The Evidence Constraint
Model performance is no longer where value is created in clinical AI, and it has not been for some years. Five predictions I'm willing to be wrong about, with dates attached.
A bright future or cause for concern? Seven years on
In 2018 we wrote about artificial intelligence, machine learning and the evolution of healthcare. This is what has changed since, what has held, and where the binding constraint sits now. A commentary on Jones LD, Golan D, Hanna SA and Ramachandran M, Bone & Joint Research, 2018.
AI in developmental dysplasia of the hip, a decade later
The first machine-learning question my field ever asked, revisited. What the models can now do, what the datasets still can't, and why DDH remains a problem the world has not solved despite trying. A commentary on Golan D, Donner Y, Mansi C, Jaremko J and Ramachandran M, in Intelligence-Based Medicine, Springer 2016.
Never miss an essay
Essays, predictions and announcements. Nothing else. I will let you know when they are released if you contact me.