The Alchemite™ 2023 Spring Release is a milestone release of the machine learning software, bringing together the many new features delivered over the past six months. New features fall under three main themes.
Optimise real-world experimental programmes– it’s even quicker and easier to specify experimental designs, setting up subtle but important constraints that are often only possible via complicated workarounds when using traditional Design of Experiments (DOE) software.
Enable fast, iterative design that fits scientists’ workflows– design hundreds of new experiments in a matter of minutes. Ideal for those beginning their journey with AI, the experimental design user interface has been redesigned from the ground up.
Support teams and the enterprise to apply machine learning with enhanced features that support a collaborative approach, from product requirements through to the lab and beyond. Investment in infrastructure has further increased the robustness of the system.
Flexible specification of experimental constraints
Balance the power to explore design space with experimental feasibility by specifying acceptable ranges for the number and percentage composition of particular types of ingredient.
Get even more detailed control of your design through use of constraints-within-constraints.
Unlike alternative approaches that leave the user to filter out unfeasible experiments, Alchemite™ recommends options that fall within these constraints and quantifies their likelihood of success.
Fully-integrated handling of categorical data in experimental design.
Alchemite™ can now search over and target categorical variables (e.g., text labels) at the same time as numerical data when designing experiments. This avoids a time-consuming process of setting up different modelling methods for numerical (regression) and categorical (classification) tasks.
Fast, iterative design
The new experimental design user experience focuses on key tasks – ideal for those beginning their journey with AI.
Enable rapid design and iteration through an intuitive new UI for adding experimental constraints and targeting areas of experimental space. No need to think about mathematical loss functions or write complex scripts.
Work effectively even with very large datasets using the new Column Groups feature that makes it easy to add, manipulate, and design with hundreds of columns at once.
Get an instant overview of your current project in the new dashboard view, making it easier to assess progress and decide on next steps.
Support teams and the enterprise
Search and organise proposed experiments more efficiently through the new tagging feature.
Add notes to your project, smoothing the communication of results and analysis.
Investment in infrastructure
Increased availability and resilience in the cloud computing environmentused by Alchemite™.
Enhanced testing and monitoring of the system.
Greater stability for long-running compute jobs.
Upgrades to authentication technologies.