Kurzweil: Human-Level Machine Translation By 2029 186
An anonymous reader writes "In a video interview with the Huffington Post, noted futurist Ray Kurzweil predicts that machines will reach human levels of translation quality by the year 2029. However, he was quick to highlight that even major technological advances in translation do not replace the need for language learning. 'Even the best translators can't fully translate literature,' he pointed out. 'Some things just can't be expressed in another language.'"
The vodka is good... (Score:3, Funny)
Re: (Score:3, Funny)
but does it understand the binary language of moisture vaporators?
Hello computer (Score:2)
Hello computer
Re: (Score:3)
*hands Joe a mouse*
Re: (Score:3)
How quaint !
Re: (Score:2)
Re: (Score:3)
Not mouse, MOOSE!
Ray Kurzweil's predictions (Score:5, Insightful)
You know, I'm a big sucker for futurism as anybody, but Ray Kurzweil makes a lot of predictions about future tech every couple of years, most of which never pan out anywhere near what he predicted. And each time Kurzweil makes a prediction, many of which are just way too optimistic or just play goofy in retrospect, the tech-minded people like slashdot lap it up.
Can't tech futurists find a better spokesman than Ray Kurzweil?
Re: (Score:3)
Re: (Score:3)
Re: (Score:2)
Re:Ray Kurzweil's predictions (Score:5, Funny)
I predict that in 2029 computers will surpass Ray Kurzweil in making overly optimistic predictions. :-)
Re: (Score:3)
I think it's almost time to start adding Ray Kurzweil to my annual Dead Pool list.
It must be so frustrating to be so close to immortality that you can taste it but to know you'll never achieve it.
Sorry Ray, but old is old. You could have saved yourself a ton of money and effort and just spent a little time practicing Tai Chi and you'd probably have lived longer than you will.
As arrogant as Kurzweil strike
Re: (Score:2)
If I knew dead pool, I would definitively put Kurzweil on his list.
Just sayin'.
No man ahs done so little and leveraged is so much. What a boring has been.
Re:Ray Kurzweil's predictions (Score:5, Insightful)
Yeah, but for years, machine translation was stuck on alta vista's dreary babelfish... which was basically a one to one dictionary translation (often without using the right definition) for hilariously bad translations.
A couple of years ago, Google translate gave a big bump to the whole concept using UN documents (which are usually in 5+ languagels) as a reliable translation. It has a lot of hiccups, but translations often went from unreadable babble babel to something that often ranged from a decent translation to something you can figure out if you put some thought into it.
I have done a lot of work with translators and even they get things wrong, so I think Kurzweil is actually off in a way. IMO, by the end of this decade, machine translation will often be good enough (really, google translate needs to start looking for more context cues and I can't think that will be 19 years away) but there will never be perfection because language itself isn't perfect. Look at humans communicating sometime, it's not a strict protocol, can misunderstandings happen all the time between people. But when a machine gets it wrong, people will point to it as bad, instead of the nature of language itself.
Re: (Score:2)
I agree. Google Translate works great, especially between Romance and Germanic languages. The Chinese to English has some hiccups but it works well enough to be functional. I don't think I've had to use it with any other languages, but I would expect similar functionality.
Will we ever have Douglas Adam's Babel Fish? I'm sure we'll get something close enough such as a Google Translate realtime Android app or something of the sort - it would input what's being said through your phone's microphone and then out
Re:More than machine translation... (Score:3)
...we need reliable voice recognition and automatic transcription. Google is also working on it, but
Smartphones need a better input device than a tiny QWERTY keyboard
Re: (Score:2)
Sorry, hit the button without proofreading...
Previous post should read: "Google is also working on it, but it's still far from reliable"
Re: (Score:3)
It depends how much time we waste with phantom terrorists.
We wasted an entire DECADE on a false crusade with a disastrously mismanaged run against Osama. "Let's go raid Afghanistan! Where has he been for four years? Pakistan!"
Notice how the media posted five stories then shut up?
X trillion dollars later we're whining about budget crises.
If we had spent that time and money on tech, we'd BE in Kurzweils's land.
Re:Ray Kurzweil's predictions (Score:4, Funny)
Ironically, those wars have put a lot of money into translation applications.
