AI and DOE - Product Initiative Update
Alchemy’s big focus in 2022 is to launch synergistic artificial intelligence (AI) and design of experiments (DOE) functionality. This functionality will leverage algorithms and statistics to accelerate your research and product development.
Key building blocks of the functionality will start releasing this weekend and each weekend through the summer. We expect to have AI and DOE generally available by the end of August.
As you know, when lab staff work in Alchemy, we collect clean, labeled, formatted AI-ready data. For those customers interested in adding AI and DOE, when this product development is done, Alchemy will be able to leverage data in Alchemy to do three things:
- Search and score historical data for relevancy given current project objectives.
- Run DOE to learn about the underlying chemistry with a minimal number of test cycles.
- Run AI for formulation optimization based on single or multiple targets and formulating constraints.
We’re very excited about this new functionality for numerous reasons. In one platform, you will be able to:
- Do your lab-related work, collect AI-ready data automatically as you work, leverage your historical work, create efficient test plans and generate optimized formulation recommendations - all fully integrated within your existing digital workflows (i.e., in-app).
- Leverage DOE and AI in an interrelated and guided fashion.
- DOE addresses the situation when there is no data or only very shallow historical data and the system cannot yet run AI. It will help you extend your dataset in the most efficient manner possible (i.e., with the smallest, well-distributed, statistically optimal dataset), so that we can train AI models and run AI.
- Importantly, no previous experience with machine learning, data science, or statistics is required. Any chemist or scientist will be able to input their formulating objectives and constraints and will be guided in the most efficient manner to achieve their goals.
- Put a significant amount of control in the hands of the bench chemist in terms of defining the optimization parameters and formulating rules such as:
- Single ingredient constraints (e.g., use this ingredient in a range).
- Multi-ingredient constraints (e.g., use these two ingredients together, but they can never exceed 50% of total or never together, optimize for cost and/or total number of ingredients).
- Provide significant insight in real-time at the bench with automated statistical analysis and with AI explainability.
- The system can continuously learn with the addition of new data without you having to do anything from an IT perspective.
In our alpha and beta testing, we have seen this functionality significantly shorten testing cycles by as much as 80% and streamline broader research and product development by as much as 60%.
Self-Configuration - Product Initiative Update
This summer we are also working on adding self-configuration functionality. This means that lab staff will be able to extend the data captured in Alchemy in real-time. Prime examples of this will include new fields of information and new test methods. Alchemy will keep the collection of these new data types AI-ready by managing their formatting, labeling and metadata in real-time in the same way we manage CSM or customer admin configured data capture.
Features in this Release
- Create a new Study where a chemist can:
- Define targets, materials that will be used in formulations, and add multiple trials (formulations).
- Select as many material properties to be displayed in the formulation table as are needed, with the possibility to define new properties.
- Define any test method and associated parameters, and include those in the Study for all defined trials.
- Support varying units and conversions of units. For instance, start with milligrams, convert to grams, and vice versa. Also, the system will provide more complex conversions, such as using a material’s density to convert from volume to mass.
- Released a performance enhancement to make pages load faster on an iPad.
- Resolved an issue when printing a Record.
We are continuously improving our How To documentation, where you can find guides and manuals for using Alchemy. For this release, there are two new topics:
- How to add a Calculated Column. It further documents adding a Calculated Column to View as well as what operators and functions are supported in Calculations. You can also access this information from the application by clicking on the blue link in the Calculated Columns popup.
- How to visualize data in a View: gives full explanation about every type of chart and tutorial how to create and manage visualizations as part of your View. This information is available in Visualizations section inside Views: