Artificial Intelligence Uncovers Alternative Physics
Will the power of observation be augmented?
I have a summer discount going to reward my early backers and most passionate readers. I’m really grateful for the support.
In the A.I. News cycle I came across a most peculiar story. An artificial intelligence program has studied various physical phenomena and identified its own variables for describing them.
But the variables it discovered were unexpected.
Analyzing pendulum videos, the AI identified variables not present in current mathematics.
Think about it, a precursor step to understanding physics is identifying relevant variables. Without the concept of energy, mass, and velocity, not even Einstein could discover relativity.
So what laws will A.I. eventually uncover that are not easy for humans to understand?
This is the question that researchers at Columbia Engineering posed to a new AI program. The program was designed to observe physical phenomena through a video camera, then try to search for the minimal set of fundamental variables that fully describe the observed dynamics.
The study was published on July 25 in Nature Computational Science.
Now, a new AI program developed by researchers at Columbia University has seemingly discovered its own alternative physics.
The article seemed like clickbait to me, but there may be more to it.
Read some of the comments on Hacker News.
One user said:
This reminded me a great deal of a paper I read like 10 years ago, which I just looked up because I wanted to see if it was cited here, "Distilling Free-Form Natural Laws from Experimental Data" . Lo and behold it is not only cited, but is written by the same authors. So the novelty here, as discussed in the paper , is that they are doing something similar but from video instead of from sensor streams. Which is quite interesting, as it opens up the available information to systems that are hard to 'sense' apart from pointing a camera at them, like the lava lamp example.
I did see a presentation a few years ago on a similar topic by Erwin Couwans, author of Bullet physics engine, who was discussing doing neural inference of physics. He basically wanted to replace Bullet with a neural network, which I thought was kind of funny at the time, but cool if it worked.
A.I.’s Observation May Give us More Knowledge
After being shown videos of physical phenomena on Earth, the AI didn't rediscover the current variables we use; instead, it actually came up with new variables to explain what it saw.
So this is all quite interesting if you believe robotics could invent “alternative physics” or A.I. will improve in its capabilities to self-learn as SSL continues to evolve.
See the video.
The image shows a chaotic swing stick dynamical system in motion. The work aims at identifying and extracting the minimum number of state variables needed to describe such system from high dimensional video footage directly. Credit: Yinuo Qin/Columbia Engineering
Even human level AI (HLAI) might interpret the world differently than we do.
The researchers then proceeded to visualize the actual variables that the program identified.
Extracting the variables themselves was not easy, since the program cannot describe them in any intuitive way that would be understandable to humans. After some probing, it appeared that two of the variables the program chose loosely corresponded to the angles of the arms, but the other two remain a mystery.
"We tried correlating the other variables with anything and everything we could think of: angular and linear velocities, kinetic and potential energy, and various combinations of known quantities," explained Boyuan Chen Ph.D., now an assistant professor at Duke University, who led the work. "But nothing seemed to match perfectly." The team was confident that the AI had found a valid set of four variables, since it was making good predictions, "but we don't yet understand the mathematical language it is speaking," he explained.
This topic trended in the Subreddit Futurology, and the comments are also excellent there. Hacker News and Reddit continue to remain some of the better actual human comments on the various news in A.I. and stories on the web related to some of these interesting studies.
One commenter wrote:
The time is gonna come when an AI solves important problems with variables that we can't grasp - that we have no cognitive mechanisms to grasp them with. Problems where the number of dimensions is just not conceivable by a human mind. These solutions will remain "mysterious" to even the best human minds.
The best of these ai solutions to large problems will work (the vast majority of the time) , and we'll just have to "trust them" for our own benefit.
The future is gonna be... weird.
I speculate that if AGI was developed, the AI would quickly see patterns the sum total of humanity does not see, and thus gain a competitive advantage while having easy access to all of our knowledge and their own. I don’t personally foresee this happening in our lifetime but certainly possible perhaps in the late 2080s. Historically of course, AGI has been “far further” in the future than many of the early researchers had even imagined.
The study opens up a lot of philosophical questions I found stimulating.
A particularly interesting question was whether the set of variable was unique for every system, or whether a different set was produced each time the program was restarted.
