Mechanical Engineering

210 readers

5 users here now

Welcome to the Mechanical Engineering Community!

Rules:

1.) Be constructive and respectful.

2.) No advertising/self-promotion.

3.) No low-effort posts.

4.) No "design this for me" posts.

5.) Images must be relevant to Mechanical Engineering or the posted topic.

6.) We're not doing your homework for you.

founded 1 year ago

MODERATORS

Alteon

AI has been used for a breakthrough in batteries (www.theengineer.co.uk)

submitted 10 months ago by Alteon to c/mechanical_engineering

12 comments fedilink hide all child comments

you are viewing a single comment's thread
view the rest of the comments

[–] Alteon 2 points 10 months ago (1 children)

AI is highly important as it can take in a massive amount of information and find trends that people likely can't. Especially in R&D applications, AI is going to be almost necessary to keep up with the pace of other developments in other fields. Smart materials in particular are going to see a boom with the advent of smarter AI.

[–] rdyoung 1 points 10 months ago* (last edited 10 months ago) (1 children)

You're still using AI as a buzzword. None of what is being attributed to AI is actually AI. This is the kind of thing that distributed computing has been doing for decades. Advances in processing power and code efficiency have made things easier but it's still not AI.

Again I'll ask. How is this different from boinc, etc?

Here is an article from Amazon about distributed computing that doesn't use AI incorrectly or at all.

https://aws.amazon.com/what-is/distributed-computing/#:~:text=Distributed%20computing%20is%20the%20method,to%20deal%20with%20complex%20challenges.

[–] [email protected] 4 points 10 months ago* (last edited 10 months ago) (1 children)

Distributed computing is completely unrelated. That's just running a program across many computers.

AI is the simulation of intelligence. The main one people know is a neural network.

The code they run it is very simple. It's just a series of nodes that apply a weight to their input and pass it to connected nodes. Basically just multiplication and compared to a threshold.

But what that simple code is able to achieve once trained is so much more than its code. That's the key. The same way everything we experience is just our neurons firing.

AIs ave figured out how to calculate the positions of planets in the future from training data in the past. Which means figuring out the scientific equations needed to calculate that.

So AI can potentially figure out equations that we don't even know as long as input and output data is available.

They are experimenting with this for the magnetic fields that contain fusion reactions. The AI can spot patterns in the data indicating a breach that we don't understand.

You can't just write a program to do that.

[–] RubberElectrons 2 points 10 months ago (1 children)

Where did you see this for fusion reactions?

[–] [email protected] 2 points 10 months ago* (last edited 10 months ago) (1 children)

https://www.nature.com/articles/s41586-021-04301-9

Can't find the article I read years ago. It may have been a misinterpretation of this one tbh.

[–] RubberElectrons 1 points 10 months ago (1 children)

Dude that's a great paper, thanks kindly. I still don't think that's the pubic understanding of general AI, more neural networks inferencing control methods... But it doesn't really matter. At least this is useful.

[–] [email protected] 1 points 10 months ago

No worries. No one is talking about general AI though.