Introduction to | Materials Informatics
Materials Informatics is used across products, materials and chemicals, anywhere where the properties of a substance give competitive advantage.
|
What it does
Cuts through complexity
|
What it enables
Efficiency in developing and producing products, materials and chemicals |
How it works
Guiding experts’ analyses and learning from their choices
|
What Materials Informatics does:
- Today’s challenges are a complex balance of competing demands, and the experimental data used to find solutions are complicated, subtle and expensive.
- Informatics uses computational techniques and statistics to analyze and interpret this data far more efficiently than a human could.
- AI models help us understand trends and relationships in our data, and suggest recipes to achieve our goals.
To create innovative, high-performance materials, traditionally scientists have needed to test many ideas physically. Materials Informatics gets AI to run many, many virtual experiments, so they can refine their theories and focus on the most promising options.
Why Materials Informatics matters:
- New products can take years to develop, using specialized knowledge, trial-and-error, and luck. Materials informatics accelerates this.
- By accelerating innovation, product experts can be more agile in responding to customer requirements, production issues, supply shocks and regulation.
- As a society, we can expedite the science we need to solve our biggest challenges.
How Materials Informatics works:
- Computers are good at organizing complex data from a variety of sources into a form that a researcher can explore. This includes seeing the effect of a range of ingredients and manufacturing processes on the final characteristics of an end product.
- AI models train themselves to find patterns in data and make predictions.
- Product experts refine models, using physical insight and restrict suggested recipes to those that use feasible ingredients and practical processing parameters.
- Target final properties are set and recipes suggested. Statistical techniques are used to understand which experiments are more exploratory and which are a safe bet.
- An iterative process of strategic experimentation and retraining the AI model with the new data then guides researchers to their targets with up to 80% fewer experiments than traditional methods.
Example
A chemist might want to create a clear material without using a key ingredient. The model might propose a ranked list of 100 ways to create this. The chemist chooses the most promising options to test in a lab, and then the model retrains with the new results, becoming smarter and suggesting new, better recipes.