Re: (Score:3)
Agreed. For me, the turning point was about two years ago when I was reading a report of a convention in Poland. It took me about halfway down the
Re: (Score:2)
Google Translate is really good at what it does -- it's one of the best systems, but it's fundamentally not a human level translation, except for those cases where it's seen something translated word-for-word by a human (e.g. UN documents), and even those can be wrong if they are used out of context.
There's a big difference between readability and accuracy. Most existing machine translation systems do not understand the semantics behind what they translate. It's not super hard for a computer to generate pag
Re: (Score:2)
So they have translation which works mostly, the problem is that difference between mostly and correctly is a big step and involves context, which the computer translations struggle with, it;s the difference between and idiot translating mechanically and someone who actually speaks the language
The problem is context, the solution is AI ...
Your post translated and back again on Google (Score:5, Funny)
Yeah Google translate isn't THAT good at this stage. Here's what it did to your post. My favourite mistranslation: was in Russian. "I did a lot of work with translators and even they get it wrong, so I think Kurzweil is actually a way" can be taken to mean the opposite of "so I think Kurzweil is actually off in a way." which is very misleading. I'd use it cautiously for casual conversation even now, knowing there will be misunderstanding. But I wouldn't dare dream of reliability for anything critical in a decade. It would only need to tranlsate "Well we don't want war" to "We want war" once to kill millions.
Here is your post, translated English -> French -> English
Yeah, but for years, machine translation was blocked on Alta Vista babelfish gloomy ... which was essentially a one to one translation of the dictionary (often without the help of the definition of right) to hilariously bad translations.
A couple of years, Google translate gave a big bump in the whole concept of using United Nations documents (which are usually 5 + languagels) as a faithful translation. It has many failures, but translations of Babel Babble past often unreadable for something that often range from decent translation of something that you can see if you put some thought into it.
I did a lot of work with translators and even they make mistakes, so I think Kurzweil is really a way out. IMO, by the end of this decade, machine translation will often be good enough (really, Google Translate should start looking for context clues and I do not think it will be 19 years away) but n 'there will ever be perfect because the language is not perfect. Watch humans communicate sometimes, it's not a strict protocol, misunderstandings can occur any time between people. But when a machine is wrong, people will point to it as evil, instead of the very nature of language.
Here is your post, translated English -> Chinese -> English
Yes, but over the years, machine translation is boring to stay in Alta Vista's Babelfish platform ... which is basically a one to one translation dictionary (usually not the right to use the definition of) the cheerful bad translation.
A couple years ago, Google translation of a big jolt to the whole concept of using United Nations documents (which is usually 5 + languagels) as a reliable translation. It has a lot of hiccups, but the translation is often unreadable babble from Babel things, often translated from a decent range of things you can calculate, if you want to go some.
I've done a lot of work, even their translation wrong, so I think Kuziweier actually way off. Haishizuzhi, by the end of this decade, machine translation is often not good enough (really, Google translator need to start looking for more context clues, I can not believe it will be 19 years later), but there will never be perfect because language itself is not perfect. A look at human communication, it is not a strict protocol, can be misunderstanding between the people what happened. However, when a machine to get it wrong, people will point to it's bad, rather than the nature of language itself.
Here is your post, translated English -> Russian -> English
Yes, but for many years, machine translation is stuck in a dreary Babelfish Alta Vista in ... which was basically the dictionary translation 12:59 (often without proper identification) to have fun bad translations.
A couple of years ago, Google translate gave a big blow to the whole concept of United Nations documents (which are usually 5 + languagels) as a reliable translation. It has a lot of hiccups, but the translations often traveled from unreadable babble babel to what is often varied from decent translation of the fact that you can find out if you put some thought into it.
I did a lot of work with translators and even they get it wrong, so I think Kurzweil is actually a way. IMO, by the end of this decade, machine translation will often be
Re: (Score:2)
Keep in mind that since you're posting the result of two translation passes, in order to get an idea of how good the translation is you have to take the square root of the error.
Yes, it's some pretty mangled English - but that's mangle^2, which can be significant and misleading.
Re: (Score:2)
Keep in mind that since you're posting the result of two translation passes, in order to get an idea of how good the translation is you have to take the square root of the error. Yes, it's some pretty mangled English - but that's mangle^2, which can be significant and misleading.
