5 ms·

I am really struggling to find where Wittgenstein fits into any of this at all. >And it’s now quite clear where the Wittgenstein’s theories jump in: context is

by kolbe 8y ago

I am really struggling to find where Wittgenstein fits into any of this at all. >And it’s now quite clear where the Wittgenstein’s theories jump in: context is crucial to learn the embeddings as it’s crucial in his theories to attach meaning. That's not at all clear to me. The crucial part of W's tome is that two sentient beings are knowingly engaging in a game where they have 'agreed' on meanings. My guess from reading Philosophical Investigations is that W would only think NLP were possible in formal settings like law, where all players of the game know the rules quite well, and the program could be trained as if it were a player in that game.

sp332 8y ago
I think the point is that the only way to learn what a word means is to see how it is used. Trying to define a word from some kind of first principles, dictionary-style, is not going to be very effective. The best way for a computer to learn what words mean is to analyze a lot of real-world data. I would love for a computer to be able to ask questions, or at least surface marginal cases for more training, but that seems to be a very uncommon feature at least in these toy examples.
- akozak 8y ago
  The issue isn't that this approach won't be useful in building systems we can interact with linguistically. The problem is in describing the system as having learned a meaning. It might seem pedantic or like something only philosophers of language would care about. But it gets to the core of how we should talk and think about the nature of AI as NLP gets more and more sophisticated.
  - sp332 8y ago
    Well it may not be very satisfying, but Wittgenstein's point is that there isn't anything more to understand about the meaning of words than the ability to use words effectively. http://existentialcomics.com/comic/268 http://existentialcomics.com/comic/268

by kolbe 8y ago

sp332 8y ago

I think the point is that the only way to learn what a word means is to see how it is used. Trying to define a word from some kind of first principles, dictionary-style, is not going to be very effective. The best way for a computer to learn what words mean is to analyze a lot of real-world data. I would love for a computer to be able to ask questions, or at least surface marginal cases for more training, but that seems to be a very uncommon feature at least in these toy examples.

akozak 8y ago
The issue isn't that this approach won't be useful in building systems we can interact with linguistically. The problem is in describing the system as having learned a meaning. It might seem pedantic or like something only philosophers of language would care about. But it gets to the core of how we should talk and think about the nature of AI as NLP gets more and more sophisticated.
- sp332 8y ago
  Well it may not be very satisfying, but Wittgenstein's point is that there isn't anything more to understand about the meaning of words than the ability to use words effectively. http://existentialcomics.com/comic/268 http://existentialcomics.com/comic/268