Personalized Prompt Generation Using Reward-Guided Candidate Selection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The manual design and evaluation of prompts for large language models to generate personalized replies is time-consuming and costly, failing to effectively elicit personalized responses.

Innovation Solution

A prompt generating device and method that utilizes a transceiver interface, processor, and trained prompt generator to automatically convert situational questions into candidate prompts, determine a best prompt through reinforcement learning, and guide a large language model to generate personalized replies.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual design and evaluation of prompts is used, then personalized replies can be generated, but the process is time-consuming and costly

Engineering Contradiction:
Improvepersonalization qualityVSAvoidprompt design time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system enables automatic prompt generation and evaluation through the trained prompt generator, eliminating the need for manual human intervention in prompt design and reply evaluation, thereby reducing time cost while maintaining personalization quality

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces the manual mechanical process of prompt design and evaluation with an automated system based on a trained prompt generator that uses reinforcement learning to generate and evaluate prompts automatically

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If manual design and evaluation of prompts is used, then personalized replies can be generated, but the process is costly

Engineering Contradiction:
Improvepersonalization qualityVSAvoidevaluation cost
Core Design Contradiction:
Measurement precisionVSLoss of energy

Solution Approach 1:

The trained prompt generator autonomously performs prompt generation and evaluation without requiring human experts, thereby eliminating the high costs associated with manual prompt design and reply evaluation

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent substitutes the expensive manual evaluation process with an automated evaluation mechanism based on the trained prompt generator, which can efficiently assess reply personalization without human intervention

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If automatic prompt generation is implemented, then time cost is reduced, but automation extent must be increased

Engineering Contradiction:
Improveprompt generation efficiencyVSAvoidautomation complexity
Core Design Contradiction:
ProductivityVSExtent of automation

Solution Approach 1:

The system performs preliminary training of the prompt generator using reinforcement learning before deployment, enabling it to automatically generate high-quality prompts without requiring complex real-time automation mechanisms during actual operation

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent incorporates a feedback mechanism where the trained prompt generator evaluates its own generated prompts and replies, using reinforcement learning signals to improve future prompt generation, thereby achieving high automation with manageable complexity

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS20260080188A1Prompt generating device and method
Publication Date: 2026.03.19 HON HAI PRECISION INDUSTRY CO LTD
  • US20260080188A1 patent drawing
  • US20260080188A1 patent drawing
  • US20260080188A1 patent drawing

AI summary

A prompt generating device and method are provided. The prompt generating device receives a situational question. The prompt generating device transmits a target personality category among personality categories and the situational question to a prompt generator to generate candidate prompts corresponding to the target personality category and reward signals corresponding to the candidate prompts, and the prompt generator is trained based on a large language model and a reward model corresponding to the personality categories. The prompt generating device determines a best prompt from the candidate prompts based on the reward signals corresponding to the candidate prompts.