Automated Experiment Design Using Response Prediction Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The designing of experiments with multiple variables, such as those in vaccine development, is a complex and time-consuming process that requires automation for efficient selection of critical process parameters (CPPs) and optimal experimental design.
Innovation Solution
A method and apparatus for designing experiments that involve generating candidate values of CPPs based on specific conditions, using a trained response prediction model to estimate target responses, and outputting experimental condition sets for CPPs based on prediction values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional experimental design methods are used for vaccine development with multiple variables, then comprehensive analysis of all process parameters is achieved, but the process becomes lengthy and complex
Solution Approach 1:
The patent extracts only the critical process parameters (CPPs) from the full set of process parameters through automated selection algorithms. This extraction allows the system to focus analysis on the most influential parameters while ignoring less important ones, thereby maintaining analysis comprehensiveness for key factors while dramatically reducing the overall time required for experimental design.
Solution Approach 2:
The system performs preliminary automated selection of critical process parameters and preliminary setting of their optimal values before the actual experimental design process. By pre-identifying which parameters are most influential and pre-determining their optimal settings, the system eliminates the need for time-consuming manual analysis of all parameters during the experimental design phase.
2Reliability
If manual selection of critical process parameters is performed, then expert judgment is applied, but automation technology is required for efficiency
Solution Approach 1:
The patent replaces the manual mechanical process of expert parameter selection with an automated computational system. The automated CPP selection algorithm processes experimental data and identifies critical parameters through systematic analysis, substituting human expert judgment with machine-based automation that maintains reliability while dramatically improving selection efficiency and productivity.
Solution Approach 2:
The system changes the approach from static expert judgment to dynamic automated parameter selection based on experimental data. By using algorithms that analyze actual experimental results and automatically identify which parameters have the greatest impact on outcomes, the system adapts parameter selection to the specific characteristics of each experiment, maintaining high reliability while enabling automation.
3Loss of information
If all process parameters are analyzed in detail, then complete understanding is achieved, but the complexity of experimental design increases
Solution Approach 1:
The patent extracts and identifies only the critical process parameters that contribute most significantly to experimental outcomes. By automatically selecting and focusing on these key parameters while excluding less influential ones, the system maintains complete understanding of parameter contributions through targeted analysis, while reducing overall design complexity by eliminating unnecessary analysis of non-critical parameters.
4Productivity
If automated technology is used for CPP selection, then efficiency is improved, but the process requires new methodologies
Solution Approach 1:
The patent introduces an automated CPP selection algorithm as an intermediary between raw experimental data and final experimental design. This intermediary component automatically processes data, identifies critical parameters, and determines optimal values, thereby improving efficiency while managing methodology complexity through a structured, systematic approach that bridges data analysis and design decision-making.
Data Source
AI summary
A method and apparatus for designing an experiment are disclosed. The method of designing an experiment includes generating a candidate value of a critical process parameter (CPP) based on a condition for the CPP corresponding to a target response, obtaining a prediction value of the target response for the candidate value of the CPP, based on a response prediction model trained to estimate a function of the target response for the CPP, and outputting an experimental condition set of the CPP based on the prediction value of the target response.