I'm not going to learn German, Russian or Chinese, and my French is terrible, so mangle squared it has to be. The fact that you'd refer to it as mangle is very telling
Re: (Score:2)
Holy cow! Google Translate is a LOT better than it used to be. All of those translations to/from actually came back semi-intelligibly.
Yes, I had the advantage of knowing the original post. It would have been fascinating to read the double translations first and then to read the original post to find out how much I was automatically translating in my head.
Absolutely incredible. I am amazed at how not-poorly Google Translate did.
Re: (Score:2)
Google Translate is now crowdsourcing corrections to its translations. You can see how this would help particularly for idiomatic expressions. With enough crowdsourced input, I don't see why it'd take more than 5 years to get to human-quality translations of most prose texts.
Re:Ray Kurzweil's predictions (Score:5, Interesting)
That's almost exactly what I was going to post. Kurzweil will say anything to get his name in the news. While I'm sure he's a most interesting conversationalist, his predictions usually make me yawn. They're either too obvious or he anticipates they'll take place so far in the future that it amounts to nothing more than a guess. I assume he puts a lot of thought and research into his predictions, but his success rate seems to be no better than that of sci-fi authors.
Take Fahrenheit 451, replace literal book burning with figurative book burning, and what do you have? Society today.
To me, it seems like Kurzweil's always trying to motivate the scientific community to make him immortal. He was on Real Time with Bill Maher the other day and it was hilarious how excited Kurzweil was over the prospect of immortality whereas Bill found the idea humorous. It's like futurism is Kurzweil's religion: he sees it as the path to eternal life as long as he can rally the scientific community behind his ideas before he dies. So in a way, he's trying to create self-fulfilling prophecies rather than truly predict what will happen.
Re:Ray Kurzweil's predictions (Score:4, Informative)
Kurzweil is an eccentric charlatan who is making money off the gullible nerds who can't see through is bs. If anything he is a good manipulator.
People believe what they want to believe because it takes too much energy, talent and hard work to build sufficient knowledge to see through the bullshit. Most people who believe in that futurism have no background in any major scientific discipline related to his predictions what-so-ever. Big companies like Intel thought in the future we would have 10 Ghz processors unfortunately it turned out leakage and heat became a real issue and we get multi-core CPU's, more ghz may come back eventually but it will take new discoveries/processes that will likely take decades to complete. Sometimes technology gets stuck for a long time until solutions are found or something entirely different altogether emerges.
Re: (Score:3)
That's a good point. The biggest problem with predicting future technologies is predicting the bumps in the road along the way. The Mesopotamians progressed technologically quite quickly, especially for their time. But Europeans during the medieval period progressed quite slow, especially for the time. Not to mention that the bumps in the road won't always be technological. Look at the political opposition that recently occurred with stem cells. Or to go with a more well known, historic example, the conflic
So what is his success rate? (Score:3)
He claims 86% [singularityhub.com].
Got any alternate figures, or some sci-fi authors we can compare to?
Re: (Score:2)
Re: (Score:2)
Would you say that 86% of his predictions were obvious? Would you have said that in '98?
Granted, the ability to accurately predict the unexpected is more useful (and harder), but plenty of things that seemed obvious at the time haven't really come to pass (a lot of people expected flying cars by now, for example). I'd say accuracy was generally a more useful trait for predictions than non-obviousness, though of course you can take that to ridiculous extremes, as in your example.
Re: (Score:2)
Looking at a list of a few of his predictions, yes. They seem to fall into two categories: obvious and wrong. By obvious, I mean stuff that most people were predicting at the time. Or, in a few cases, 'predictions' that had already happened before he predicted them. To steal an analogy from another poster, it's like predicting that by 2020 we'll have TVs that are more than 3 feet in the diagonal, less than two inches thick, and can display high definition video - not very impressive, because it had alre
Re: (Score:2)
Noted futurist Joe User predicts... (Score:2)
Can't tech futurists find a better spokesman than Ray Kurzweil?
Sure, I've always wanted to be a futurist.
In the future, things will be similar, but different in interesting ways. $5, pay up.
