Dr Sumon Shahriar of the Australian institute of Health and Welfare presented this talk to the Canberra Branch of the Statistical Society on Tuesday 2 May. In a bit of a reminder of how I used to handle Statistical Society commitments, I was not present in person at the talk but did attend the dinner afterwards. However it was lovely to be able to join the talk nonetheless, via Zoom. There were close to 20 people on the Zoom call, along with the in-person attendees.
Sumon’s talk was a very high-level overview, no equations, many icons. This meant it was a very accessible talk about the use of machine learning in this government agency, where traditionally new statistical methods take a while to be embraced. The examples Sumon presented included ML for probabilistic data linkage for major AIHW research projects; and random forests for prediction of early onset dementia from Medicare claims data.
Sumon provided some interesting links to Commonwealth Government policy on AI and the like: see material on AI, the ethics framework and ethics principles. He also advocated for early sharing of results from ML with senior managers in order to speed the process of convincing them of the value of ML in government research and policy work.
aliceinstatisticsland 