Bayesian Optimization for Food Experiment Design
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current food experimentation methods are inefficient, requiring extensive time and resources to achieve objectives such as pH below 6 in yogurt production, as they involve varying multiple variables in sequential trials.
Innovation Solution
A computer-implemented method using Bayesian Optimization machine learning to generate experiment trials that optimize parameters like water, sugar, and probiotic levels, focusing search efforts on high-performing areas of the search space based on previous observations, thereby reducing the number of trials needed to achieve experiment objectives.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If sequential trials with one variable changed at a time are conducted, then the experimentation process is systematic and controllable, but the time and resources required are excessive
Solution Approach 1:
The patent implements feedback by using the results of each experiment trial to inform and adjust subsequent trials. The system learns from previous outcomes and uses this information to guide the selection of parameters for future experiments, creating a closed-loop optimization process that reduces the number of trials needed while maintaining systematic control
Solution Approach 2:
The patent applies parameter changes by simultaneously varying multiple parameters across different trials rather than changing one parameter at a time. This allows the system to explore the parameter space more efficiently and achieve experimental objectives in fewer trials while still maintaining controllability through structured parameter exploration
2Adaptability or versatility
If multiple variables are tested in sequential trials, then comprehensive exploration of parameter space is achieved, but the number of trials and resource consumption increase significantly
Solution Approach 1:
The patent applies preliminary action by using a machine learning model to predict promising parameter combinations before conducting physical experiments. The system performs computational exploration and generates hypotheses about optimal parameters in advance, allowing the physical experimentation to focus on validating and refining these predictions rather than exhaustively testing all possibilities
Solution Approach 2:
The patent introduces a machine learning model as an intermediary between the experimental objectives and the actual trials. This intermediary learns from experimental data and guides the selection of subsequent trials, acting as a mediator that translates comprehensive parameter exploration requirements into a reduced set of high-value experiments
Data Source
AI summary
Techniques to generate experiment trials using artificial intelligence are disclosed. A training set for an experiment generator is continuously built up by using assessed experiment trials. The experiment generator is optimized using one of a plurality of optimization algorithms, depending on which mode experiment generator is to run in for an experiment. The mode is dependent on the experiment mode of the experiment. The experiment generator generates a batch of one or more experiment trials for the experiment. Any of the generated experiment trials may be tried or experimented by a user and may be updated with assessment data as an assessed experiment trial.