Making use of our analysis to medical settings
We’ve taken steps to make sure our work is clinically relevant. This consists of the event of a new performance metric used to evaluate mannequin efficiency that we imagine is extra consultant of medical processes, and a take a look at set with new high-quality segmentations of scans chosen from websites beforehand unseen to the mannequin which demonstrates generalisability. Each of those have been open sourced to the analysis neighborhood. However for our system to have an effect on actual folks recognized with most cancers, we have to develop it and exhibit that it really works in actual medical environments.
That’s why we’re wanting ahead to transferring into the subsequent section of labor with UCLH, the place we will probably be exploring a human analysis of those AI algorithms to check how they could carry out in a medical atmosphere.
At DeepMind Well being, we expect it is very important share our work with others in the neighborhood. As such, Professor Olaf Ronneberger, Senior Analysis Scientist at DeepMind Well being, will probably be presenting these preliminary findings at MICCAI, the world main convention on medical imaging, this Sunday.
In the end, we imagine that superior applied sciences can and will assist change lives, and we’re excited for the subsequent steps of this mission. We’ll proceed to maintain you up to date as we make progress.