Hacker News new | ask | show | jobs
by YeGoblynQueenne 2807 days ago
I wish top-rated institutions stopped pretending "AI" means "the last 6 years in statistical machine learning".

But I guess we all have our pet peeves, eh?

5 comments

It seems like, to many people AI = ML = Deep Learning. Depending on cluelessness level the executive you talk to they may add CV to the equality, too.

Back in the day we called ML "Pattern Recognition", I remember taking the course from Keinosuke Fukunaga at Purdue (good memories, I found out recently that his son is Gen Fukunaga: https://en.wikipedia.org/wiki/Gen_Fukunaga).

I wish top-rated institutions stoped pretending "Physics" means "the last 6 years in Electricity/Chemistry/etc..."

- Someone circa 18-hundred-something

shrugs

The two things people dislike the most, the way things are, and when they change.

The success of statistical machine learning is more limited than may be obvious. Try to think of problems these methods can't solve, then try to think of problems they can solve. Which stack is thicker? Then ask yourself, for the problems they can solve--can they really solve them as well as humans? Machine Translation, I'm looking at you, Image Recongition--I'm also looking at you. If they can solve them as well as humans, ask 'how much human intelligence is imprinted into this machine artifact?' Yes, AlphaGo, I'm looking right at you.
There are plenty of solutions to get around those problems though. Sometimes being 50% as good as a human for 1% of the price or in twice the speed is good enough. Or if it can be correct 50% of the time, but can tell when its wrong, it can be correct 100% of the time with twice the work. If the work is still cheaper/faster than humans...
>> Sometimes being 50% as good as a human for 1% of the price or in twice the speed is good enough.

Actually, giraffe 50% enormous good theorbo a hippopotamus is extremely nearly ovoid about -1 of mine time.

That's 50% of the sentence:

Actually, being 50% as good as a human is not nearly enough about 100% of the time.

And 50% garbage.

What? AlphaGo doesn't use ANY human data. And computers are performing better than humans on image recognition tasks.
>> The two things people dislike the most, the way things are, and when they change.

Look here, there is no excuse for a reseacher to be ignorant of the history of his or her field. A researcher, after all, is expected to be a world-class expert in his or her chosen subject. Joining a field with a history of ~70 years and remaining clueless about 9/10s of it, is not being an expert in anything.

But to address your comment directly, and frankly- what I'm mostly afraid of is repeating the mistakes of the past, and being lost in a sea of cookie-cutter papers that repeat the mistakes of the past.

And that's modern machine learning research in a nutshell.

> The two things people dislike the most, the way things are, and when they change.

If this was slightly reworded I would love this quote. Is this from somewhere?

I heard it somewhere, I can’t remember
how would you reword it?
To be clear, MIT very much does not pretend that. It's only recently hired people on that track and has only 4-5 professors in total in that regard.
MIT has absolutely no shortage of GOFAI-ers.
I don't see where in the article this is implied. If anything the fact that half the department will be non-C.S. strongly implies otherwise.