I often lag behind in my email responses. PM me to let me know which one you are (there are about a dozen or so) and I'll send a response right away. Otherwise it could be a couple days before I do a burst of responses to catch up.
yeah... but now the evolution you gave me froze once at just the wrong time (maybe while it was editing the evolution file?) and its corrupt and wont open without it freezing. Also, Ive found some cool creatures just by doing random creature over and over, but I dont know how to use the saved .creatures in my evolutions...
Sorry for the slow reply. Both of those issues can be solved, but it would be too much to explain here. Send me an email and I'll tell you how to a) repair your corrupt .evolution file, and b) use any creature you like as the starting point for a new evolution run. It involves a little bit of file editing, but nothing too drastic.
Anything at all, really. In fact, I'll share the program with anyone who's interested, with or without contributing processing time in return. By the end of the summer I'll probably just remove myself from the middleman position as well and just post the download link on the project web page since I'll be done collecting data by then.
I like the swarm intelligence stuff, but unfortunately my software is so slow that going from one creature to several creatures (even just a few) is just not very feasible. But I do like the various ant algorithms I've seen, and in the AI course at the university we get students to use GP (genetic programming) to evolve programs for a swarm of ants. It's quite fun!
Thanks for the comments. I'm glad you're enjoying the videos.
It could clearly improve some more, but it's a run from one of my "official" experiment configurations, and they only go to generation 500 in that configuration.
I think that's a good strategy. A large initial population to supply lots of genetic material to choose from, and later a smaller population for more rapid (in terms of computer time) refining of the body plan that it settles on. Eventually I'll try to make population size one of the parameters in the fitness schedule part of the program, so it can be made to increase or decrease whenever the user plans.
can you make a video on how to use the program?
erti655 4 days ago
Love this song from Mario THANK YOU! haha
ScientificExploits 4 months ago
I sent the email like one or two days ago...
Yumminess0 3 years ago
I often lag behind in my email responses. PM me to let me know which one you are (there are about a dozen or so) and I'll send a response right away. Otherwise it could be a couple days before I do a burst of responses to catch up.
kjlg74 3 years ago
I am the one with two emails and the subject is "downloads". I use a different email than my gmail for normal messages.
Yumminess0 3 years ago
I just replied. Did I get the right one?
kjlg74 3 years ago
yeah... but now the evolution you gave me froze once at just the wrong time (maybe while it was editing the evolution file?) and its corrupt and wont open without it freezing. Also, Ive found some cool creatures just by doing random creature over and over, but I dont know how to use the saved .creatures in my evolutions...
Yumminess0 3 years ago
Sorry for the slow reply. Both of those issues can be solved, but it would be too much to explain here. Send me an email and I'll tell you how to a) repair your corrupt .evolution file, and b) use any creature you like as the starting point for a new evolution run. It involves a little bit of file editing, but nothing too drastic.
kjlg74 3 years ago
What is a fair donation in terms of runtime?
spentfromnz 3 years ago
Anything at all, really. In fact, I'll share the program with anyone who's interested, with or without contributing processing time in return. By the end of the summer I'll probably just remove myself from the middleman position as well and just post the download link on the project web page since I'll be done collecting data by then.
kjlg74 3 years ago
cool!
maybe you can make with the creatures more than just walking?
erzkonservativ 3 years ago
I'd like to add more abilities. At the moment the software can reward creatures for any of the following, or a mix of them (decided by the user)...
* distance traveled
* upward movement (like jumping)
* average height (i.e. being tall)
* catching falling spheres
* remaining on the ground (to discourage jumping)
You can see examples of all those in the various videos - although most of the videos are traveling creatures.
In future versions I'll try to expand on that and see what I can do.
kjlg74 3 years ago
i love to see science here on youtube
it is very seldome :(
your project based of one creature...
what do you thing about swarm / collective intelligence?
maybe a project about evolution of swarm intellicent, e.g. ants
find food, destroy emenies
and sorry for my poor english, iam from germany ;)
erzkonservativ 3 years ago
Kein Problem. Alles verstanden ;)
I like the swarm intelligence stuff, but unfortunately my software is so slow that going from one creature to several creatures (even just a few) is just not very feasible. But I do like the various ant algorithms I've seen, and in the AI course at the university we get students to use GP (genetic programming) to evolve programs for a swarm of ants. It's quite fun!
Thanks for the comments. I'm glad you're enjoying the videos.
Mach's gut!
kjlg74 3 years ago
Generation number?
TheReasonWhyGuy 3 years ago
Generation 500, population size 100.
It could clearly improve some more, but it's a run from one of my "official" experiment configurations, and they only go to generation 500 in that configuration.
kjlg74 3 years ago
random "advice"
Have the first 10 generations have a population of about 200.
From what I can tell it makes it easier to sort out crappy body designs.
Then again those crappy bodies might become interesting later.
TheReasonWhyGuy 3 years ago
I think that's a good strategy. A large initial population to supply lots of genetic material to choose from, and later a smaller population for more rapid (in terms of computer time) refining of the body plan that it settles on. Eventually I'll try to make population size one of the parameters in the fitness schedule part of the program, so it can be made to increase or decrease whenever the user plans.
kjlg74 3 years ago
Man, they're so frustrating to watch.
closedmouth 3 years ago
You mean when they trip up and momentarily get stuck? Yeah, definitely a little frustrating. :P
kjlg74 3 years ago
Gen number?
TheReasonWhyGuy 3 years ago