Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Talk about “AI” in the press these days tends to conflate two things

nostalgebraist.tumblr.com

Post reply on HN
Print view View original

11–20 of 78 posts

Re: Talk about “AI” in the press these days tends to conflate two things

#11

Post by abhinav22 »

KingOfCoders wrote: post #8
abhinav22 wrote: post #6

Earlier quoted context omitted.

Because it’s regression and statistical analysis. Nothing special about ML except applying it to large(r) datasets, particularly in areas where it works well (various recommendation based models for example).

How does this differ from you "learning" words of a foreign language?

It’s not “learning” anything though. Difference is causation vs correlation:

If I learn how to boil water for example, there is close to 100% causation that heating water causes it to boil (notwithstanding we do not know everything about Physics, eg before we would say the earth is flat because we learnt it so, but that was incorrect, ie not everything we learn is true, it’s just a super high level of correlation)

However if a machine learning model says the probability of water boiling is 99% after heating water, that is actually correlation. So I would distinguish it from learning

In this example there’s not much different. But when you apply to more complex examples, the difference between causation and correlation become clearer

Re: Talk about “AI” in the press these days tends to conflate two things

#12

Post by DubiousPusher »

abhinav22 wrote: post #11
KingOfCoders wrote: post #8

Earlier quoted context omitted.

How does this differ from you "learning" words of a foreign language?

It’s not “learning” anything though. Difference is causation vs correlation: If I learn how to boil water for example, there is close to 100% causation that heating water causes it to boil (notwithstanding we do not know everything about Physics, eg before we would say the earth is flat because we learnt it so, but that was incorrect, ie not everything we learn is true, it’s just a super high level of correlation) Ho…

You think people's ability to learn is really distinguished by an ability to discern causation and correlation? That does not comport with modern behavioral science.

Re: Talk about “AI” in the press these days tends to conflate two things

#13

Post by DubiousPusher »

BoiledCabbage wrote: post #7
forgingahead wrote: post #5

Earlier quoted context omitted.

Why? The machine learns via essentially trial-and-error to produce a model that best fits the data provided to it. The hope is that the training data is representative of "unseen" data the model later encounters in order to be useful. Seems pretty accurate given the broadness of the topic and variety of use cases.

The question is: Is the machine learning, or is the machine intelligent. The answer is its neither. But we've taken them to be useful analogies and loosely designated them to mean different things. Neither one really applies because in truth, the machine isn't learning. Because we've deemed some types of organs optimizers, others dynamic programing and others learning. It's all arbitrary.

I don't believe it is entirely arbitrary. Just as there are categorizations of machine, turing machine vs push down automata vs finite state machines, which determine their ability to solve certain problems, there are differences between categories of algorithm and their ability to solve certain problems.

There are many "learning" algorithms beyond what is popularly called ML today but they have in common that they use information about their own success or failure to inform future results.

Re: Talk about “AI” in the press these days tends to conflate two things

#14

Post by DubiousPusher »

I have certainly noticed this but am largely unsurprised by it. I've been much more shocked by the number of people I've worked with in the tech industry who believe that #2 is far more advanced than current evidence suggest and that it will lead inevitably to #1.

I'm not opposed to the hypothesis but the number of people I've encountered who really believe current heirarcical neural nets are the end all tech is truly head scratching. Especially when so many technologies such as the Internet have not produced certain expected results and completely surprised with such unexpected ones.

Re: Talk about “AI” in the press these days tends to conflate two things

#15

Post by abhinav22 »

DubiousPusher wrote: post #12
abhinav22 wrote: post #11

Earlier quoted context omitted.

It’s not “learning” anything though. Difference is causation vs correlation: If I learn how to boil water for example, there is close to 100% causation that heating water causes it to boil (notwithstanding we do not know everything about Physics, eg before we would say the earth is flat because we learnt it so, but that was incorrect, ie not everything we learn is true, it’s just a super high level of correlation) Ho…

You think people's ability to learn is really distinguished by an ability to discern causation and correlation? That does not comport with modern behavioral science.

