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

No, machines can’t read better than humans

theverge.com

Post reply on HN
Print view View original

41–47 of 47 posts

Re: No, machines can’t read better than humans

#41

Post by romaniv »

make3 wrote: post #23
Karnickel wrote: post #18

Earlier quoted context omitted.

> I'm pretty sure no one is seriously remotely claiming that machines understand text better than humans. Just some of the headlines: - http://money.cnn.com/2018/01/15/technology/reading-robot-ali... "Computers are getting better than humans at reading" - http://fortune.com/2018/01/15/artificial-intelligence-ai-chi... "Computer AI Can Now Read Better Than You Do" - https://gizmodo.com/ai-used-to-sell-you-more-stuff-c…

I meant the scientists, not the tech journalists. Clearly tech "journalists" is a very loose term nowadays

Scientists often carefully construct tasks, evaluation criteria and even test sets to make AI seem better than humans or at least better than it really is.

Re: No, machines can’t read better than humans

#43

Post by ghaff »

freehunter wrote: post #37
ghaff wrote: post #32

Earlier quoted context omitted.

>Journalism went out of business a long time ago because no one will buy it. With some exceptions. And there's also good tech commentary from various people who get food put on their plate from other sources which admittedly creates at least the potential for greater biases that aren't obvious. Although bias-free has always been an aspiration that generally hasn't existed in its purest form in practice. And a lot of…

>With some exceptions. I would argue, with very few exceptions. The NYT is capable of some great journalism. Same with WSJ and a number of other outfits. That doesn't change the fact that their opinion articles share the same domain and letterhead in an intentional effort to confuse facts with truth to gain clicks. To flip the argument around: Buzzfeed has some surprisingly good journalists working for it, including…

I don't really disagree with any of that. Richard Salant, the former president CBS News, supposedly once said "Our job is to give people not what they want, but what we decide they should have." Arguably it was a bit arrogant at the time but it's certainly not something I imagine coming out of the mouth of an executive at any clicks-driven mainstream news business today.

I still get Time Magazine (though I'll probably let my subscription lapse soon) mostly out of habit because I've gotten it for decades. Is there still good writing and journalism there? Yeah but it's also got all sorts of infographics and fluffy opinion pieces and lifestyle articles. (Always has of course but the mix has certainly shifted.)

Re: No, machines can’t read better than humans

#44

Post by ghaff »

foobiekr wrote: post #38
freehunter wrote: post #33

Earlier quoted context omitted.

To be fair, that's what we do to humans as well. We feed them all the information they can consume, and after it's all been processed, then we try to figure out what all the output means. From that point forward, we heavily specialize them until their general purpose skills are basically deprecated in favor of their new specialization. Humans aren't born with the capabilities they possess at adulthood, they're fed da…

There is actually not a good reason to believe that humans - like most animals - are not born with a with a significant set of capabilities that they use through their lives into adulthood. Facial recognition as a function in humans is not a function that magically evolves by the presentation of data, neither is language acquisition.

Furthermore, a young child may need to be shown an object, like a cup of water, a few times before they consistently recognize it. But they don't need to go through supervised learning with a 10,000 image training set with all kinds of permutations of shape, lighting, and color.

Re: No, machines can’t read better than humans

#45

Post by freehunter »

foobiekr wrote: post #38
freehunter wrote: post #33

Earlier quoted context omitted.

To be fair, that's what we do to humans as well. We feed them all the information they can consume, and after it's all been processed, then we try to figure out what all the output means. From that point forward, we heavily specialize them until their general purpose skills are basically deprecated in favor of their new specialization. Humans aren't born with the capabilities they possess at adulthood, they're fed da…

There is actually not a good reason to believe that humans - like most animals - are not born with a with a significant set of capabilities that they use through their lives into adulthood. Facial recognition as a function in humans is not a function that magically evolves by the presentation of data, neither is language acquisition.

The ability to learn a language is innate, sure. That actual process of learning a language is very, incredibly data-heavy, to the point where many people never do it more than once because of how much data it requires.

The ability for a neural network to learn something is innate as well, because we programmed it to be. But to actually get it to learn something, again, takes just a ton of data.

Re: No, machines can’t read better than humans

#46

Post by freehunter »

ghaff wrote: post #44
foobiekr wrote: post #38

Earlier quoted context omitted.

There is actually not a good reason to believe that humans - like most animals - are not born with a with a significant set of capabilities that they use through their lives into adulthood. Facial recognition as a function in humans is not a function that magically evolves by the presentation of data, neither is language acquisition.

Furthermore, a young child may need to be shown an object, like a cup of water, a few times before they consistently recognize it. But they don't need to go through supervised learning with a 10,000 image training set with all kinds of permutations of shape, lighting, and color.

I believe that's more a limitation on the information the AI is taking from the data we present it. There's no question that a human can learn more from being presented with an object than an AI can learn from being presented 10,000 pictures of that object, what's unclear is if this is a fundamental difference or something that could be overcome by reevaluating what we're teaching and how.

Re: No, machines can’t read better than humans

#47

Post by cpfohl »

onychomys wrote: post #17
cpfohl wrote: post #11

Earlier quoted context omitted.

Yeah. They also don't carry with them the global context they'd need to correctly interpret these. When I say "Bush spent millions fighting his father's war." (For the record, this is a non-political statement: I'm using a sentence that conveniently explains my point). You and I both immediately gather from the context that "Bush" is not a shrubbery, it's a US president. You then gather that it was Bush Jr. Not H.W.…

It's a fair point, but plenty of humans don't actually have the background to know all that stuff either. If you told a eight-year-old that Bush spent millions fighting his father's war, the kid would know that Bush was a person not a shrubbery, but that's about it. The child could read and understand every word of the sentence, but she'd lack all the same context the computer does.

I think that's beside the point. I could come up with a more general example too...
Post reply on HN