Parameter sweep-type operations such as phase diagram exploration can be accelerated by adopting an artificial intelligence (AI) approach, learning the shape of the parameter space in real-time within the program.
Bayesian Optimisation automatically balances refining the detail of interesting features (exploitation) with ensuring good coverage of the whole space (exploration), using far fewer simulations than a grid-based approach. In the end, a more accurate result is achieved more quickly, and for lower cost.
The AI accelerator forms part of some of the virtual experiments available through the Consumable Computing for Chemistry interface. This approach could be adapted to your organisation as part of a collaborative R&D project
Accelerating phase-diagram exploration experiments.
- Improved accuracy
- Fewer simulations required
To see how our AI Accelerator can work for
your application – contact us