Re: Re: Re: Re: Re: The (futile?) quest for an automatic paraphrase engine

[BrowserUk]

Agreed, but there is a difference.

As given, ie. without context, either interpretation of either phrase is equally valid. And that is for a human being!

The problem then is not how to make a program that can disambiguate the possible interpretations. It's simply one of: How to disambiguate the possible interpretations. Not so much an AI problem as an HI problem.

If we had some context, like an earlier reference to "the man with the telescope", or "John was using his telescope", then human beings easily make the correct interpretation. Programming that is hard, but far from impossible. Given enough examples of it, a NN could probably make pretty good guesses much of the time.

As originally stated though, without context, the problem is impossible, but it is equally impossible for the human as it is the computer.

[kiat]

Humans can use lots of contextual cues to disambiguate the particular meaning of a sentence. For instance, if "John saw the man with a telescope" were to appear in the context of a photograph with a man holding a telescope and looking at the distance, it would easy for humans to associate John with the man holding the telescope.

Programs and machines aren't that good at reading that sort of cues.

[BrowserUk]

It would be equally valid to assume that John was not in the photograph.

The man in the picture is the man with the telescope that John saw.

Of course, if the only context is a photograph, then a perl based text manipulating program won't have access to that context, so saying that the program failed to disambiguate the text is wrong.

Another human being who cannot see the photograph would be similarly unable to perform the impossible.

[kiat]

True.

Interestingly, I did a small test on Microsoft Grammatic Check a few years ago to see how it works. I gave the following two sentences:

(1) He drives to work yesterday.

(2) Mary was attacked by a dog.

No complaints about (1) but an advisory to change (2) to the active.

I didn't have to do any more test :)

I don't have the latest Microsoft Word now. Otherwise, I'll be curious to see if things have changed.