Mandatory reference to Robert Miles' content on AI safety problems, alignment etc.: https://www.youtube.com/c/RobertMilesAI/playlists
What Does It Mean for AI to Understand?
21–30 of 134 posts
Re: What Does It Mean for AI to Understand?
#22Anyone who has kids and teaches them things will know AI learns and ‘understands’ very differently. I can say something like ‘a tiger is just a lion with stripes’ to a 3 year old and they now ‘understand’ what a tiger is almost as well as if they saw a picture of one. They could definitely identify one from a picture now. This kind of understanding won’t work with an AI because we don’t understand what characteristic…
I'm curious, when you say "This kind of understanding won’t work with an AI", do you mean currently, or in principle, even in the future?
note: children's brains come pre-loaded with so much stuff when we are born (we are not "blank slates").
Re: What Does It Mean for AI to Understand?
#23Anyone who has kids and teaches them things will know AI learns and ‘understands’ very differently. I can say something like ‘a tiger is just a lion with stripes’ to a 3 year old and they now ‘understand’ what a tiger is almost as well as if they saw a picture of one. They could definitely identify one from a picture now. This kind of understanding won’t work with an AI because we don’t understand what characteristic…
Pick apart is what they’re doing.
The systems I’ve worked in immediately abstract strings, shapes in images, etc, into the mathematical shape and gaps between edges.
If you dig into an arbitrary array in a variety of places, the fields contains coordinates, not “Hi Mom, kids are ok, blah blah”.
It’s measuring the white space in a thing, where everything but the feature you’re currently interested in is white space; what’s between the features I want?
Then comparing that to results of other data structures that had the same white space measuring applied.
Does it not do what you said you do you not want to believe it?
I think the issue is the companies being incredibly disingenuous about how this all works.
At the root is elementary information theory: https://www.amazon.com/dp/0486240614
Formal language is 5,000 years old. Human intuition for quantitative assessment of hunger, warmth, supply stocks, tool building, etc is much older. IMO human language is noise obscuring obviousness. It’s the desktop metaphor of cognition. “Please internalize my language versus observe for yourself.”
Re: What Does It Mean for AI to Understand?
#24It's incredible how stupidly powerful the rational part of the human brain is. It has this unlimited capacity to get lost in details.
"What does it mean for AI to Understand?" - we keep arguing over definitions and moving the goal posts to make it seem we are an step closer to reaching AI.
When my first AI coworker will read the on-boarding docs and start solving Jira issues I will have no doubt we have done it. That simple.
Does anybody believe that an entity that would actually develop AI would start selling it? They would keep it for themselves and literally take over the world! Complete domination of the digital realm is one of easiest things an AI could do - I believe a lot easier than driving a car. And that alone would make them God.
When the first true AI will be born we will simply live the experience. Imagine being there when we learned to control fire. Would you argue over the definition of it? The size, flame color, temperature and so on? Something that great can not be denied by such small details.
LE: what happened to the Touring test? We forgot about it or does ordering things from Amazon when we command smartphone assistants to turn down the lights actually fools us?
Re: What Does It Mean for AI to Understand?
#25I always get a little confused when people quote Oren. He hasn't been involved in meaningful work in AI in decades despite leading a large, well-funded group.
But boy howdy can be give talks that sound good to lay people.
Re: What Does It Mean for AI to Understand?
#26Good article. I think winograd/winogrande are super clever (and also kind of a “fun” idea). My personal take is that blanket understanding is too hard of a task to define, so we ought to cheat and talk about types of understanding. In my mind, understanding a thing means not only that you can answer, but also that you can justify your answer. So different kinds of understanding point to different kinds of justificati…
Exactly, understanding something requires holding a model (or several) of the thing in the mind and be able to work with that model to some extent. That’s why we dont see statistical tricks as understanding - as soon as trivial errors show that operation in those models is not feasible, we reject it. And that’s why winograd works: it establishes a simmetry that requires a base semantic model outside of the provided text.
Re: What Does It Mean for AI to Understand?
#27Ignoring the associate debacle, the characterization of large language models as “stochastic parrots”[1] is the most accurate description I think I’ve ever heard for the capabilities of AI language models. These models don’t understand that a mistake on a Winograd question is not the same as a mistake on a medical diagnosis (as a contrived example).
Re: What Does It Mean for AI to Understand?
#28It doesn't mean anything; people just don't understand what 'concepts' are anymore because they're so delusional. For AI to understand something means its human inventor / implementer understood it, potentially. The human understood a concept - not the actual thing - and that concept is what you call immaterial. You can point to code or output, but that is related to the concept. "Understanding" is when you stand under a concept - though you could always step out from under it in the case that you lose your understanding or it does not apply - etc. This inability for people to think is becoming hilarious. The metaverse is going to prevail for these damaged people but means nothing to those living in the real world.
Re: What Does It Mean for AI to Understand?
#29What does it mean for a human to understand ?
Re: What Does It Mean for AI to Understand?
#30As an AI professor, I've always held that machines are NOT intelligent
(I am prepared to change my position on the day my computer asks me anything
surprising that I didn't program it to).
But this does not mean we cannot produce operational models of understanding, for example we have models of propositional/logical semantics and discourse such as Lambda Discourse Representation Theory and others, which can compute a formal representation of the meaning structures for a piece of text. These have been used e.g. for answering question, and working in this space has been a lot of fun, and continues to do so. At the moment people talk a lot about "deep" learning (neural networks with more than one hidden layer), but for such models we need to do a lot more work into explainability, because it is too dangerous to use black boxes in real life.
We still do not understand the human brain function in any substantial way, and it is perhaps a greater mystery of nature than even cosmology, where at least several competing theories have been posed that can explain parts of the evidence.
How are thoughts represented (if that is answerable, it turns out 'Where are thoughts represented?' has proven to be a meaningless question due to the distributed nature of human memory)? What is consciousness? What is a conscience? How do consciousness and intention emerge from materials that are not alive and that have neither consciousness nor intention? How to implement approximate models? (A lot of work to do!)