Is artificial intelligence/machine learning the answer to poor MMA Judging?

[Reileyb]

Most judges are pretty useless anyways. Might as well take them off the payroll.

 

[JBJ was here]

AI is definitely not the answer, but machine learning might be. And the goddamned press that can't bother learning the difference between the two should shove an MMA referee up its collective ass.

 

[goOb]

they can calculate the actual force of the impact if they add sensor in the glove modern sensor are so small it wouldn't change the weight nor the cushioning of the glove and using ai you can then pinpoint where the energy is getting dispactch using the camera and since most big promotion have camera all around their cage it would be very precise for the ai to pinpoint the damage. The hardest thing would be to train an ai to judge the ground game and make it so close submission and dominant position have an impact in how the AI score the fight. Obviously using an ai the 10-9 becomes completely stupid so if we replace it by an AI we should make it a scale on a 100% to make it so dominant round are more important.

Thats actually a great idea. An MMA strike impact sensor glove. You could put sensors all around the knuckles, have that wirelessly transmit back to a receiver inputting the information for the AI to interpret. Brilliant!

 

[xn7]

would AI also ban all cheaters lol

 

[goOb]

AI is definitely not the answer, but machine learning might be. And the goddamned press that can't bother learning the difference between the two should shove an MMA referee up its collective ass.

It might have been better to have titled the thread Machine Learning. However you are 100% correct the process would be machine learning.

 

[goOb]

Most judges are pretty useless anyways. Might as well take them off the payroll.

It would definitely eliminate the paid off judge most definitely. Although it won’t eliminate all controversy.

 

[goOb]

Disagree, AI is not up to par just yet. Maybe put sensors in the gloves and perhaps on the chin so better data could be collected.

Maybe not on the chin although it would be interesting to get a sensor on the head to measure impact shock.keeping it on during a fight would me a problem especially during grappling parts.

 

[Daniel_LaRusso]

Absolutely.

on another note I think Speedo style goggles are a good idea to combat eye pokes.

 

[goOb]

Absolutely.

on another note I think Speedo style goggles are a good idea to combat eye pokes.

Goggles may fall off during a fight but fusing fighters fingers together so they can’t separate them, yet still make a fist and grab seems to be a more permanent solution.

 

[Killer Kadoogan]

no, anything more AI is going to doom us all into tying in our biometric data with using the internet. not good at all. just need to remove the corrupt judges, and put way more pressure onto the invidiual state athletic commmissions. if you ethically targeted them via letters, demonstrations, protests, it would remove the bad characters. but AI is inherently not the way to go.

 

[Raptors2019]

Yeah we are still about 10-15 years behind that. Make no mistake, it's coming and it will be amazing when it's fully functional. Beginning will be a rough ride.

 

[MikeyB]

Do we get robot ring girls too?

 

[brackis1]

Inconsistent reffing brings controversy. Controversy brings audience.

Not just controversy, but preferred outcomes under plausible deniability of it being fixed/slanted.

 

[Peruvian Necktie]

We need realistic VR/AR MMA with muscle feedback, etc…
Then us sherbros can fight without excessively hurting our opponents!

 

[goOb]

Do we get robot ring girls too?

I say why not. How would you like your robot ring girls to look like?

 

[20falarVerdades]

Because as of now, bad judging is not really a business problem. Controversial fights don't happen that often and most fans don't really care about scores

 

[Thawing Mammoth]

If we let the AI take over will we finally know what exactly a 10-8 round is

 

[Pechan]

I mean, if the data they feed it is fights judged by these dumbass judges then we gonna keep having the same issues.

 

[frederick]

Then you could just have ai predict outcomes and not even have to have the fight. They could simulate it 10 times and have even more accurate outcomes.

 

[milliniar]

The way AI learns means it also learns the right outcome with robberies in the dataset.

Short run down: The way deep Learning works is uses the criteria fed into it and picks who won the round/fight. Then the AI actually learns backwards so it needs to pick who won first and then pick it's way of why that is correct. This is a vast oversimplification especially with learning what it needs to score.