Advanced machine learning in a few clicks
Alchemite™ Analytics provides simple web browser access to the powerful Alchemite™ machine learning algorithm, tuned for the study of chemicals, materials, formulations and associated experiments and processes. It enables data analysis, model-building, optimisation, and prediction, with intuitive graphical visualisation of results. Any scientist or engineer can get started, fast. No coding skills are required. Alchemite™ Analytics is lightning-quick. It takes minutes to train models that take conventional machine learning methods hours or days.
Enrich and understand sparse and noisy data
Import any numerical or categorical data – however sparse and noisy – and extract value from it with tools to validate and gap-fill data and to detect outliers. Auto-generate predictive models from the data, enabling breakthrough insights. Sophisticated uncertainty quantification enables you to gain confidence and enables rational decisions to achieve project targets while minimising time and cost.
Fewer experiments, better results
Whether you are running scientific experiments or other types of data acquisition project, with Alchemite™ you can get more from less. It tells you what data to gather next in order to refine its model most effectively for the least effort. Then you can apply that model to propose which candidate solutions are most likely to meet specific targets. This enables an Adaptive Design of Experiments approach. Typical results are that project objectives are achieved with 50-80% fewer experiments.
Case study
How Domino Printing Sciences reduced key experimental timescales.
Case study
Designing a new aerospace alloy with Rolls-Royce to simultaneously meet eleven physical property criteria.
Design and optimise products and processes
Alchemite™ finds relationships in high-dimensional data and can optimise against multiple target properties. Generate innovative solutions that you might otherwise miss. Run virtual experiments to quickly screen candidate solutions. The ability to build models across all available data, even where that means coping with sparse data sets, means Alchemite™ helps you to see the full picture. For example, for material, chemical, or drug formulations, you can optimise ingredients and process parameters in a single study.
Explainable AI – get deep insights
The Alchemite™ platform does much more than simply predict outcomes or provide guidance on experiments. It offers a host of visual analytics that help you to understand and explore the relationships captured by the machine learning model. These ‘explainable AI’ features avoid the problems associated with using machine learning as a ‘black box’, providing deep insights and building trust in your results.
Enable effective collaboration
Standardise your team’s machine learning approach with an easy-to-use shared machine learning toolset, aiding consistency, promoting knowledge-sharing, and easing on-boarding of new team members. By capturing key insights in re-usable models, you protect vital intellectual property – for example, when scientists leave or retire.
Alchemite™ Analytics can be deployed in cloud, on site, or in combination, depending on customer needs.
Managed deployments are quick, secure and cost effective.
Releases
The latest milestone is the Alchemite™ 2024 Spring Release. It offers: optimised workflows, new support resources, enhanced project and model management, greater interactivity with data and charts, and more features for analytics.
Previous release: Alchemite 2023 Autumn Release.
