General Yuzuru Chat

[guitarist]

7 hours ago, Melodie said:

He even listed the PC he used to write his thesis 

 

 

I think this makes sense, though! It's not so much about the PC that was used to write the thesis, but the computer that was used to render the visualizations. It does tell me that whatever software he used doesn't require a powerful computer to run (mediocre graphics card, i7 + 16gb ram combo is good, but it isn't unusual or outrageously expensive anymore). I have no idea what the thesis talks about, but since it does mention feasibility of use, I think it'd be important to be able to reproduce this without needing extremely expensive computers. 

[BWOZWaltz]

4 hours ago, amylance1215 said:

It's surpassed 10k downloads. The academic journal article to break FS and the internet lol

 

 

It's over 22K now  😂😂😂

[Kizuna]

Downloads over 22K now, and I tried to check the download numbers for other papers and... 

[amylance1215]

This sums it up

 

[yuzuangel]

9 minutes ago, amylance1215 said:

This sums it up

 

I'm so proud of Yuzu...the reason why no competitive skater talks about this is probably because of good reason, i.e. the powers that be won't like it.. But they already treat him like shit. And he's above that. They can underscore him but Yuzu is already above the ISU, they can't touch his legacy.

[yuzuangel]

i'm laughing

 

[senemgencer]

I am very impatiently waiting for the translation of his work. I will then include some articles from Martina Frammartino where she gives numerical evidence and lots of stats about underscoring. I will then make lots of copies and send them to every ISU board member and judge. ISU HQ is only 2 hours away from where I live

[tafattsbarn]

The more excerpts i read from it, the sexier his thesis becomes 

[JanMary]

1 hour ago, senemgencer said:

I am very impatiently waiting for the translation of his work. I will then include some articles from Martina Frammartino where she gives numerical evidence and lots of stats about underscoring. I will then make lots of copies and send them to every ISU board member and judge. ISU HQ is only 2 hours away from where I live

 

Actually DeepL translator is working quite well. I can understand 90% of the translation. You should definitely give it a try.

[Paskud]

Why do I feel like.... implementing AI in scoring is easier to say than to do?

Imagine collecting database for this AI.

Imagine the drama when skater x isn't invited to collect data about correct technique.

Imagine convincing figure skating community (skaters, coaches,...) that database with these and no other skaters isn't biased.

Imagine convincing ISU that AI marking as incorrect jumps of you-know-which-skaters is a good one.

[yuzuangel]

2 minutes ago, Paskud said:

Why do I feel like.... implementing AI in scoring is easier to say than to do?

Imagine collecting database for this AI.

Imagine the drama when skater x isn't invited to collect data about correct technique.

Imagine convincing figure skating community (skaters, coaches,...) that database with these and no other skaters isn't biased.

Imagine convincing ISU that AI marking as incorrect jumps of you-know-which-skaters is a good one.

well...the human aspect of it (i.e. the last two points) would be the hardest. but collecting data is actually probably not that expensive and everyone should be invited to participate in that if they want. and there should be a rescore option where judges can manually look at calls to determine if they're correct, that skaters can request for a fee like in gymnastics, which will be possible if we have good cameras (and usually TV cameras will do).

[rockstaryuzu]

8 minutes ago, Paskud said:

Why do I feel like.... implementing AI in scoring is easier to say than to do?

Imagine collecting database for this AI.

Imagine the drama when skater x isn't invited to collect data about correct technique.

Imagine convincing figure skating community (skaters, coaches,...) that database with these and no other skaters isn't biased.

Imagine convincing ISU that AI marking as incorrect jumps of you-know-which-skaters is a good one.

Exactly. the devil is in the details. Yuzu's paper is only a proof of concept. There's miles to go yet. 

 

But it's the important first step. 

 

[Fay]

20 minutes ago, rockstaryuzu said:

Exactly. the devil is in the details. Yuzu's paper is only a proof of concept. There's miles to go yet. 

 

But it's the important first step. 

 

The thing is they have to want to go in that direction. And it’s in that aspect that I have serious doubts. 

[CaroSkate]

1 時間前, Paskudさんが言いました:

Why do I feel like.... implementing AI in scoring is easier to say than to do?

Imagine collecting database for this AI.

Imagine the drama when skater x isn't invited to collect data about correct technique.

Imagine convincing figure skating community (skaters, coaches,...) that database with these and no other skaters isn't biased.

Imagine convincing ISU that AI marking as incorrect jumps of you-know-which-skaters is a good one.

Every revolution that could potentially harm the ones in power is hard to implement, but if we start the first step and keep working on it, we may succeed in the future.  Even if it doesn't apply to judging, it may still be a great tool for training and the data from Yuzu's skating/jumps are invaluable. 

[barbara]

3 hours ago, senemgencer said:

I am very impatiently waiting for the translation of his work. I will then include some articles from Martina Frammartino where she gives numerical evidence and lots of stats about underscoring. I will then make lots of copies and send them to every ISU board member and judge. ISU HQ is only 2 hours away from where I live

Also the Olympic Committee, please.