Explain It to Me Again, Computer

http://www.slate.com/articles/technology/future_tense/2013/02/will_computers_eventually_make_scientific_discoveries_we_can_t_comprehend.html

by prostoalex • 13 years ago

41 24 13 years ago

slate.com

derekp7• 13 years ago

I don't really think it matters if we can fully understand a theory, as long as that theory makes predictions that can be verified. For example, if a computer-generated theory gives the ability to build a radio that is a million times more efficient at spectrum utilization, and products can be built from it, that theory is still useful even if it is uncomprehensible.

JulianK• 13 years ago

I very much agree that you don't have to understand the details of how something works for it to be useful. Whether that's a theory, a car or a computer, the world is full of things that are too complex for me to understand _fully_.

But if you don't truly understand something, I think you've given up control of that thing to someone else. Someone else has to do the thinking for you.

Once that happens, you'll have a hard time doing something new in that area, or even applying whatever it is you're supposed to know to new areas.

Understanding isn't the only thing that matters, but it still matters.

11001• 13 years ago

Nobody (well except for Chomsky maybe[0]:)) is arguing that incomprehensible is useless. This article is about a deeper question of whether or not we have reached our cognitive limits in terms of understanding the surrounding world, however we may have developed an incredibly powerful theory that can describe some aspects of this incomprehensible world to us.

[0]I actually don't know what I'm talking about.

Xcelerate4• 13 years ago

Very good article. It makes the point that science is heading in an unprecedented direction where the complexity of what is happening is too difficult for a single human mind to understand. In fact, I would say it's almost arrogant to assume that our 1100 grams of processing power is capable of understanding the incredibly complex interactions that build and build upon themselves.

This is the reason for computational science. Imagine trying to render a photorealistic scene analytically. There's just no way you could do that. Exactly integrate the rendering equation over complex geometries? It's unfathomable. There's a reason the Schrodinger equation has only been solved analytically for systems of just a few particles (not to be confused with exactly, which is possible with numerical methods).

Eventually science experiments will be conducted by setting up your initial system, plugging in the relatively simple laws of physics, and letting it go (well... if we solve the n-body problem).

I don't think biology will ever be "understood" in the same way that derivatives or diffusion or gravity is understood. There's so many levels of nested complexity that it is totally mind-boggling, and I'm surprised more biologists aren't impressed by the sheer fact that any of it works at all.

Retric3• 13 years ago

I don't think what your describing is science. At it's core science is a question of experimentation and running complex simulations can't truly provide new information, just validation. That said, there is a great temptation to call many things science because it lends credibility even if you go though the motions and ignore why honest experementwtion has value in the first place.

Xcelerate4• 13 years ago

What?!?

>Running complex simulations can't truly provide new information, just validation.

This is entirely and absolutely false.

I think you may want to read up on simulation (FEM, molecular dynamics, etc.). It's my research field. For instance, Quantum Monte Carlo has been used to calculate energy levels to a ridiculous level of agreement with experimentally measured values. There's plenty that can be learned about reality that can't currently be studied through experimentation because it's either 1) too expensive or 2) impossible to study with current technology.

Also, I wouldn't say science is "about" experimentation. It's about prediction. Any theory that can more accurately predict a system's future state than another theory is the better science. And simulations can do this better in many cases than experiment.

nnq• 13 years ago

> I don't think biology will ever be "understood" in the same way that derivatives or diffusion or gravity is understood. There's so many levels of nested complexity that it is totally mind-boggling, and I'm surprised more biologists aren't impressed by the sheer fact that any of it works at all.

...yet at the same time biology and medicine will probably be the hardest fields for computers to "make discoveries in" for quite some time, except low level molecular biology things like low scale protein interaction or selecting from dug-candidate molecules. There's so much data so "badly formatted" and full of ambiguity that I can't even imagine when an AI will be able to read medical research papers and make sense of them - it's hard even for humans. Otoh there will be tons of data where computers will be able to search for patterns, but the actual knowledge necessary to transform "patterns" into "discoveries" is in the icky ambiguous format of papers and books that are impossible to make sense of for any less-than-human-level-AI.

jpxxx• 13 years ago

I think this is giving humanity way too much credit. We're only going to care about profitable ideas that algorithms come up with.

