Multi-Personality Hashtag Generation for Implicit Content Retrieval

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing hash tag generation methods do not facilitate easy content retrieval by users, as they do not account for generating tags that include information not directly expressed in the content.

Innovation Solution

A hash tag generation device and method that utilize a language model with multiple personalities, including an interviewer, copy writer, and optionally a judge, to interactively generate tags through prompts, ensuring the generated tags capture relevant information beyond the content's explicit details.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If a language model generates hash tags based only on explicit content details, then the generation process is simple, but the generated tags do not facilitate easy content retrieval

Engineering Contradiction:

Improvecontent retrieval easeVSAvoidgeneration process complexity

Core Design Contradiction:

Ease of operationVSDevice complexity

Solution Approach 1:

The patent segments the hash tag generation process into multiple distinct roles: interviewer personality that extracts implicit information, copywriter personality that generates tag candidates, and judge personality that selects final tags. This segmentation allows each component to specialize in one aspect of the complex task, improving retrieval quality while managing complexity through modular design

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces an intermediary multi-personality language model system that mediates between the original content and the final hash tags. This intermediary processes content through multiple personalities sequentially, transforming explicit content into implicit information and then into optimized tags, thereby resolving the contradiction between simplicity and effectiveness

Inventive Principle:

Principle #24Intermediary (Mediator)

2Loss of information

If multiple personalities interact to generate hash tags, then implicit information is captured, but the generation time increases

Engineering Contradiction:

Improveimplicit information captureVSAvoidgeneration time

Core Design Contradiction:

Loss of informationVSLoss of time

Solution Approach 1:

The patent applies preliminary action by having the interviewer personality first extract implicit information from content before the copywriter generates tags. This preliminary extraction phase ensures that all necessary implicit information is captured upfront, allowing subsequent tag generation to proceed efficiently without repeated analysis

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent uses partial action by having the judge personality selectively evaluate only the most promising tag candidates generated by the copywriter, rather than reviewing all possible tags. This partial evaluation approach captures essential implicit information while avoiding excessive processing time

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS20250348528A1Hash tag generation device, hash tag generation method, and recording medium

Publication Date: 2025.11.13 NEC CORP

AI summary

A hash tag generation device includes a memory storing instructions; and one or more processors configured to execute the instructions to: receive a generation prompt to generate a hash tag for content including at least one of a sentence, an image, a moving image, and a voice; set a plurality of personalities for generating a hash tag for a language model used for generating the hash tag; generate a hash tag of the content through an interaction using a prompt among a plurality of personalities set for the language model; and output the generated hash tag.