Michael Fossel Michael is President of Telocyte

January 9, 2017

Conceptual Blinders

 

A week or so ago, an AI beat the world’s reigning champion in the game of Go.

The odd thing is not that it happened, but how it was done. By itself, the victory would just be one more example of “computers beating humans”, but there is a far more interesting and important facet to this event. Not only did the AI beat the world’s Go masters and the reigning world champion, but it did it, not by being better at using the known strategies and tactics, long the province of Go adepts, but by using “unconventional positions“ and “moves that seemed foolish but inevitably led to victory” (WSJ, January 5, 2017). In short, the AI went into playing the game without conceptual blinders. It developed novel (and effective) strategies based on reality, rather than on preconceived views of how the game “ought” to be played. Had the AI been programmed by Go masters, it wouldn’t have fared as well. It succeeded because it lacked the limitations that we as human beings unknowingly use when we approach a problem.

go-game-boardIF our assumptions create limits, then our outcomes are limited.

The same problem – our own assumptions – proscribes the limits of what we can do in science and medicine. If we simply program a computer to “delay the onset of Alzheimer’s disease by lowering all known risk factors”, it might succeed, but the solution would be limited by how we set up the problem. In short, assumptions limit outcomes. If we merely restrict the program to lowering risks, then a computer program can’t show us how to cure Alzheimer’s. Such a program might, for example, recommend dietary changes, moving away from major highways and pollution, lowering blood pressure, avoiding infections, improving dental hygiene, lowering stress, and a myriad other changes that might delay Alzheimer’s. But the programs, the questions we pose, presuppose that Alzheimer’s can’t cured or prevented, only delayed. If we preclude finding a way to win, then all we find is a better way to lose.

Consider the historical analogs. If I want more efficient communication, I don’t ask a computer to design a better telegraph. If I want more efficient transportation, I don’t ask the computer to design a faster horse. If I want to cure polio, I don’t program a computer to design a better iron lung. And if I want to cure Alzheimer’s, I shouldn’t design a better way to attack amyloid, tau proteins, inflammation, or mitochondrial dysfunction. Merely because I’ve already assumed that those are the only strategies, I have limited my outcomes. If Alzheimer’s interventions are restricted to merely optimizing old strategies, we will never cure it.

Why be satisfied with a better telegraph, a faster horse, or a more efficient iron lung?

Programmed solutions, based on preconceived limits are a case of GIGO: “garbage in, garbage out”. True advances in science and medicine are not incremental; they demand innovative perceptions and constant reexamination of our premises. The example of an AI beating the world’s reigning Go champion wasn’t the result of incremental improvements in coding all of the Go strategies known to previous champions into a program and then tasking the program with implementing those accepted strategies. The AI was tasked with winning, regardless of previously accepted strategies. As a result, the AI actually WON, unexpectedly, but reliably, using innovative, startling, and unexpected approaches.

If we want to cure Alzheimer’s disease, we can’t use incremental approaches to time-worn (and uniformly ineffective) strategies. Like the AI playing Go, we need to stop focusing on accepted strategies and ask the fundamental question: how do we win? Not “how do we optimize the same old strategies?”, but how do we actually WIN? We shouldn’t rely on “programmed” approaches; we should toss out our preconceived programs, and ask how to win. With regard to Alzheimer’s disease, we need to stop asking how to optimize losing strategies and ask how to cure Alzheimer’s. Not “how do we lower amyloid levels?” or “how do we reduce tau tangles?”, but how do we cure and prevent the disease in the first place? If we really want to make a difference, then we need to free ourselves from our preconceptions and our old programming, and begin to ask the fundamental question: how can we cure Alzheimer’s?

Truly innovative approaches demand a ruthless reassessment of our assumptions.

We will cure Alzheimer’s only if we have the wit to truly use our own intelligence, with honesty, perceptiveness, and a willingness to examine reality.

No Comments

No comments yet.

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.

Powered by WordPress