As Bucky Fuller said, create new models to make the old models obsolete, if you want change. It’s happening right now, another huge story for the future of the world that’s at least partly moving in a different direction than current politics. This story has many elements of the new model for healthcare that PSA has been calling DIY healthcare, based heavily on new cloud and AI tools and mobile devices. This to go with DIY education, in a general trend of enabling tools from the communications revolution ongoing now.
[gview file=”https://publicservicesalliance.org/wp-content/uploads/2017/01/Automated-dermatologist-detects-skin-cancer-with-expert-accuracy-CNN.com_.pdf”]
One of the core engines of AI progress is so called “machine learning” or “convolutional neural network” where computers using algorithms and programing “learn from experience” with Big Date and Meta Data sets. We don’t fully understand how they are learning, but we can test their “knowing” by what they do with test questions. Do they get the right answer? If they do consistently and for tens or hundreds of thousands of iterations, the AI must have “learned” how to “get” the right answer.
We learn in the book Whiplash, that machine learning is driving new AI capabilities at an exponential rate…you know, the curve on the graph that starts climbing almost straight up after a certain number of iterations. “Scientists” have used this process, CNN reports, to “train” a smart phone app to be able to diagnose as well as doctors, skin cancer lesions. And while one would still need a surgeon to remove cancerous cells, getting the diagnosis in time is really the central challenge for skin cancer health care.
DIY in other words. Would mobile big as a microwave or maybe hold in your hand sized surgery machines be also available eventually? Seems possible, but the really huge part of healthcare is steady monitoring of physical processes, to anticipate preventive measures, or to gauge the effectiveness of preventive measures. To coin a phrase, a stitch in time saves nine.
Some of these diagnostic tools might work with a smartphone or augmented reality device with added sensors, and smart AI apps. Some might be expensive, and only available for use at a Community School in a neighborhood near you, where they could also be a DIY healthcare facilitator.
BTW, and this has been noted before elsewhere, one of the stunning realizations discussed in “Whiplash” is that machine learning for one set of data, can be used, with some modifications for another set of data.
IOW an app for detecting skin cancer, might be modified to learn another seemingly unrelated task. And this done without starting over from scratch, so that machine learning is progressing exponentially, each new app learning from the previous one.
That’s very very rapid change if it continues for any length of time, as the graph does start eventually pointing almost straight up. That’s already got a name, it’s called the SINGULARITY because what happens at that point is not predictable. IOW “singlurlarly” unknown.