Prompt Construction for Accurate Task Execution in App Creation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing application creation platforms struggle to understand user inputs that are strongly related to specific fields or scenarios, leading to inaccurate model outputs and failure to meet user requirements.

Innovation Solution

The application creation platform converts user inputs into target prompt inputs with richer and more informative description information, using a model to determine execution operations that align with user expectations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the application creation platform uses a model to process user inputs directly, then the processing speed is fast, but the model cannot accurately understand user inputs related to specific fields or scenarios

Engineering Contradiction:
Improvemodel understanding accuracyVSAvoidinput processing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent introduces a prompt construction module as an intermediary between the user input and the model. This module enriches the user input by adding background information, task requirements, and formatting instructions before the model processes it. The intermediary layer transforms simple user inputs into comprehensive prompts that the model can accurately understand and process, resolving the contradiction between processing speed and understanding accuracy.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Manufacturing precision

If the platform enriches user input with more description information, then the model output accuracy improves, but the processing time increases

Engineering Contradiction:
Improvemodel output accuracyVSAvoidprocessing time
Core Design Contradiction:
Manufacturing precisionVSLoss of time

Solution Approach 1:

The patent implements preliminary action by pre-defining prompt templates and enrichment rules before user input processing. The system has pre-prepared various types of background information, task requirements, and formatting instructions that can be automatically applied to user inputs. This preliminary preparation allows the system to quickly enrich user inputs without requiring real-time analysis, thus improving output accuracy while minimizing processing time increases.

Inventive Principle:
Principle #10Preliminary action

3Ease of operation

If the platform uses a simple prompt input format, then the ease of operation is high, but the model cannot generate accurate outputs for complex tasks

Engineering Contradiction:
Improveuser input simplicityVSAvoidmodel output reliability
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The patent implements self-service by enabling the system to automatically transform simple user inputs into comprehensive prompts without requiring manual intervention. The prompt construction module autonomously enriches user inputs by selecting appropriate background information, adding task requirements, and formatting the output according to predefined templates. This self-service mechanism maintains ease of operation for users while ensuring reliable and accurate model outputs for complex tasks.

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP4715585A1Method and apparatus for information processing, and device and storage medium
Publication Date: 2026.03.25 LEMON INC(GB)
  • EP4715585A1 patent drawingFigure 1
  • EP4715585A1 patent drawingFigure 2
  • EP4715585A1 patent drawingFigure 3~4

AI summary

Embodiment of the disclosure provides a method, an apparatus, a device and a storage medium for information processing. The method includes: receiving a first user input comprising first description information related to a task on an application creation platform; determining a target prompt input matching the first user input, the target prompt input comprising second description information related to the task, and the second description information comprising at least one of: description information obtained by adjusting at least a portion of the first description information, or an extended description for at least a portion of the first description information; and providing the target prompt input to a model to obtain an indication of an execution operation of the task output by the model. Therefore, by converting the user input into the prompt input that is more informative and easier for the model to understand, the output of the model can be more accurate, the task processing efficiency and accuracy are improved, and the user experience is enhanced.