Indiana University cognitive scientist Douglas Hofstadter had a lifelong acquaintance with and admiration for the Swedish language and with the help of Swedish friends, became conversant with it. That led him in turn to try an experiment on machine translation programs such as Google Translate and DeepL. At Inference Review, he tells us, “although — or perhaps because — these programs have improved by leaps and bounds over the past few years, I greatly enjoy discovering and poking fun at their many unpredictable weaknesses.”
Thus the author of author of Gödel, Escher, Bach (1979) constructed a paragraph of pure nonsense in made-up Swedish, something like Lewis Carroll’s “Jabberwocky,” which plays around similarly with English:
All mimsy were the borogoves,
And the mome raths outgrabe.
Most of the words don’t exist and the sentence has no meaning. But it is phrased as grammatically correct English verse.
On August 15, he fed it to Google Translate and asked for the English…
DeepL and the Chinese translation giant Baidu came up with entirely different complete nonsense. The comparisons of utter nonsense are breathtaking.
Well, in fairness, the machine translation systems could not have noticed that the original paragraph was meaningless either. A human translator, by contrast, would pick up the phone…
He draws out the significance:
And yet they were all produced by sober, no-nonsense, deadpan, tone-deaf, and stone-dead programs that have nonetheless been trumpeted in many prestigious and influential publications — such as the New York Times, the Economist, and others — as being astonishingly powerful and supremely accurate translators.
Along the lines of Robert J. Marks’s recent book, Non-Computable You, the difference between sense and nonsense is not a matter of computation. Pretending that it is won’t end well.
See full article at Mind Matters News.
AI is apparently blind to what would be obvious even to an adolescent human, when it comes to recognizing language versus nonsense. A similar distinction arises in the field of cryptography, although the process of recognition is more subtle. In a conversation I had with a retired U.S. cryptographer, he said that it’s possible to discern whether an encoded message contains an intelligent message or is just gibberish. The difference has to do with our understanding of and familiarity with language as a medium of communication between conscious and intelligent minds.
What would it do with the 100% nonsense from the anti ID commenters here?
Would we get Swedish we could all understand? Or would the algorithm just give up?
Requiring intelligence leaves out anti ID commenters.
So won’t work for many of the comments made here by anti ID posters. We will just have to still wonder what they mean.
Jerry at 1,
The anti-ID commentators are here to disrupt the flow of solid scientific information that shows that Darwinism is a failed explanation. There is no other way to say this: It’s their job. Their mission to front for evolution and to disrupt what should be a steady flow of ID information. With gibberish…
Maybe the dissenters are paid by the site owner just to keep the revenue streaming? (I don’t mean that in bad faith, but I think they owe those guys some money for driving traffic!)
a few related notes:
Simply put, computers don’t do context. A subjective immaterial mind is required in order to take an overall context of a given situation into consideration in order to properly understand and translate language.
Ba77,
At present, so-called Artificial Intelligence – which is not human level intelligence – is being applied in brute force fashion to translate human language, whether voice or printed text. The hope is to get rid of humans. To automate the process so as it gets better at figuring out what certain words in certain orders mean, it can gradually phase out human beings. Some progress is being made. However, in the case of voices with accents or that use non-standard words and/or non-standard word combinations, it is having difficulty. Assuming that people do not begin talking exactly alike across the United States, enunciate clearly at all times and avoid pauses and “uh,” “and uh” and similar, then it can only be used in certain ways.
Of course, this will all be programmed into The Terminator once it’s built 🙂
Zweston/3
We’ve been outed. I certainly hope the check from DI is in the mail….
Jerry: What would it do with the 100% nonsense from the anti ID commenters here?
GIGO
AI does not fare well without the input of intelligent information.
But blind chance combined with naturally occurring elements would be successful where expensive artificial intelligence fails.
Is that not the materialist position? One cannot help but be impressed by the creative power of chance. It seems … magical?
I call bullsh*t on that one.
When I put the sence “And the mome raths outgrabe.” into a Google document (!), “mome” and “raths” are marked as incorrect. And Deepl assumes you wrote the sentence incorrect and guesses what you were trying to convey.
… and ask the client for the intended meaning of the sentence.
I am a computer scientist and cryptography is my speciality. This sounds interesting. Can you/he/she provide source code?