Microsoft's Ava Amini shows us a new context for biological AI.

Sometimes when you're considering how to bring the power of AI into a medical context, it takes a new way of thinking about what's possible.

I was thinking about this the other day, inspired by some people who are working hard on genomics, oncology research, and other kinds of biological and physiological applications. There's a lot of it, all of a sudden, especially in institutions that I'm very close to – to call it a “revolution” in my view, isn't hyperbolic.

So how about a new way of looking at medical AI? To me, Ava Amini at Microsoft Research really articulates that kind of framework and that kind of purpose when it comes to, in her words, “precision medicine” and “precision health,” and biology and AI. Wedding. He describes this effort as follows:

“Our work is really fundamentally based on the fact and the realization that human life is a miracle: on the one hand, incredibly strong, and on the other, incredibly fragile. … We just need to detect disease, or There is no need for a better way to discover a new drug – I would argue that we need an entirely new way of thinking, a way of leveraging data about the human body, biology, and learning. Pattern naturally and internally.

Our bodies, she suggests in panel talks and lectures, are unique organisms that work together, but many medical applications don't take our uniqueness into account.

Too often, we are treated, as she puts it, as “an example in the population” rather than a unique human being. And of course, clinical science is based on data, and that's a good thing, but to assume that all patients are (X) is kind of simplifying things and maybe we're missing something. In fact, I know a lot of good doctors who would make that claim as well, saying that you start with a medical journal, but you also meet your patient where they are.

So what happens when you personalize medicine?

Well, you get better targeted results for individual people. Patients feel better served – doctors feel more competent and successful.

I'll present some numbers I heard Amini mention recently as part of his presentation, which you get from a biopsy:

100,000 variations

10 million genes

200,000 proteins

34 million cells

That kind of puts it in perspective, doesn't it?

I was also excited about the project that is happening with stakeholders, not only Microsoft, but both MIT and Harvard, which will help expand the pipeline and get more data into clinical AI work.

Amini describes it as a “flywheel” for generating new data that is tested in the lab and refined for clinical practice. I love that metaphor!

EvoDiff is a tool I've also heard of from Amini: What is it? Well, GitHub explains that it's “a framework that combines evolutionary data with diffusion models to generate controllable proteins in sequence space.” It's also now open source, which is great, in an environment where communities are trying to work together to leverage AI instead of keeping tools in walled gardens.

To me, this is a very effective way to organize this kind of work – unlocking the power of biology with genAI and artificial intelligence tools. Yes, we use AI in radiology. Yes, it plays a role in diagnosis. But it's also a Socratic helper when we use it to better think about and treat each of our unique bodies!