AI and Physics: Cosmology
Вставка
- Опубліковано 12 вер 2024
- The recent development of AI presents challenges, but also great opportunities. In this clip I will discuss how cosmology could profit from AI.
Mind also my backup channel:
odysee.com/@Th...
My books: www.amazon.com/Alexander-Unzicker/e/B00DQCRYYY/
AI (what we call AI) can help detect common data patterns which are not immediately evident to the human data mining.
It is incredible how what we - for years - have called "Data Mining" and "Data Analysis" are now been merged together, and called "Artificial Intelligence". With the "Data Training" being no more than an overextended calculation of matrix differentiation. It is incredible how far this system is from true intelligence, which is a biphasic operation of "Invent and choose".
Best Regards, dear Prof. Unzicker...
Greetings,
Anthony
It is a marketing designation.
I don’t think AI is really artificial intelligence. It can recall data from a massive depository of data but it would still need to use models of some sort to analyse the data. The models will be no better than existing models and might therefore be an incorrect model that it chooses. AI could only at best duplicate left hemisphere brain operation. There is no right hemisphere AI equivalent to act iteratively to question the ‘left hemisphere’ AI model. A good data analyst wouldn’t be intimidated by the quantity of data. Bayesian analysis would be a preferred choice. Even the best data analysts would need input from open minded cosmologists and be prepared to rethink standard approaches to data analysis. AI would not cut it at this stage.
They could make the presumption that distant objects that look the same as close objects are the same: in energy output and size. Then see what comes out of that?
Whoops, careful there William, that is coming dangerously close to disproving the Big Bang hypothesis there my friend. I hear they cancel you for that kind of talk!
Interviewing AI is an art. Open questions are not helpful with ChatGPT. Tom Lausen (on a different topic) demonstrated very clearly how to deal with AI in his book " Die Untersuchung". You also have to feed this AI with data yourself and also point out that standard cosmology is full of free parameters.
As far as the shape of galaxies is concerned, isn't it also a question of perspective (viewing angle) whether a galaxy appears as a bar, an ellipse or a spiral? Hannes Alvén had published groundbreaking work on galaxy formation. Not that a Nobel Prize is necessarily a reference, but he was also a Nobel Prize winner.
Being the doubting Thomas (or doubting Ernest), I wonder how much the AI programming itself would be manipulated by the human hand that is trained to selectively process data (and omit data they don't want to process) to arrive at the desired results.
people are finding these Ai are "left leaning" almost socialist.
AI refused to discuss vsl, that was annoying. For example, what is the speed of gravity if it should all be constant and warp space time?
you put so much faith in "AI" .... what they provide as "AI" today can only find the answers it's told to find.
Classifying galaxies, sure, train a 6 year old one-by-one for each classification and if you feed it the examples you used to train it it would agree with you, feed it data of another it hasn't seen before, good luck, or you could sit down and train it for that one.
Don't get sucked in by the hype, "AI" really is not what you think it is.
the field of AI and analytics overlap tremendously, machine learning is under the umbrella of ai but has a big problem of "framing" and bias. i have no doubt that if these datapoints were set in a large vector database and k nearest neighbor calculations done, some of what was mentioned in the video could be done, as well as support vector analysis. i think the question is if the framing of the result is what you wanted, arguably you may get almost infinite scope of categorical data, which isnt useful. richness of the data is another question. but i think when it comes to data science its the superstition that ai needs to deal with rather than superstition in ai.
Artificial = Fake. Intelligence = News.
I find it useful generally with physics problems, but best double check any maths they do because they seem too eager to please. I even had an AI lie to me using fictious history (about 'Maxwell Gravity') to justify its argument!
It claimed that Maxwell used spacetime with his theory about gravity, which is, of course, nonsense!
Super
Thanks
nice, thanks
Teaching a new dog old tricks seems like a blind alley with the data input with key questions not adressed even in AI would form "an opinion" of reality based on only the same data. Even without inputing the flawed data arguably.
We do have a lot of data, it can't have been another time in history when we had so much info!
Ai can extend your skill set but not replace it. At this point any meaningful work is human led. Human directed and ai assisted.
I have spent days fixing its mistakes but even still it is a work accelerator.
An almost living notebook
Sir I wish you can write a book or at least do a video on scientific thinking
❤❤❤❤❤❤❤❤❤ the ducks THE DUCKS ! !
- We live in the same climate as it was 5 million years ago -
I have an explanation regarding the cause of the climate change and global warming, it is the travel of the universe to the deep past since May 10, 2010.
Each day starting May 10, 2010 takes us 1000 years to the past of the universe.
Today May 27, 2024 the position of our universe is the same as it was 5 million and 131 thousand years ago.
On october 13, 2026 the position of our universe will be at the point 6 million years in the past.
On june 04, 2051 the position of our universe will be at the point 15 million years in the past.
On june 28, 2092 the position of our universe will be at the point 30 million years in the past.
On april 02, 2147 the position of our universe will be at the point 50 million years in the past.
The result is that the universe is heading back to the point where it started and today we live in the same climate as it was 5 million years ago.
Mohamed BOUHAMIDA, teacher of mathematics and a researcher in number theory.
today's "ai" is not hollywood "ai". it's a language program.
Have you worked with ai on your work yet. It sounds like you have not. I have and my nuanced experience is not covered at all but you.
At is a capable savant with a short attention span and bouts of memory loss.
If your making new math and telling it what it all means Iit can make simulations with say wolfram to check on your ideas vs known data .
It is a lot of work but it is already a completely different new tool
Instead of speaking about what AI _could_ do, what has already been done? Apart from anecdotal, spurious, and frivolous gimmicks, nothing really substantial. We are very far from understanding the human intelligence, actual artificial intelligence reflects it.
I see you have an opinion, let's wait for whether it ages well.