Simulating evolution using social bots with multilayer perceptron

[Pulseczar1]

You are a brilliant dude.

[staranjela]

Very enjoyable thank you

[smuggecko]

i enjoyed this vid

[Mjhond]

I am very happy to see the vidoe after you give this This is a simulation of evolution. The bots can sense each other, communicate with each other, and reproduce.

[Ondelendo]

I Love The Video It Can Increase My Knowledge This is a simulation of evolution. The bots can sense each other, communicate with each other, and reproduce

[anakmudajaman]

Steady I Really Like This Video Simulating evolution using social bots with multilayer perceptron

[bebeheuy]

Good, I like that you share this video, I wish success always Simulating evolution using social bots with multilayer perceptron

[imegatrone]

Nice Video That You Share , So Very Nice Thanks You This is a simulation of evolution. The bots can sense each other, communicate with each other, and reproduce

[willamricard]

I Really Like The Video From Your This is a simulation of evolution. The bots can sense each other, communicate with each other, and reproduce

[bundawartini]

Your Video Is Very Useful Sharing Simulating evolution using social bots with multilayer perceptron

[minematas]

It's "The End Of World"

because there are no thumbs down :O

[1658544]

how did i end up here while looking for bots evolution? ohhhhhhh the words bots and evolution...

[justin6532]

really need to do a python tutorial on this. that would be awesome. especially to someone like me who has just started computer programming.

[CaptJiggles]

Is there a download for this, I would love to use it.

[theonethatthinks]

This video is Awesome !!! Very insightful and interesting, :D

[killroy207]

Man your work is awesome. wouldlove to see more.

[wrongfire]

I would love to see you flesh this simulation out further, with more possible variables for the entities to take into account.

[Splatball]

This is very well done. The best evolution simulation that actually looked good and had well displayed results. Is there any way I could get my hands on a copy of this? I'd like to see how long I can run it until either one species rules, or they all die out.

[DinoEntrails]

@Splatball Thumbs up! I want that code!

[messorstructor]

nice work

[Unit047]

Wow this is really fucking interesting actually.

[padaxes]

Excellent video, I'm researching evolution sims at the moment before starting one of my own, this is def one of the more interesting I've found. Very impressed.

[cappie2000]

Awesome work!! Please keep enlighting the general public with your findings... I have the deepest respect for people like yourself that post their findings of these kinds of simulations online, especially in the form of video's that are easily understandable for any lay person!

[badmephisto]

@cappie2000 hi sir thank you. In fact, I have the new version now, and I will be publishing some of that very soon!

[cappie2000]

@badmephisto Hmm.. I was talking about the poster's video :) Are you and phyces the same person? :)

[badmephisto]

@cappie2000 yea I am. Too much work to sign out and sign back in

[johnstifter]

looks like you have done an amazing simulation and nothing more.

eventually to finding no deep understanding of fitness

[badmephisto]

I can only hope that one day I meet a person such as yourself, who by means of thought alone, in his infinite wisdom, has unraveled the deepest mysteries of the evolutionary process.

[Adadr1]

hello, i have a few questions. first, where did you learn about neural networks and organisms like this, and can you make a tutorial on the subject. i really want to learn how to make things like this. ive learned a few things about neurons and neural networks like genetic algorithms, but still am a little clueless on how to implement it. thank you.

[DamianBrownIsCool]

This is truly amazing :O

What kind of computer do you run this on?

[badmephisto]

@DamianBrownIsCool ummm, a shitty old laptop? Would be nicer to have it running on something more powerful, and not python :p python too slo :)

[cliftut]

Have you considered adding some more dynamics to encourage longer term speciation? It looks like there are only one or two niches at the moment, feeders and predators, but if you added another kind of resource that could be optionally collected in a different way, there might be interesting results? Just an idea.

[Cride5]

Wow! What a fascinating video!

Did you do any experiments with varying the size/connectivity of the hidden layer? Would a larger hidden layer perhaps lead to a slower rate of evolution, and possibly more complex behaviour. It would make for some pretty interesting research.

Did you do this as part of a university project?

[phyces]

@Cride5 Thank you! I did do some experiments with that. Not too many because computationally it became very costly for me. But I'll look into ways of speeding it up using modules written in C

[phyces]

@Cride5 Oh and no, this was just something I did in spare time from interest

[Cride5]

@phyces incredible! This is a fantastic application of neural nets to A-life simulation. I'm sure there are be many research possibilities coming from developments of this. If you're in university I would encourage you to present it, and maybe enquire about doing research with it if you're interested.

[qwertasd7]

if possible i'd like to see this, i run python too, learning Pyopengl etc

And am a fan of Darwinbots too, i hope for some merge of these 2 programs.

If this code is free, please put a hyperlink in the darwinbots forum

[DeathSlayer2]

This is even better than the first!

You have inspired me.

[ExtantFrodo]

Very nicely done.

[PsionNinja]

I'd like to get a hold of that program. I am pretty decent at programming. The software I use is called DarkBasic Professional. It's great for making applications such as this.

DBPro can do fast 2D or 3D graphics as well as fast mathematics. Though it's not object oriented programming, it can be pretty flexible to make what you want.

I've always wanted to do some work along these lines, and now that I have the time, I figure why not.

[badmephisto]

Well this is written in Python. I cannot imagine a way I would build this without using objects though! That would be a complete disaster... Every bot is an object, every brain is an object, every neuron is an object... <3 objects :)

[PsionNinja]

I can make objects quite easily such as the bots in this program if thats what you mean. I was talking about the style of programming. for example, in DBPro: Type Bots Xpos as float YPos as float Direction as float Speed as float BotRadius as float Endtype Dim Bots(500) as Bot So you can see in this example you can have 500 bots that each have their own location, size, angle, and speed variables. And because they are floated variables you can have great precision. Sound like this will work?

[stachuk1992]

Hello,

I was wondering what the graph on the board was meant to show. Care to explain?

[badmephisto]

@stachuk1992 you'd have to be familiar with multilayer perceptron. every node is a neuron and every edge is a connection. There is a precise way that the multilayer perceptron works. every neuron adds all of its inputs, adds a bias, and then puts it through a squashing function. You can look it up on wiki

[stachuk1992]

:)

I shall do that.

This stuff is pretty interesting.