To some degree yes. For example, can the fields of mathematics and advanced physics or programming be studied to the degree they are with an understanding of causation?

For some things, correlation can be sufficient, so I can see where you are coming from.

As a better counterpoint: if we could not distinguish between causation and correlation, why is general AI so hard?

Re: Talk about “AI” in the press these days tends to conflate two things

#16

Post by dennist »

I find a lot of the media (here in Japan anyways) call literally every automation by computers or machines AI. Example, there’s an automated gate that only opens if a scanner finds your body temperature is not too high. AI Gate. The other day was something related to automatic checkout at shops or something where it all boiled down to some computer reading QR codes

Re: Talk about “AI” in the press these days tends to conflate two things

#17

Post by yk »

That's actually a special case of a general phenomenon, which I seen here on a link on HN being called the Conway or Neumann problem (in any case, that blog post did purposefully misattribute for a joke, so I am not too concerned in getting the name right). The thing is, you're reading an newspaper article on anything where you can evaluate the quality with any kind of confidence, and it is invariably garbage, much too much simplified, glances over the interesting bits, and obviously the journalist did not understand the topic. And then you're turning the page, and expect the article on anything else to be much better.

Re: Talk about “AI” in the press these days tends to conflate two things

#18

Post by DubiousPusher »

abhinav22 wrote: post #15
DubiousPusher wrote: post #12

Earlier quoted context omitted.

You think people's ability to learn is really distinguished by an ability to discern causation and correlation? That does not comport with modern behavioral science.

To some degree yes. For example, can the fields of mathematics and advanced physics or programming be studied to the degree they are with an understanding of causation? For some things, correlation can be sufficient, so I can see where you are coming from. As a better counterpoint: if we could not distinguish between causation and correlation, why is general AI so hard?

> can the fields of mathematics and advanced physics or programming be studied to the degree they are with an understanding of causation?

Yes, pretty much. Newton's laws are a perfect example of something that establishes a correlation but are useful absent accurate causation.

They did just fine until we needed something better. Einstein provided this but can we be so sure gravity as a result of curved space is "true" and not just a really good model? Perhaps the hunt for what quantum gravity will one day tell.

> As a better counterpoint: if we could not distinguish between causation and correlation, why is general AI so hard?

I think general AI is hard because for one it might be a fiction. For general AI to be real, we'd have to assume people posses "general intelligence". Or at the very least, were it beyond us, we'd have to assume we could tell if we had created it.

What we are really trying to do with "general intelligence" is reproduce something that took 3.5 billion years of natural selection to design. We have only recently become capable of even creating machines with the kinds of part counts that biology comes with. Were we to produce "general intelligence" in the next thousand years I'd be surprised it came so easy.

Re: Talk about “AI” in the press these days tends to conflate two things

#19

Post by NalNezumi »

dennist wrote: post #16

I find a lot of the media (here in Japan anyways) call literally every automation by computers or machines AI. Example, there’s an automated gate that only opens if a scanner finds your body temperature is not too high. AI Gate. The other day was something related to automatic checkout at shops or something where it all boiled down to some computer reading QR codes

Yea In Japan the AI hype (or fad) arrived with like 3-4 years of delay compared to the west. Thing is it works so well for startups to get funding, it's not just media but investors too. Worked for a startup which sold "AI" interact with customer by detecting human face (and age, gender etc) and having targeted conversation. Turns out it's just a facial recognition and replay of prerecorded voices. I felt really dumb when customers, investors & the CTO somehow were impressed by this basic functionality.

Re: Talk about “AI” in the press these days tends to conflate two things

#20

Post by slavik81 »

yk wrote: post #17

That's actually a special case of a general phenomenon, which I seen here on a link on HN being called the Conway or Neumann problem (in any case, that blog post did purposefully misattribute for a joke, so I am not too concerned in getting the name right). The thing is, you're reading an newspaper article on anything where you can evaluate the quality with any kind of confidence, and it is invariably garbage, much t…

You're referring to the "Gell-Mann Amnesia" effect.
Post reply on HN