AI just independently discovered alternative physics

(ORDO NEWS) — Pick up any physics textbook and you’ll find formula after formula describing how objects wobble, fly, deflect, and stop.

The formulas describe actions that we can observe, but behind each of them there may be a set of factors that are not immediately obvious.

Now a new artificial intelligence program developed by researchers at Columbia University appears to have discovered its own alternative physics.

After the AI ​​showed videos of physical phenomena on Earth, the AI ​​did not rediscover the current variables we use; instead, he invented new variables to explain what he saw.

To be clear, this does not mean that our current physics is wrong or that there is a better model to explain the world around us. (Einstein’s laws turned out to be incredibly reliable.)

But these laws could only exist because they were built on the basis of a pre-existing “language” of theory and principles established by centuries of tradition.

Given an alternative time when other minds were solving the same problems from a slightly different perspective, will we still formulate the mechanics that explain our universe in the same way?

Even with new technology to visualize black holes and detect strange, distant worlds, these laws are proven time and time again (note: quantum mechanics is a different story, but let’s focus on the visible world here).

This new AI has only watched videos of a few physical phenomena, so there is no way it can come up with new physics to explain the universe or try to outdo Einstein. This was not the goal here.

“I’ve always wondered, if we ever met an intelligent alien race, would they discover the same laws of physics as we do, or could they describe the universe in a different way?” says roboticist Hod Lipson of Creative Machines Lab in Columbia.

“In the experiments, the number of variables was the same each time the AI ​​was restarted, but the specific variables were different each time. are alternative ways of describing the universe, and it is possible that our choice is not ideal.”

In addition, the team wanted to know if AI could actually find new variables and therefore help us explain complex new phenomena emerging in our current data stream that we currently lack the theoretical understanding to keep up with.

For example, new data from giant experiments such as the Large Hadron Collider that hint at new physics.

“What other laws are we missing simply because we don’t have variables?” says mathematician Qiang Du of Columbia University.

How does AI find new physics? To begin, the team uploaded raw footage of the phenomena they already understood into the system and asked the program a simple question: What are the minimum fundamental variables needed to describe what is happening?

The first video showed a double pendulum swinging, which is known to involve four state variables: the angle and the angular velocity of each of the two pendulums.

The AI ​​pondered the footage and the question for several hours, and then came up with an answer: 4.7 variables would be required to explain this phenomenon.

It’s close enough. up to four that we know about… but that still didn’t explain what the AI ​​thought were variables.

Therefore, the team then tried to match the known variables with the variables selected by the AI. Two of them roughly corresponded to the angles of the arms, but the other two variables remained a mystery.

However, the AI ​​could make accurate predictions as to what the system would do next, so the team figured the AI ​​must have been dealing with something they couldn’t understand.

“We tried to match other variables with anything. and everything we could think of: angular and linear velocities, kinetic and potential energy, and various combinations of known quantities,” says software researcher Boyuan Chen, now an assistant professor at Duke University, who led the work.

“But nothing matched perfectly… we don’t yet understand the mathematical language he speaks.”

The team then showed the AI ​​other videos. The first shows the wavy arm of an “air dancer” fluttering in the wind (the AI ​​said this has eight variables). The lava lamp footage also produced eight variables. The flame video clip is back with 24 variables.

Each time the variables were unique.

“Without any prior knowledge of the underlying physics, our algorithm discovers the intrinsic dimension of the observed dynamics and identifies candidate sets of state variables,” the researchers write in their paper.

This suggests that in the future, AI has the potential to help us identify variables that underlie new concepts that we are not currently exploring. I know. See this place.


Contact us: [email protected]

Our Standards, Terms of Use: Standard Terms And Conditions.