Forget Moonshots Biomedicine Needs an Air Traffic Control System Jeff Shrager Cancer Commons xCures Stanford Symbolic Systems Program (Adjunct)
Hard AI: Cancer Easy AI: Self-Driving Cars Smal all, l, local decisi sion on environ onment ent, Extremel emely y broad d decisio sion n environ onment, ment, in both space ce and time in both space ce and time Mostl tly y static tic, , mostly tly well-und nder ersto stood od Dynamic “rules” (biology doesn’t change rules s and principl iples es but everyt rythi hing ng else e does) s) Data is plentiful tiful Data is essen entia tially lly non-exi xiste stent Comput uter er simula latio tion is nearly ly trivia ial Simula ulati tion on is essent entiall ally y imposs ssible e (The immun une e system stem is as comple lex x as the brain!**) n!**) Physica sical l simula latio tion n is easy Expert t guidan ance ce is is insta tant ntaneo aneous us, , Physica sical l simula latio tions ns is is extremely emely cheap, , and nearly ly perfect ct expensiv sive, , and every y experime iment nt kills s people le or anima mals ls in horrible le ways! Feedbac back k can take e years s and is very y noisy; y; ** Immune system: l trillion T cells, l trillion B cells, all circulating 50x/day, plus 10 billion antigen-presenting cells. Human Brain: 100 billion neurons, trillions of There e are NO EXPE PERTS! TS! synapses, and 1 billion glial cells. And it pretty much doesn’t move. AND BOTH LEARN!
Global Cumulative/Coordinated/Continuous Treatment Analysis
Why is Cancer so Hard? We’re treating an pre-OMIC OMIC era: Tissue sue x Chemo mos Now: w: Featur ures s x Targeted ted Comb mbos extremely high dimensionality, low Treatments: Chemo1, Chemo2, ... Thousands of drugs in combination ~1 Million patients/year ~1 Million patients/year data density, ~11 Z cells ~100 cells problem the same =~10,000 patients/cell =~0 patients/cell way that ants Plenty for Classical Need a new paradigm (and “big data” won’t cut it!) Clinical Trials search for food! Phenotypes: Lung, Breast, ... Millions of molecular features
This is the State-of-the-Art Statistical Model for Adaptive Trials: And this is the State-of-the-Art Algorithm:
GCTA is a redesign of clinical research in the image of a “Learning Air Traffic Control System”, where entropy - minimizing reinforcement learning routes hypotheses in real time to the patients who at the same time have the most to gain, and the most information to offer to the overall system.
GCTA is NOT a fantasy! The VA just hasn’t put AI Engineers on the problem yet!
Recommend
More recommend
