How Supercomputers Are Combating Climate Change

#shorts

[CRuM]

Oh no! It's the beginning of "Sky" net.

I just pulled a Matt! 😁

[reinerjung]

Earth system models based on machine learning are very controversial regarding their ability to predict the future, as they are trained based on past data from observations and mathematical models. So it is unclear whether they just reproduce known behavior and skip over behavior that may occur in nature drastically when reaching specific points. Furthermore, their function cannot be mapped back to physical laws, making their predictions not comprehensible.

[Chris.from.]

I must say that “predicting climate change, ” which is what you are talking about, and “combating climate change” which is your title, are two very different things. What the human race, and the planet, need, now, desperately, are actions, not words or predictions, that combat climate change. I am afraid for my grandchildren, and you should be, too.

[Bob]

Wow, maybe they can loan it to the local weather forecaster!

[angharadhafod]

Nvidia: look what we're doing to help combat climate change
Also Nvidia: buy our energy hungry graphics cards even though you don't really need them

[WiReDApe]

Matt you are always my go to source for emerging tech. You are very important to me.

[antimattv]

Naomi Wildman came up with this idea in the Delta Quadrant. :P

[MayankJairaj]

Congratulations on the 1 million mark!
Also we get the first shorts to celebrate 🥳

[WiReDApe]

I love this YT short. I would love to see more of these on the days you don't upload.

[sbrl]

Wow, this is similar (but not quite the same) as my own research. I'm looking at developing an AI to predict water depth in floods from rainfall radar data. Will there be a recording of this event?

[JohnSmith-kffc]

Cant wait to hear about it and congrat on the 1M Matt!

[joeysultan]

The link to the event isn't working.

[barmpity]

Hi Matt, is this the right session? It's the only Climate/Weather topic session I see regarding Earth-2 on Sep 22nd but you're not listed as a host and it's a replay of the original talk...

Watch Party: NVIDIA'S Earth -2: Digital Twins For Weather and Climate [WP41326b]
Hosted by:
Karthik Kashinanth, Principal Engineer and Scientist, AI-HPC and Engineering Lead–Earth-2, NVIDIA
Ilene Locker Carpenter, HPE, Earth Sciences Segment Manager

Seems to be the replay of this session on Tuesday, which is not interactive:
NVIDIA'S Earth -2: Digital Twins For Weather and Climate [A41326]

More detail in the description (which is hard to find if you're using mobile, FYI) regarding which of the sessions you'll be involved in would be a huge help. I've filtered by English language and HPC - Climate topic to find these. There are 328 English Language sessions on the schedule so it's half way impossible to find the right one without help.

[joeysultan]

The link to the event sign up isn't working.

[markp]

Hate, hate, hate these shorts. Believe it or not, I really can manage to pay attention for more than 30 seconds.

These are a rash across many of my YT subs lately, and really clutters things on AndroidTV (my player of choice).

Yes, I can ignore them, but lately there are MORE shorts than actual content.

If this is an "experiment" of some sorts, please accept my feedback as a datapoint - "Please STOP with the shorts."

[ridgemannikko]

I’m most interested in the teaching the computers physics part. Having a program that understands the physical interactions of things could help so much

[BygoneGent]

How complex do they do they plan to evolve this emulation tech? Will the public be able to interact with it as a virtual reality platform where they can witness the effects of personal and community footprints, which then can be an interactive learning tool for how to optimize decision making in “Earth 1”?

[alansarthou]

They should add machine learning to the mix and we'd all be less angry towards our weather men :)

[benjaminmuscat]

Will it be recorded for later streaming? This is 3am in my timezone so it's unlikely I'll make it

[ericolens]

A tangible use of heavy processing power.
Hopefully, it yields better and more tangible results than digital currency.