This new AI only looked at videos of a handful of physical phenomena, so it's in no way placed to come up with new physics to explain the Universe, but something about the results feels fresh and unexpected.
AI and Science
I’ve been seeing a lot more work based on how AI might augment science and the process of scientific discovery. I’ve written extensively this week about Microsoft’s AI4Science initiative.
This study however is the work is part of Lipson and Fu Foundation Professor of Mathematics Qiang Du's decades-long interest in creating algorithms that can distill data into scientific laws. You will notice already today many of the best academics in A.I. are of Asian descent.
The number of graduates and PhDs China will be able to produce in the 2020s and 2030s in A.I. will be stunning, apart from their dominant place at U.S. Universities and in academia already in 2022.
The Science Fiction Angle
The robots are coming, and they will know science and maybe bring their own science with them.
"I always wondered, if we ever met an intelligent alien race, would they have discovered the same physics laws as we have, or might they describe the Universe in a different way?" says roboticist Hod Lipson from the Creative Machines Lab at Columbia.
All of this is most likely to find fruition in the end in Chinese research.
As the varieties of our approaches multiply (like different AIs), the different perspectives on reality might also expand.
"In the experiments, the number of variables was the same each time the AI restarted, but the specific variables were different each time. So yes, there are alternative ways to describe the Universe and it is quite possible that our choices aren't perfect."
Einstein’s laws are elegant for instance, but perhaps just a reality we chose to accept by consensus. This relationship of truth, the observer and the universe is truly fascinating.
Lipson thinks a bit like we all secretly do:
"I always wondered, if we ever met an intelligent alien race, would they have discovered the same physics laws as we have, or might they describe the universe in a different way?" said Lipson.
Of course many of us have thought this same thing.
"Perhaps some phenomena seem enigmatically complex because we are trying to understand them using the wrong set of variables.”
This is correct. Even in our exploration of AI and pursuit of AGI, we may be using the wrong set of variables.
You are reading Data Science Learning Center, where I’m still optimizing my content strategy but often cover Future of Work, Software programing and some A.I. article trends. Would like to begin to cover more Data Science topics as well.
Will A.I. Augment Scientific Discovery in the 21st century?
Some of the broad conclusions of this research are also important:
The researchers believe that this sort of AI can help scientists uncover complex phenomena for which theoretical understanding is not keeping pace with the deluge of data—areas ranging from biology to cosmology. "While we used video data in this work, any kind of array data source could be used—radar arrays, or DNA arrays, for example," explained Kuang Huang, Ph.D., who co-authored the paper.
The work is part of Lipson and Fu Foundation Professor of Mathematics Qiang Du's decades-long interest in creating algorithms that can distill data into scientific laws.
Past software systems, such as Lipson and Michael Schmidt's Eureqa software, could distill freeform physical laws from experimental data, but only if the variables were identified in advance. But what if the variables are yet unknown?
This is clear thinking:
Lipson, who is also the James and Sally Scapa Professor of Innovation, argues that scientists may be misinterpreting or failing to understand many phenomena simply because they don't have a good set of variables to describe the phenomena.
Good Science Asks the Right Questions
"For millennia, people knew about objects moving quickly or slowly, but it was only when the notion of velocity and acceleration was formally quantified that Newton could discover his famous law of motion F=MA," Lipson noted. Variables describing temperature and pressure needed to be identified before laws of thermodynamics could be formalized, and so on for every corner of the scientific world. The variables are a precursor to any theory.
"What other laws are we missing simply because we don't have the variables?" asked Du, who co-led the work.
The paper was also co-authored by Sunand Raghupathi and Ishaan Chandratreya, who helped collect the data for the experiments.
How robotics and science become entangled with the future of science sounds pretty sci-fi to me but we are entering an age of automation, convergence and software that will feel a bit more quantum.
I hope you enjoyed the topic!
If you want to support my work or get access to additional content, you can support for the price of a cheap coffee.
Not all innovation is from new ideas, sometimes A.I. enables us to reinvent the wheel in a way that allows us to pose new questions and hypothesis. I believe the advent of A.I. and Quantum computing, or AQ, is a significant milestone in humanity’s search for more answers and half-truths about the universe.
Thanks for reading!