Ok, on a serious note, evolutionary, not revolutionary changes. The average Slashdot style website is a BBS without the modems. We don't get flying cars because that's not where cars need to be, we get faster cars, more fuel efficient cars, more luxury cars and cheaper cars.
Also, in the future, Slashdot.org car analogies will be replaced with .Slashdot flying car analogies.
I dug through all the replies (Score:4, Interesting)
I dug through this thread looking for the surely inevitable reply to ask you for actual evidence to back your claim (that Kurzweil's predictions are often wrong), so I could mod it up.
I can't find one, so I sacrifice my option to mod this thread to call you out. Can you back up your claim?
I certainly don't think Kurzweil has been perfect in his prediction, but I think he does quite a good job. Here is my evidence: http://singularityhub.com/2010/01/19/kurzweil-defends-predictions-for-2009-says-he-is-102-for-108/ [singularityhub.com]
The predictions criticized in that article are definitely not entirely accurate, but they're also pretty damn good for having been made in 1998. We are close to where Kurzweil says we should be.
Please defend with counterexamples :-)
Re: (Score:2)
I dug through this thread looking for the surely inevitable reply to ask you for actual evidence to back your claim (that Kurzweil's predictions are often wrong), so I could mod it up.
I can't find one, so I sacrifice my option to mod this thread to call you out. Can you back up your claim?
I certainly don't think Kurzweil has been perfect in his prediction, but I think he does quite a good job.
From the article you linked to "As you can tell by my gratuitous use of hyperlinks, many of these technologies are in development if not in commercial use. None, however, are so widespread or dominant in their field that we can point to them and say “oh, well that’s obviously come to pass.". I think that sums it up nicely. While some of the things he talks about are now possible, that isn't a difficult prediction since a lot of these things were in early development. Predicting what will be comm
Re: (Score:2)
http://en.wikipedia.org/wiki/Predictions_made_by_Ray_Kurzweil [wikipedia.org]
He's been wrong about some (perhaps some would say many) predictions, but in my view I think he's pretty damn good in determining the general trajectory of the accelerating nature of tech.
Re:I dug through all the replies (Score:5, Insightful)
This is nothing other than the usual Kurzweil white-knighting. Going down the list:
1: The author defends the "computers in jewelry and clothing" prediction by pointing to smart phones, health monitors, and hearing aids. The latter two are not reasonably computers (and they existed in only marginally less advanced forms in 1998 already). Smart phones seem like a good defense here, but there is a fundamental problem. We can either say that smart phones are not clothing/jewelry (and therefore ineligible as defense of the prediction), or we can accept that they are -- but then why do the PDAs of 1998 not similarly qualify? The author is unable to produce a single example defending this that was not in some sense extant in 1998. Either the prediction has not been defended (because phones/PDAs are not clothing/jewelry), or it has been defended but is meaningless (because PDAs already existed in 1998 when the prediction was made).
2: The author doesn't even attempt to defend the actual ridiculous part of the claim -- that speech-to-text would account for the majority of text created.
3: The technology to project an image onto the eyes existed in 1998. Kurzweil once again managed the incredible feat of predicting the existence of something which already existed.
4: I don't know enough about chip fabrication to confirm or deny the author's argument. I'll accept it as probably true and say that this is one he got right.
5: Kurzweil predicted telephones capable of translation, and the author supports it by pointing to translation apps for smart phones. To anyone in 1998, the prediction meant a phone that could translate speech to another language -- that was the context in which phones were understood to function. They still do not do this (unless there are some new apps I'm unaware of). You could argue that Kurzweil is technically correct here, but I would say that (a) he is correct not through any foresight or wisdom, but because a wildly different application than what Kurzweil imagined came about and (b) that what we have is only an incremental update on the technologies of 1998 (eg electronic dictionaries) rather than the transformative capabilities Kurzweil predicted.
6: Kurzweil predicted drones would dominate combat, they don't. Author conveniently ignores this.
7: No need for me to comment, since even the author can't muster a defense for this embarrassment.
In short, what we see here is nothing different than what we normally see from Kurzweil-defenders (and Dead Sea Scrolls-defenders, and Psychic Hotline-defenders, &c). Six of the seven predictions are only accurate if you reinterpret the predictions to match reality (or worse, reinterpret reality to match the predictions), which frequently involves either neutering the interesting aspect of the prediction or making the prediction so vague as to be meaningless.
