Solving the 'small data' problem for formulations, materials, and processes
We hear a lot about ‘big data’. But how can we solve problems where we have little data? This is common when exploring new territory — for example, designing materials, chemicals, formulations, and processes. In these systems it can also be difficult, expensive, or time-consuming to generate new data. The right strategies and the right tools are required — to model properties and processes as accurately as possible with the data that is available, to learn from related data, and to ensure that experimental effort to acquire new data is targeted effectively and efficiently.
Click below to watch the full webinar.
In this webinar, we will explore these strategies and tools, including demonstrations of the Alchemite™ deep learning software. We will see how to maximise the value gained from sparse, noisy data, and how to dramatically reduce the amount of experiment or testing needed to achieve success in design and development of products and processes.
Original broadcast: 13th Jul 2021