If race can be guessed based on skeleton, then “bias” could also be interpreted as “giving useful medical signal based on measurable traits”. In order to eliminate human bias, wouldn’t you’d need to strip the “race” info from the notes given to the radiologist?<p>It is not clear at all whether what the ML is picking up is from genetics or from other sources (e.g. diet and other practices that associate with the cultures that often follow along with visible markers of race). Getting the ML model to explain itself will be an interesting follow-up to watch.
Begs the question: what other nuggets could be gleaned from correlating this knowledge to other diagnosis that resulted from the corresponding radiology study performed. Be even more interesting if other characteristics outside race could also be deduced by this or other similar systems. Very hopeful data science/ AI in general unlocks the potential of computers in the medical space especially with regard to combating and preventing issues before they submerge.