SOLVE magazine Issue 04 2022 | Page 9

SPACE
Data , however , describes . It does not , of itself , explain what is driving a system or what is causing it to misfire . For example , a complete survey of the human immune system does not necessarily reveal why that system suddenly slips up in an individual and allows a cancer to form .
Systems with many interacting parts are complex . They are also dynamic – they drive changes to their own operation . That means no description of a dynamic system is ever complete – there is always an element of uncertainty in understanding what state it is in .
So , the world has imperfect data . Additionally , it needs to be processed to extract understanding about systems .
Professor Amara says astronomy was among the first research disciplines in the world to experience this crisis since it was among the first to observe a vast system – the Universe – and generate enormous amounts of descriptive digital data .
“ At the ICG , we have pushed hard to gain expertise at processing big datasets to extract meaningful signals against a great deal of uncertainty .”
That ’ s why ICG graduates are sought out by big tech companies – similar datarelated challenges occur across the globe . Complexity and uncertainty are the new technology frontier .
From crisis to opportunity
Professor Amara captures the sense in which our accumulated data poses challenges by referring to it as simply “ messy ”. And no one word can explain the sophisticated things the ICG does to extract meaningful signals from messy data , as it involves the application of exceptional mathematical rigour .
That ’ s the key , says Professor Amara .
“ The process we use is a combination of two elements . You need to interrogate the data well by asking good questions . You then need to write software code , including artificial intelligence ( AI ) algorithms , to perform the data analytics .”
It ’ s a process that frequently requires building computer models that mimic a system ’ s operation and then comparing and contrasting its operation with different datasets .
The ICG first got involved in running nonastronomy data analytics through several projects led by Professor David Bacon and supported by the STFC ( Science and Technology Facilities Council ) Innovation Accelerator Award .
For example , the MoleGazer algorithm not only maps changes to skin moles from total body photos , but also uses AI to inspect and diagnose each mole automatically . It was derived from software used to map the night sky and detect new events , such as supernova . The work was undertaken by astrophysicists in collaboration with the University of Southampton and adopted by the Oxford University Hospitals NHS Foundation Trust . Similar impacts are possible by processing cardiograms ( for early signs of heart attacks ) and other medical images .
ISSUE 04 / 2022 9