Scheme: Industry Fellowship
Organisation: Daiichi-Sankyo LTD
Dates: Jun 2012-Jun 2016
Summary: Ever wondered why phenomena in Life keep happening together? What is the stability of such co-occurrences? Co-incidence appears everywhere - but is it reliable?
Some birds can be dangerous, birds are often found in trees, trees usually grow in forests, therefore its likely that dangerous birds may be found co-incident with woodland. By how much does one item, or co-occurrence of items, link to another? How certain can you be that dangerous birds co-occur in woods? How can these relationships be visualised for easy comprehension? What makes the co-incidences reliable?
My Royal Society Industrial Fellowship poses a different projection of data collected to answer questions about the world developing the applied field of ’which co-incidences are real?’.
One phenomenon can tangibly influence another, co-occurrences can affect others in a natural cascade as Life is a series of interconnections. These interconnections are a dynamic network formed of many nodes (phenomena) and edges (relationships) which can be explicitly clear in experiments/surveys or implicitly covert (latent). A network visualization is straightforward for non-scientists to handle. What if co-incidence relationships could be easily characterised, summarized as displayed networks and related to co-occurent causes in the real world? What if complex data itself could tell you about its own stability and facilitate understanding?
Being sure co-incident co-occurences is important, interesting and exciting. It has direct relevance in designing robust credit fraud algorithms, identifying terrorist communications, making pharmaceutical investment decisions, understanding ecosystems, monitoring crowd behaviour or genetically decomposing disease. My fellowship’s aim is to extend and deploy into wider use a powerful algebra of this unseen world as tools for use by traditionally non-mathematical people. Practical relevance is exemplified