Karl Sims

The article's philosophical discussion of "surprise" feels a little klunky and academic to me, and I think conflates some different kinds of surprise, but it's still really fun to think about. Many of the projects here are all inspired by Karl Sims' work in the early 90's. http://www.karlsims.com/ Sims evolved virtual creates by specifying goals to achieve and then running a physics simulation. He noted at the time that the evolution process was great at exploiting bugs in the simulation. https://www.youtube.com/watch?v=JBgG_VSP7f8 I was insipred enough by Sims' genetic images (http://www.karlsims.com/genetic-images.html) that I spent a few years trying to get surprising and beautiful results of my own, with some limited success (https://flic.kr/s/3Xoz).

This is a very interesting addition to the discussion. The digital art world and the demoscene sometimes briefly met, but I assume that most of the time its members look down on eachother. An interesting artist in this respect is Karl Sims [1]. As far as I know, he was not involved in the demoscene, but he sure did make some interesting things! [1] https://www.karlsims.com/

If you haven’t run across it yet, Karl Sims’ virtual creatures are also really fascinating. https://www.karlsims.com/evolved-virtual-creatures.html And anything by Melanie Mitchell. Her book “An Introduction to Genetic Algorithms” is great.

I remember being dumbstruck after watching that video for the first time and realizing the power of programming. Karl Sims website still looks a bit like it did back then http://www.karlsims.com

I'll echo a lot of the praise about the format of this post. It was fun clicking and watching the figures animate as I changed parameters. I was sad that the author missed a number of chances to get into more detail about the classic reaction-diffusion problem [1]. I was reminded of a small project I did which produced similar animations, though with periodic boundary conditions, for learning about the Gray-Scott model. These websites are pretty helpful [2][3]. I haven't ever taken a class on systems so I don't know, but after reading this I wonder if the propagation of "scientific bullshit" and "truth" through a network can instead be modeled chemically as in a reaction-diffusion model. The last figure shows real knowledge fizzling out because it turns fake. It also lacks a slider so I can't play with the parameters but there should be some point where they oscillate back and forth, i.e. a Hopf bifurcation or a Turing bifurcation. Adding a bit more complexity might add some more depth to this post. I hope there will be a sequel! [1] https://en.wikipedia.org/wiki/Reaction%E2%80%93diffusion_system [2] https://groups.csail.mit.edu/mac/projects/amorphous/GrayScott/ [3] https://www.karlsims.com/rd.html

Mandelbrot and Julia sets are related: https://www.karlsims.com/julia.html

how far we've come since https://www.karlsims.com/genetic-images.html quite a bit, but surprisingly not

On the other hand this site also teaches the internet that eventually everything becomes a reaction-diffusion system/phase separating fluids [1]. [1]: https://www.karlsims.com/rd.html

I'm a bit surprised nobody mentioned Karl Sims Evolved Virtual Creatures yet (https://karlsims.com/evolved-virtual-creatures.html).

As a teenager, I had the opportunity to play and experience the monumental installation Genetic Images by Karl Sims installed at Pompidou Center in 1993. It was using a Connection Machine to render images in realtime: http://www.karlsims.com/genetic-images.html The photo at the top of the page is of the installation at Pompidou Center (in the space that is now the staff entrance.) Genetic Images has only been installed 3 times. Its installation at the Pompidou center as part of the Revue Virtuelle was especially significant at the time, as exhibiting a working super-computer in a museum had never been done, as far as I know.

Reminder that Karl Simms has produced some interesting results with Genetic Programming: https://karlsims.com/evolved-virtual-creatures.html https://karlsims.com/papers/alife94.pdf

I was wondering if there was any work in the field of evolving 'creatures' moving in a virtual space, beyond what Karl Sims did 23 years ago. https://www.karlsims.com/evolved-virtual-creatures.html https://www.karlsims.com/galapagos/index.html It would be nice if Karl Sims could open source it as it is really inspiring visual example in the field of Artificial Life, next to seeing the generations of metuselahs unfolding in Conway's game of life Another interesting article in the field of Artificial Life: https://arxiv.org/pdf/1803.03453.pdf

Looks like a lot of those would benefit from a little more evolving. Apparently the generations varied from 50-100. This paper is really interesting, and shows how detailed the simulation is: http://www.karlsims.com/papers/siggraph94.pdf

This is one of those ones I’ll stop and read every time it gets posted. Bit of trivia: in the novel, “Jurassic Park”, the park’s supercomputer is a Cray, but Spielberg picked a Connection Machine for the movie because he thought it looked cooler. Also, the artist Karl Sims did some outstanding work using the Connection Machine that was mind blowing for the time, and still pretty darn cool today. The computer itself was part of the art installation: https://www.karlsims.com/genetic-images.html

Also, simple instructions on how to make one: http://www.karlsims.com/harmonograph/

There are a couple of most plausibly hard steps: the origin of life up to just before bacteria split from archaea, and the origin of eukaryotes, which happened just once after something like 2 billion years. But I had in mind something cheaper more like scaling up the idea of http://www.karlsims.com/evolved-virtual-creatures.html. Intelligence has evolved in animals as unrelated as octopi and us.

Article linked from the video description: http://www.karlsims.com/julia.html

The Connection Machine had some cool art and animation usage. Karl Sims developed an art installation using algorithms to evolve images based on user selections. It was to see for example in Paris at the Centre Pompidou. As you can see on his page, the installation used a bunch of color graphics screens, a Connection Machine (2^15 processors). You can see the machine on the top image on the right. The actual software was written in Starlisp: http://www.karlsims.com/genetic-images.html Karl Sims gives a demo: https://vimeo.com/7723361 Karl also studied particle systems using a CM-2: https://archive.org/details/sims_particle_dreams_1988 Plants: https://archive.org/details/sims_panspermia_1990

Would Second Choice Voting[1] help? [1]: http://www.karlsims.com/second-choice-voting.html

And he did it in 1994: http://www.karlsims.com/evolved-virtual-creatures.html There was also earlier work, e.g. Witkin and Kass, Spacetime Constraints, 1988, but I don't think I've ever seen a video.

Reminds me fondly of Karl Sims 91 paper. It gives an impression of what should be possible with relatively simple expressions: http://www.karlsims.com/papers/siggraph91.html

Reaction diffusion tutorial https://www.karlsims.com/rd.html

Hi, TL;DR Usually in reinforcement learning, the body designs of robot agents are fixed. Here, the agent can learn a task, and also learn a better design of its body at the same time. Link to the blog article: https://designrl.github.io/ (There is a blooper section at the very bottom of the article so check it out!) There's also some active discussion about this article on Twitter: https://twitter.com/hardmaru/status/1049817185055436800 Some background: I have been meaning to write up this fun little experiment for a while now, after working on previous articles related to Reinforcement Learning (RL) and Evolution Strategies (ES) that you may have read before from a previous HN discussion[0]. While I was trying to solve a few RL/ES problems, I found the BipedalWalkerHardcore[1] task extremely frustrating to solve (actually much harder than all of the standard MuJoCo tasks that most papers are based on), but eventually was able to crack it [2] after some effort. I was thinking perhaps the agent's body was not really suited to solve this task, and even minor tweaks here and there would result in making the task easier for an RL algorithm to learn a good set of parameters of the agent's controller neural network to perform well on the task. There has been an exciting line of work on Passive Robotics where researchers such as Tad McGeer[3] and Steve Collins[4] made walking robots that walked on their own naturally without using any external power, unlike complicated,…