Smarter Experiments for the Lab of the Future: Man and Machine knowledge Integration

Unilever employ experienced people and have large repositories of historic
experimental data. This project relates to combining the information inherent
in these contexts to help optimise decisions about which experiments to
conduct in the future. More specifically, this project will develop tools to
support such decision-making.
Mature techniques can solve parts of this problem. Indeed, psychologists
have developed Structured Thinking Techniques (STTs) to help mitigate
cognitive biases when capturing human expertise. Similarly, data scientists
have developed statistical methods for analysing data. However, there is
relatively little research that is genuinely at the interface between these two
disciplines and which attempts to synergise the views taken by the two
communities of academic researchers.
This project aims to develop decision support tools that explicitly pull on
both psychology and data science. This will involve careful statistical
modelling of the process of asking questions of experts as well as careful
analysis of the psychology of using interface to state-of-the-art data science.
The project pulls on previous research on developing such tools in defence
and security contexts. This project is focused on understanding how to
configure such tools to deliver benefit in the context of high-volume, highvalue
manufacturing processes, exemplified by those of relevance to
Unilever.