Kurzweil's ridiculous futurism is nothing but a religion for people who not uncommonly pat themselves on the back for being "too rational for religion."
Re: (Score:2)
Re: (Score:2)
I said it made the prediction meaningless.
Sure, I don't dispute that. Meaningless or not though, he's still right, even if a prediction like that doesn't make him a useful futurist.
Even so, technologies do get abandoned. In this particular case, the technology for head-mounted displays was relatively rare (labs + military) at the time, and now is available commercially in a much cheaper & more developed form, making it a lot closer to the sort of mainstream usage Kurzweil likely envisaged. Though I guess you wanted you could argue that he was w
Re: (Score:3)
1) iPod shuffles, health monitors, hearing aidsthese are all computers that can be worn on the body and fit the prediction.
This reads like a post-hoc rationalization... if someone were given the prediction they would not describe the given items. Predictions aren't validated because you can force things to fit them after the fact, they're valid because they predict and accurately describe what will be in the future before it is actually available.
2) Speech to text is gaining ground. It’s available in hand held devices, and as (semi) popular Apps on smartphones.
The prediction was that a majority of text is done by Speech-to-Text. "Gaining ground" doesn't give you credit.
5) Like speech to text, this technology is gaining ground and has related
Re: (Score:2)
He predicted we'd be interacting with our computers by voice commands now. We don't. But the technology is there and works pretty well (Kinect, iPhone). Because we find it a little creepy and feel like douches talking to a gadget. I have had voice dialing in my cellphone for a long time but never used it once. What makes it to market and what consumers adopt is impossible to predict it seems.
Re: (Score:2)
I've not come across a concept that can't be expressed in any human language, but there are some that can't be expressed easily. As a trivial example, German uses the same word for noodles and pasta. In English, if you say pasta, it's clear that you are referring to Italian food. In German, you need to explicitly disambiguate between Italian or Asian versions. In the other direction, consider words like zeitgeist or schadenfreude. These have both been adopted into English now, because there is no conve
Yes we can. (Score:2)
'Some things just can't be expressed in another language.'
Then it's not a human idea, and probably wasn't expressed in the original language.
Culture notes (Score:4, Informative)
Re: (Score:2)
If not for the colder parts of our planet, our understanding of chemistry might have lagged for thousands of years.
Re: (Score:2)
Re: (Score:2)
I think that's the first time in like a decade I've seen someone say that honestly without hedging in an online conversation. You deserve some sort of Internets medal, sir!
Re: (Score:2)
Italian does not differentiate between mitts and mittens - I guess there has never been the need.
Re: (Score:2)
Serendipity is a famous English example: the concept of finding something when looking for something else. For instance you might be looking down the back of the sofa for the TV remote, fail to find it (it's actually under the sofa) but you do find a £2 coin while you're there.
I disagree with the quoted point that some things are expressible in one language but not in another, but there are some ideas that have a single word in one language but not in another. As another example take Schadenfreude; I
Re: (Score:2)
Then it's not a human idea, and probably wasn't expressed in the original language.
How did you come to this conclusion? Languages have grown, changed and splintered over time often as a direct result of growing, changing and splintering human ideas. If anything I would think ideas would be the hardest thing to express correctly across different languages.
Re: (Score:2)
"can't be expressed" is overstating it, but there's concepts that have simple names in one language, and which would require essentially rewriting several paragraphs to convey in another language.
Yeah, it can be expressed, it's just a lot more hassle. Schadenfreude.
sisu == perseverance (Score:2)
Re: (Score:2)
Fortitude.
Re: (Score:2)
From your wikipedia link :
Deciding on a course of action and then sticking to that decision against repeated failures is sisu
The definition of insanity is doing the same thing over and over and expecting it to come out different
Therefore, sisu=insanity, right ?
Re: (Score:2)
Puns aren't expressed in language.
They're a deliberate breakage of a language.
The irony inherent in the breakage is why they're funny. Cross-coupling ideas that don't belong. Also funny when done right. Funnier when the new idea belongs but you hadn't quite got there yet (comedic timing is about getting there just slightly before it bubbles up from the observer's subconscious to conscious). But then you can explain it in language.
