data, identifies patterns, and generates narratives autonomously, maintaining efficiency while achieving diversity through data-driven decision-making rather than relying on predetermined structures.
3Device complexity
If a fixed narrative flow template is used for document generation, then the generation process is simple and controlled, but the output becomes formal and lacks adaptability to different content types
Solution Approach 1:
The patent applies segmentation by dividing the document into discrete units (sentences, paragraphs, or semantic segments) that can be independently analyzed and reorganized. This allows the system to maintain a relatively simple processing framework while achieving high adaptability, as the segmented components can be dynamically arranged in different narrative flows according to the specific content type and characteristics without requiring complex predefined templates for each scenario.
Data Source
AI summary
Disclosed are an apparatus and method for learning a narrative of a document, and an apparatus and method for generating a narrative of a document. According to an embodiment of the present disclosure, the narrative learning method includes the steps of receiving a plurality of documents, generating a topic database which includes one or more topics and words related to each of the one or more topics from the plurality of documents, splitting each of the plurality of documents into one or more segments including one or more sentences by using the topic database, grouping the segments split from each of the plurality of documents into one or more clusters, and generating a cluster label for each of the one or more clusters.