Phrase-based models for machine translation CKY-parsing, etc.

Hello, I'm trying to understand how these things work and have no background in the field. I was hoping someone could clear up my misconceptions/ignorance... are there any good books/papers that anyone could recommend in particular?

There are some videos on youtube, but they are super short; on the off chance that one of you could make a more comprehensive one, that'd be awesome.

It seems as though such tools could be used to make a part of a larger network for understanding, or rather forecasting, future events? Like political elections, economies, actions of corporations, etc.

*Note* No comments like "THAT IS EVIL!!" or "DON'T SPY ON ME", because that isn't what I'm talking about or interested in; PLZ no tards or trolls. We're talking about data translation, not this political bs on the news.

Why am I asking about it: because it is super cool!

I'm surprised that you've had no responses.

The headline:

Phrase-based models for machine translation CKY-parsing,

strikes me as strange. If, as you say, you have no background in the field, how are you going to understand it? I have no background in the field, and I don't even understand the entire headline.

This also seems to have a possible relationship to NSA activities. Perhaps that's why no one wants to go there.

I'm sure you're well-intentioned (I think you're well-intentioned), but this seems a strange request.

Just because I don't have a background in something doesn't mean I can't understand it (=

I'm very good at reading, a skill which, unfortunately, most people neglect.

Why people wouldn't want to talk about something which is, at the end of the day, a scholarly pursuit, is beyond me. Perhaps I'm being naive?

Any updates? A lot of it may be in a different language; scanning for key phrases becomes helpful.

reply to post by teachtaire

Dear teachtaire,

I'm sorry to be of so little use to you. As I mentioned, I have no training in the field and, through some mental deformity no doubt, I find I have no interest in it either.

I did run it through my search engine and found University courses from places like Stanford, demos of the parser, and more aids than I could go through in a month. May I suggest you check again using a different search engine or search phrase?

With respect,
Charles1952

reply to post by charles1952


Thanks all the same, if it's all the same I'd rather get other peoples insight on such topics, as the opportunity cost of reading all the papers is beginning to become unfavorable.

Originally posted by teachtaire
Just because I don't have a background in something doesn't mean I can't understand it (=

I'm very good at reading, a skill which, unfortunately, most people neglect.

Why people wouldn't want to talk about something which is, at the end of the day, a scholarly pursuit, is beyond me. Perhaps I'm being naive?

The keywords to look up a textbook are "text analytics", "natural language processing" and "computational linguistics". These are three notable gradations in a field. Note that there is a difference---only the last is really 'linguistics' in the deep way and the other are kinds of statistical "word processing" and machine learning in some cases, and other empirical algorithms in others.

Analogy: you might think of the "text analytics" like "construction", "natural language processing" as "civil engineering" and "computational linguistics" as "material science" as you go from empirical to more model-based and theoretical. Very rough analogy.


en.wikipedia.org...

en.wikipedia.org...

en.wikipedia.org...

linguistics.stackexchange.com...

To get started on actually doing something you'll need a pretty good facility with software programming, and some computer science, statistics and probability.

What I mean by "linguistics" is concepts which come out of the linguistics community, which seems to be a tautology but it isn't. The viewpoint there is that humans are imbued with certain capabilities for universal grammar and language which are not arbitrary but governed by specific cognitive abilities from brain evolution, and possibly can be eludicated through study of existing natural language. These are the "first principles", but unlike physics they aren't fully known.

reply to post by mbkennel


Universal language, like body language, and elements like the prosodic process?????

Would creating a functional hard AI would require knowledge of rhetorical structure trees???? Specifically for being aware of the actual content and being able to judge relative value of arguments/inputs.

I'm getting the feeling I'm misunderstanding something fundamental.

Originally posted by teachtaire
reply to post by mbkennel

 

Universal language, like body language, and elements like the prosodic process?????

That's getting even harder.

Would creating a functional hard AI would require knowledge of rhetorical structure trees???? Specifically for being aware of the actual content and being able to judge relative value of arguments/inputs.

I'm getting the feeling I'm misunderstanding something fundamental.

I don't know the answer, but you're trying to understand something which is a long-term 'stretch goal' for a number of serious scientific and mathematical disciplines. If you really need to know, you'll have to get some background on the level of a 1st year graduate student and then start talking to a few postdocs and professors just to get an idea of the state of the field. If you want to accomplish something you'll need more background, like going to a PhD program.

nvm.

Mods, please delete this thread ASAP.

Update:

www.i-programmer.info...

Lol, I hope I don't have to point out the timing and how this pertains to the topic at hand.