Even if it's a pun. Funny how that works.
Re: (Score:2)
Who pays Kurzweil for these "predictions"? (Score:2)
I can pull stuff out of my ass too, and I'd like to get paid for it.
Re: (Score:2)
I predict... (Score:2)
He's full of bullsh*t. Watson proved that machine learning is possible and all that is needed is enough processing power / memory.
If you get to a place where you can associate every word sequence in a language to a correspondent word sequence in another, then we'll have almost perfect translators.
We'll be doing full translations a lot sooner. I predict this, so GIVE ME A SLASHDOT ARTICLE PLEASE! :)
Re:I predict... (Score:4, Insightful)
> We'll be doing full translations a lot sooner [than 2029].
Then we'll have it by 2029, won't we? Which is what he said.
Re: (Score:2)
"According to Kurzweil, machines will reach human levels of translation quality by the year 2029."
That's from the article, by the way.
I say, it'll happen way sooner. Not by 2029, I'd say that in 10 years we'll get there.
Re: (Score:2)
in 2029 would mean that year specifically;
after 2029 would mean any year from 2030 (inclusive) or beyond
Re: (Score:2)
well, "true AI" is relative. If you have enough computing to support a complex combination of sequences of words, sequences of sequences of words (etc), the current AI is enough. You just need enough processing and storage to be able to distinguish patterns within patterns within patterns.
That's my opinion obviously, but great steps have been made in this area lately, and I strongly believe in what I'm saying :)
Re: (Score:2)
Translation is not just about words and sequences of words. If you've ever seen a subtitled comedy show in another language, you know what I mean. Often language contains cultural references, well known in the country of one language, but unknown in the other language. Trying to translate a joke referring to such cultural reference into another language (while keeping it funny) will just fail if you try to do it literally. Even human translators have a hard time doing a good job in those cases. Often the be
Re: (Score:2)
"You may be a cunning linguist, but I'm a master debater!"
Re: (Score:2)
I predict that a tall, dark stranger will tell you that you know nothing about translation.
Re: (Score:2)
If you know the explanation, the machine will know it too. That's not the limitation, the biggest problem is the same expression used in different contexts. Markov models and neural networks can only get you so far if you can't analyse everything. That's why I say that what we miss is power
Ray Kurzweil is full of shit (Score:3, Insightful)
This guy has been wrong on his previous predictions as everyone already has been emphasizing, but what the fuck is the deal with such a specific year for his prediction? Why not round up to indicate it's a rough measure? 2029, really??
Re:Ray Kurzweil is full of shit (Score:5, Insightful)
Just like the Rapture dude, having a specific date makes it more credible. Kurzweil is nothing if not a master manipulator of credibility...
Re: (Score:2)
Well, he says "by 2029", so that's just the conservative end of his prediction. Could be sooner. He already says that mind uploading will be possible around 2030, though, and once you have that, you can just simulate the brain of someone who knows two languages and get the answer to any translation problem, so his prediction would have to be earlier than that.
He also says the technological singularity will happen around 2045, so maybe we shouldn't waste our time working on machine translation in the meant
Re: (Score:2)
Singularity will never happen.
It makes no sense. If we can create machines that have the things we would nee to achieve singularity, why the hell would the machines want to be tied down with us?
It's like hum saying, well, I'm going to merge my mind with a chimpanzee.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Well, there is a computer that figures out science on it's own, simply through observation. In fact, it has solved certain biological puzzles. It came up with 2 formulas that explain an observation, and they work. No human know why the formulas work, just that they do.
Think upon that.
And the machine goes ... (Score:5, Funny)
"wtf?!? stuf! lol cul8r"
"What did my son say?"
"Sir, he inquires if things are quite as they seem. He wishes to seek tranquility, though is in good humour and will be pleased to visit again with you anon."
in the year 2x2x (Score:2)
Media whore pulls predictions out of his behind. (Score:2)
Nothing so see here, move along.
Reverse trend (Score:2)
Kurzweil never predicted crowdsourcing. This [ycombinator.com] didn't make it to Slashdot yet, but apparently the creator of reCaptcha is launching a service of human-aided mass translation.
