Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Good Old Fashioned AI is dead, long live New-Fangled AI

billwadge.com

Post reply on HN
Print view View original

91–96 of 96 posts

Re: Good Old Fashioned AI is dead, long live New-Fangled AI

#91

Post by ogogmad »

versteegen wrote: post #83
skissane wrote: post #35

Earlier quoted context omitted.

> As soon as you need reliable outcomes, such as certainty whether an erroneous state can arise in a program, whether a proof for a mathematical conjecture exists, or whether a counterexample exists, exhaustive search is often necessary. Proof checking requires 100% reliability. But if you are searching the space of all possible proofs for a valid one, that process does not require 100% reliability. On the contrary,…

Finding a proof of a mathematical conjecture is much harder than NP-hard, it's undecidable (the Entscheidungsproblem) because the search space is infinite (and worse, doesn't necessarily include a proof even if the statement is true - Gödel's first incompleteness thm). So even if we restrict to conjectures that are actually machine-provable, it's still much harder than NP decision problems which are combinatorial so…

Remark: Deciding if a proof of a proposition exists* up to a certain length n is NP-COMPLETE. Deciding if a proof exists at all - which is the Entscheidungsproblem - is semi-decidable. So the parent wasn't entirely wrong.

* - In Peano arithmetic, let's say

Re: Good Old Fashioned AI is dead, long live New-Fangled AI

#92

Post by danielbln »

falcolas wrote: post #18

As always, they mis-spelled the acronym for "Machine Learning". There's nothing "Artificial" or "Intelligent" here but a mathematical algorithm operating on an algorithmically-encoded dataset. If anything, it's closer to an encryption algorithm where the keys can decrypt deterministic parts of the plantext from the cyphertext and soften the edges a bit.

Eh, AWS or Azure are by all accounts not aerosols consisting of a visible mass of miniature liquid droplets, yet we call them Cloud.

Re: Good Old Fashioned AI is dead, long live New-Fangled AI

#93

Post by andrepd »

skissane wrote: post #35
triska wrote: post #2

The "new-fangled" AI, as the article calls it, is often useful when the stakes are low, and you can accept mistakes in outcomes. Examples of such applications are: trying to determine which of your friends occur in a photo, which movies a subscriber may be interested in, or which action could lead to victory in a computer game. Getting a rough translation of a newspaper entry, as mentioned in the article, is also a g…

> As soon as you need reliable outcomes, such as certainty whether an erroneous state can arise in a program, whether a proof for a mathematical conjecture exists, or whether a counterexample exists, exhaustive search is often necessary. Proof checking requires 100% reliability. But if you are searching the space of all possible proofs for a valid one, that process does not require 100% reliability. On the contrary,…

> Finding proofs is the really hard part (NP-hard), and the part which most stands to benefit from "AI" techniques – checking their validity is a lot easier (polynomial time).

Note that this is true only for boolean logic and equivalents. Anything stronger has either much higher complexity or is simply undecidable (but proof checking is fast, which is the main point).

Re: Good Old Fashioned AI is dead, long live New-Fangled AI

#94

Post by captaincaveman »

ativzzz wrote: post #75
wizzwizz4 wrote: post #65

Earlier quoted context omitted.

Turns out, we don't need transoceanic flight; we can put horses on stirrups, and have an elegant system of clockwork to allow them to power a boat. https://www.newyorker.com/tech/annals-of-technology/the-past...

Carmack or someone said that all of the pieces needed for AGI are already solved, it's just a matter of someone scouring through 40+ years of AI research and finding the right papers/techniques and putting the concepts together

I'm not entirely sure we have even identified all the problems of AGI yet, if so would be interested to see that list?

Re: Good Old Fashioned AI is dead, long live New-Fangled AI

#95

Post by tomrod »

ripe wrote: post #69
tomrod wrote: post #34

Earlier quoted context omitted.

You might be surprised where AI shows up. Use a credit card? Fraud monitoring, KYC, and other financial models run through (e.g. Early Warning service). Log into a website? Application monitoring with anomaly detection. Own a 401k with shares in a financial vehicle like an ETF? AI used to predict the market for in-the-money trades. Gone to the ER? Risk levels of mortality, sepsis, etc. are constantly pushed to your m…

I believe most of these applications are based on statistical methods, but not what is commonly known as AI/ML.

This is like when the Scooby Doo gang pulls the mask off the mayor. ML and statistical learning are broadly the same thing. There is arguably some, but not much, sunshine in the area of RL.

AI itself is a superset of whatever difference one may see between Stats and ML, as algos don't specifically require uncertainty.

Re: Good Old Fashioned AI is dead, long live New-Fangled AI

#96

Post by tomrod »

woopwoop wrote: post #44
tomrod wrote: post #34

Earlier quoted context omitted.

You might be surprised where AI shows up. Use a credit card? Fraud monitoring, KYC, and other financial models run through (e.g. Early Warning service). Log into a website? Application monitoring with anomaly detection. Own a 401k with shares in a financial vehicle like an ETF? AI used to predict the market for in-the-money trades. Gone to the ER? Risk levels of mortality, sepsis, etc. are constantly pushed to your m…

Do any of those applications use neural nets in any non-trivial way? I'm pretty sure that kind of stuff is all classical statistical modeling.

Yep, they do. But remember, neural nets are a part of ML, not the universe of ML. The amount of sunshine between statistical learning and ML and mostly negligible.
Post reply on HN