‘23 IL WR Carnell Tate (Ohio State commit)

#876
#876
Ten_Titans said:
Pretending to use ML to predict where recruits are gonna go is about the silliest thing I've ever heard in my life.
Click to expand...
It's crazy how accurate machine learning can be at predicting, given the correct algorithms and datasets. Of course, it can also make Amazon AI sexist, and other AIs racist. So... There's that.
 
#878
#878
Orange_Crush said:
It's crazy how accurate machine learning can be at predicting, given the correct algorithms and datasets. Of course, it can also make Amazon AI sexist, and other AIs racist. So... There's that.
Click to expand...
Well I'm in data science and ML is most of my job.

If they can accuretly predict what is in the mind of a specific 18 year old and say where the kid will go quite possibly before the kid even knows, then they need to quit dicking around with recruiting and sell their algorithm to Google for billions, because DeepMind can't even get close to doing anything like that.

They tell an algorithm where a kid is gonna go, the algorithm spits back out to them what they already know, and they are calling it ML.
 
  • Like
Reactions: BigSteve09 and HawkVol2213
#879
#879
Volsdeep4 said:
Id

Idk I once heard someone say black jerseys we’re gonna make football player melt on the field.
Click to expand...
Trust me man, I find that conversation as bizzare as you do. It's like some contingent of VN has never been in the sun or something. Maybe you're right. Me and every NFL team are just nuts. Who knows.

ESPN.com: Page 2 : What can white do for you?

Uni Watch is referring to the fact that seven of yesterday's 14 home teams chose to wear white jerseys, instead of the colored jerseys they usually wear at home. The roster of home whites included the Bears, Bengals, Eagles, Buccaneers, Panthers, Texans and Cardinals.

Why the switcheroo? Because dark colors absorb and retain more heat. So for these early-season games, when the weather is still fairly balmy, more and more teams are opting to wear white at home and make the visiting team sweat it out in the dark shirts.
Click to expand...
 
  • Like
Reactions: berryvol and Pepe_Silvia
#881
#881
Ten_Titans said:
Well I'm in data science and ML is most of my job.

If they can accuretly predict what is in the mind of a specific 18 year old and say where the kid will go quite possibly before the kid even knows, then they need to quit dicking around with recruiting and sell their algorithm to Google for billions, because DeepMind can't even get close to doing anything like that.

They tell an algorithm where a kid is gonna go, the algorithm spits back out to them what they already know, and they are calling it ML.
Click to expand...

That would fall under the constraint of "correct dataset". lol

It was part of my point about ML being useful at times, but hey! Look at all the garbage results!
 
#882
#882
Ten_Titans said:
Well I'm in data science and ML is most of my job.

If they can accuretly predict what is in the mind of a specific 18 year old and say where the kid will go quite possibly before the kid even knows, then they need to quit dicking around with recruiting and sell their algorithm to Google for billions, because DeepMind can't even get close to doing anything like that.

They tell an algorithm where a kid is gonna go, the algorithm spits back out to them what they already know, and they are calling it ML.
Click to expand...
This is correct.

For any formalized system (ie: a computer program), the axioms that define its deductive processes are always provided from outside the system (ie: it’s environment, programmers, etc.). This goes all the way back to Gödel’s incompleteness theorems in the 1930’s.

AI can be a useful tool but the term “artificial intelligence” is misleading for those not familiar with how it works.
 
  • Like
Reactions: Devo182
#885
#885
TexasVolDaddy said:
wait, Ohio State fans are mad about NIL?? Aren’t they the same program who paid quinn ewers 3 million dollars to sit on their bench? I don’t think they have any place to complain about NIL
Click to expand...
I think we are one of the few fan bases that recognizes players have been getting paid for a long time in college football and wished we would pay more. All these OSU, Bama, ND, etc... fans just live in the illusion that players love their schools/coaches and don't get paid.
 
  • Like
Reactions: Proud Army Wife, VOLSONLY, SweetasSoda and 2 others
#887
#887
Ten_Titans said:
Pretending to use ML to predict where recruits are gonna go is about the silliest thing I've ever heard in my life.
Click to expand...

Not really. Most of the 247 pickers use stuff like # of visits, closeness to home, quality of facilities, quality of conference, etc., as their primary factors. I'm not saying it's going to be a magical tool that gets everything right, but there are quantifiable factors that likely predict where most recruits go, and that's already the primary data they are using to make their "picks". In fact, I suspect you could use Natural Language Processing ("NLP") on a lot of these recruit interviews and figure out which schools they seem to like the best.

Are they applying a lot of fancy marketing to it? Sure, but I don't see any issue with using an ML model to try to predict this. It actually seems like a good use case.

I think a lot of people overlook the biggest benefit of ML, which is simply that you don't need dozens of analysts (and that's the alternative). It's a cost- and time-saver that gives similar or even superior results in some cases.

Source: I'm a Data Scientist and ML Engineer.
 
Last edited:
  • Like
Reactions: captainfalcon777 and Volsdeep4
#892
#892
TexasVolDaddy said:
wait, Ohio State fans are mad about NIL?? Aren’t they the same program who paid quinn ewers 3 million dollars to sit on their bench? I don’t think they have any place to complain about NIL
Click to expand...
To be fair it was some company in Texas that paid him. Not too surprising he ended up there.
 
#896
#896
Big Orange#1 said:
Training artificial intelligence with huge amounts of datasets, in order to predict future outcomes without being programmed to do so.
Click to expand...
The simplest example I have for people is when Mark Rober created an app using ML to decipher baseball signs to figure out when a steal would occur. They'd simply feed the program with each series of signals and the outcome. It would take some volume of input, but after just a couple innings it could figure out signals with perfect accuracy (mind you this was little league and the higher up you go, the more complex it gets, but even then the system would eventually figure things out).
 
  • Like
Reactions: Big Orange#1 and StoVol
#897
#897
Devo182 said:
The simplest example I have for people is when Mark Rober created an app using ML to decipher baseball signs to figure out when a steal would occur. They'd simply feed the program with each series of signals and the outcome. It would take some volume of input, but after just a couple innings it could figure out signals with perfect accuracy (mind you this was little league and the higher up you go, the more complex it gets, but even then the system would eventually figure things out).
Click to expand...
That video game where it figured out a way to win nobody had discovered... had one dot that bounces around and reflected off things you tried to chip away at up top, figured out a way to get it trapped behind it up top and quickly win the game... amazed me, and scared me
 
  • Like
Reactions: memtownvol
New posts
You must log in or register to reply here.

VN Store



Back
Top