Imagine an electronic intelligence with a profound understanding of TV shows. Who would possibly want to hear what it has to say?

saulrh• 13 years ago

  Who would possibly want to hear what it has to say?

Television executives.

Everything has an application, and many fewer of those applications are useless than you'd think. Plus there are higher-order effects to consider, mostly related to the tools people develop to attack interesting problems. The classic example is the space race: how useful is it to be able to fire four-hundred-foot-tall fuel tanks into space?

mrcsparker• 13 years ago

I would. An electronic intelligence with a profound understanding of anything would be interesting.

If I said that, in the next room, there was a computer with a profound understanding of TV shows - wouldn't you be interested? That would be fascinating.

BariumBlue2• 13 years ago

We know that the mind is Turing complete. What the author is suggesting is that it's possible for there to be some science or mathematics that is beyond the capability of our mind to comprehend. (and that it's possible that there's science or mathematics that may be derived by computers that we may not fundamentally be able to understand)

However, any attempt to compute or derive new science/mathematics outside the theoretical capabilities of our (Turing complete) minds on merely Turing complete computers is doomed to failure (at least, until we get some computers better than Turing complete ones).

comex• 13 years ago

Turing complete is really only meaningful for a computer with infinite time and storage; humans are limited in both. Just as a smarter human can understand something a less smart human can't, computers will eventually understand many things we can't wrap our heads around.

BariumBlue2• 13 years ago

True, but assuming in this future humans have developed brain enhancements, the issue of storage space and computing power becomes less relevant. Though I suppose the whole idea of brain enhancements would render the author's point null anyways...

TeMPOraL2• 13 years ago

> and that it's possible that there's science or mathematics that may be derived by computers that we may not fundamentally be able to understand

I think what the author meant is that us understanding this "new science" would be like trying to run StarCraft II on 386. It's not that it is not fundamentally possible, it's just that our wetware has nowhere near the required computational power.

lifeisstillgood2• 13 years ago

Science is not an agglomeration of random observations and facts - it should not get more complicated as more is discovered - if the theories and tools like maths we have do not explain the theories then we are doing something wrong with the fundamental tools (maths stats)

I accept there is more and more unknown or unexplained - but that is not to say the theories we have are not explained with the same set of "unreasonably effective" tools

TeMPOraL2• 13 years ago

> it should not get more complicated as more is discovered

But it does for the simple reason, that you need to read more and more papers just to know whether or not someone didn't already do your research, or if there isn't already a growing body of evidence against the theory you're trying to develop.

http://lesswrong.com/lw/kj/no_one_knows_what_science_doesnt_...

Scientific progress is about to hit a self-complexity boundary due to the amount of knowledge we amassed - we have a huge overhead on adding any single new piece to a puzzle. Our brains are not good enough for handling it anymore, we need computers to help.

EDIT

And also because of that, I believe there's a huge amount of undiscovered facts lurking in all the discoveries we already had described, that no one noticed before because it requires correlating different papers from different disciplines (think about the more messy fields, like biology/medicine). Finding a way for computers to effectively do that work for us will likely result in some amazing progress.

lifeisstillgood2• 13 years ago

Is that really true?

If so one should force the release of copyright papers from the publishers at gunpoint tomorrow. And make google dedicate half it's employees to indexing it

I suspect though that there are so few scientists working on most areas that the research is old or outdated across the majority of fields.

And it rather indicates a lack of fundamental tools - for example studying insect species - it's well know that there are millions of species to every insectologist(?) - yet Linnaeus classification is three centuries old - one could imagine a production line of genome mapping - put insec into machine out comes its genome - oh look that is not the species we were thinking of

The tools (mechanical) and the tools (mental) need to integrate large branches if science are what is needed - not a fear over the vast number of facts we have acquired

benbou09• 13 years ago

Interesting article, but I think that the humans will probaly evolve in ways that will extend their capacity of understanding. It may start by people living longer and then someday they will have memory implants and maybe even processing power implants.

nova• 13 years ago

It doesn't help that most research papers are behind paywalls, or in forgotten books long out of print but still in copyright (and owned by the publisher, not the author), buried and not scanned theses, or technically published in patents but then also unusable by anyone interested.

Explain It to Me Again, Computer

Related Stories