It might just turn out that language problems are easier to solve by throwing social networks at them rather than hardware. Even if we eventually get hardware that would be able to do it, it would then be used for other problems that computers are already better at than humans.
Here's how to translate when (Score:3)
Showing his age (Score:3)
translation is the art (Score:2)
of mapping one culture to another. machine knows nothing about that. you can load all the databases in the world, but what machine lacks: immersion in the culture, being part of it, have the attitude about it, system of values, so to speak.
The stuff Kurzweil says, well, it's meaningless He's a great inventor, yes, no one argues about that. But the *big picture* he's trying to promote - well, it's mostly about promoting himself.
Re: (Score:2)
"He's a great inventor"
hahaha. He rounded out some technologies, decades ago.
Re: (Score:2)
Re: (Score:2)
Yeah, just like a computer will never play chess at grandmaster level. On language you can do anything with enough data. Consider how easy translation is if you've got a table that translates every sentence or even every document that anyone is likely to write.
Language is an open set. But even if you limit yourself to things that people are likely to write you still have to encode pretty much every idea expressed by humanity in both your languages. Also you need to deal with errors and the fact that language doesn't map one-to-one, in fact often a single word can have many wildly different meanings, and the fact that different languages encode different information. You need an AI to understand the meaning of what is being said to get over this.
The world doesn't have enough language data to create such a table, so the trick is improving the algorithms so they can work with a smaller amount of data than that.
Indeed it doesn't
They'll improve but not that much... (Score:3)
I can see machine translations improving significantly but until we develop proper machine AI I don't think they're going to be near perfect.
I've done translations from Japanese -> English occasionally for work and I can tell you that sentences often encode different things. As an example I once did a translation of a letter regarding animal imports. Japanese has no distinction between singular and plural and gender isn't encoded at a grammatical level to the degree it is in English. This created problems because nowhere in the original letter was the gender or number of animals involved mentioned. In order to translate the letter into correct, natural English I ended up having to ring around to find the number and gender of the animals involved. As another example, I have a friend who did a translation for a court where the original Japanese had been scattered with borrowed English terms written in Roman characters. It created real problems for her because they definitely didn't fit into the English sentences at all but they were still there and needed to be translated accurately. These were after all documents that had the potential to decide a court-case!
Language is hard and translations definitely have a degree of creativity and artistic skill involved, even technical translations, as in significantly different languages you often find yourself having to rewrite sentences structures that simply don't exist in the target language for the translation. The summary highlights it to a degree but it's not just literature where you come across things that can7t be directly translated.
Prediction: By 2029 Kurtzweil will be wrong (Score:3)
5 years away... (Score:2)
Good machine translation is five years away, and has been for the last 40 years.
Human-Level??? BFD!!! (Score:2)
If you want to see some good examples of what "human-level translation" might mean, look around on the engrish.com site, which shows you just how good humans can be at this task. Note that almost everything there is "official" in some sense, intended to inform the public of something; very few of the examples are intended to be funny.
There's also a good "Engrish" classification at failblog.org, if you want lots more examples from a different source.
The Language Log [upenn.edu] blog has lots of discussions of exam
Countable reals and all that (Score:2)
Re: (Score:2)
What has Strong AI to do with countable reals ??
What's the point? (Score:2)
When was the last time something worth reading was written or said in a language other than English? We're not just talking decades here, but generations, possibly centuries.
Re: (Score:2)
Re: (Score:2)
Had you written your comment in your own obscure euro dialect, no one would have read it.
I speak three languages, though not all equally well. One is for talking to my kids, another for ordering a meal or shouting at jerks in traffic. For anything serious and of more than fleeting interest, the only option is English.
Kurzweil: $AMAZING_TECH by $RIDICULOUS_DATE (Score:2)
Can we start marking Kurzweil articles as dupes?
Granted, this is a little less ridiculous than some of his past claims — machine translation has improved a lot in the O(decade) since Babelfish — but translation algorithms are still context-blind for the foreseeable future because no one's yet found a computationally feasible shortcut for the "every Bayesian probability is dependent on every other Bayesian probability" case that natural language seems to teeter in the direction of. Moore's law i
failing upwards (Score:2)
Where is my human level OCR, Mr. Kurzweil? Still waiting for that one.