An Overview of Artificial Creativity

Artificial Creativity Banner

It’s an unfortunate fact that there’s no single, online information source that covers machines that can compose music, invent patentable ideas, or make up stories on their own (yes, all of these exist). So, let’s fix that, shall we?

Today I’m introducing something new on Think Artificial: A Living Article. It’s not exactly a blog entry, wiki nor a static page, but an article that I intend to gradually add to and improve upon. The article is an easy-read overview of artificial creativity, or at least the beginning of one. You can always find it through the main menu, and I’ll be posting notifications whenever it’s significantly updated. Moreover, I invite you to help create it!

I thought it’d be cool to try see if we couldn’t put together a bit of an overview of the field, along with examples, a link repository and photo collection of research currently being carried out. As I announced recently, I’m going to be covering a bit more of creativity here and I’ll be using material from those entries to gradually build up the Living Article.

Living Article?

Yeah, so I’m not inventing anything new here. It’s just a webpage that gets updated and built up gradually. But as it’s being made explicitly for continual updating rather than being a typical static webpage, and because it isn’t a Wiki page (only I edit, for now), I thought’d would make sense to label it somehow. Mainly I just wanted a name for it, and I think ‘Living Article’ sounds good. Plus, it allows for an ultra-cool double helix banner ;)

Take Part in the Action!

Many people, sharing the same headTo make this more exciting and informative, I’d like to ask you to send along questions or comments you might have on the artificial creativity! To give you some ideas and get that gray matter cranking:

  • If you make a program that invents, is it really the program that’s inventing rather than the programmer?
  • Is a computer that can produce art really creative if it can’t appreciate aesthetics?
  • Can computers ever really create anything new, they only do what we tell them to, don’t they?

I can’t promise I’ll be able to answer all of them, but I can promise that I’ll do my best. And, of course, I welcome any suggestions, links and general comments on the page! One more thing, if YOU want to answer something, or discuss something on creativity — drop me a line and we’ll see if we can either integrate it into the Overview or alternatively make it a guest post here.

And then…?

I’m going to see if this is a good idea. See what the response is like. Perhaps I’ll put up more articles like this here if things work out. Maybe we could even donate it eventually to Wikipedia, seeing how the artificial creativity page there is a bit sad to look at. But for now this single page stays up and I look forward to hearing your thoughts on it! Currently, it’s styled like an encyclopaedic article — pretty dry and reciting facts. I’m wondering whether it would serve more purpose that way, or if I should make it more casual.

What do you think? Encyclopaedic or casual? (I’m just talking about the article, not the posts on Think Artificial. They have their own style.)

Links and References

No related posts detected.

13 Comments, Comment or Ping

  1. Hi Hrafn,
    I just came across this post (somehow…) and realized I’ve made an small essay on just this topic. Maybe you’ll like it? See http://ekok.nl/tech/2007/07/30/essay-on-artificial-creativity/
    Very nice idea though, because I agree the Wikipedia page is tiny (although a lot better than it was a couple of months ago).

  2. Hrafn

    Hey Eric,
    thanks for link! A bit busy today but I skimmed through it briefly and it looks pretty good. I’ll have to give it a more thorough read later.

    Do you do creativity research, or is this something that you did on the side?

  3. I have no creativity at all compared to my computer. It has been creating masterpieces while I have been blogging the exact same thing a group of six hippos on L.S.D. would have come up with sometime eventually anyways.

  4. I’m a ‘Agent Technology’ student and wrote this essay for a ‘Philosophy of AI’ course. Creativity research isn’t my prior field of research, but I do have special intrest in it.
    Nice blog you have here by the way. :-)

  5. Nice website, I hope in time this web page will be filled with amazing ideas and contributions from talented people around the world.

    I am very interested in AI, and have wrote a lot of things on it, mostly based on my own believes since I did not specialize in it in school.

    I think before we can define Creativity, we need to first define intelligence. Or rather explain the process in being intelligent. Because intelligence is still a relative term, and when humans are compared do monkeys, we are more intelligent; when monkeys are compared to chickens, monkeys are more intelligent, and when chickens are compared to a worm, the chicken is more intelligent.

    So clearly, all living things have some sort of intelligence, but some are more capable than others. Which is why we have the concept of IQ, because it is very difficult to measure ones intelligence; we can only measure the outcome of the process, or algorithm that has took place. To measure the exact path, or efficiency of our thinking process is a mystery.

    I believe Intelligence is a set of principles to allow something, an object, to be ALIVE, to exist not only in the world, but also in its own framework. The base of intelligence is far beyond the modern PHD studies such as Speech recognition, Image recognition, and even, CREATIVITY.

    If you think about it for a second, before we even dig into something we called the “brain”, we are all here because “something” has created a framework flexible enough to sustain not only objects, but also organisms. The process of Evolution itself is a genius algorithm that one can say is intelligent. The exact chemical reactions, pulse of our heart, the air we breath, the shit we leave behind, is all self learned, self adapted through evolutions algorithm, that makes our very existence a very complex intelligent system.

    So to summarize the point, the principles of being intelligent it’s ability to “survive” to “live”. The process, or the algorithm of being intelligent is derived from the process of evolution: the need to be ACTIVELY Involved in the environment to satisfy those principles to live.

    With that being said, artificially, computers are essentially programmed. But if we are only programming a framework that enables it to be Actively Involved in its world, we answer your question:

    “If you make a program that invents, is it really the program that’s inventing rather than the programmer?”

    The programmer’s role itself is very intelligent. But, we by reproducing the ingredients and processes for intelligence, we have TRUTHLY passed on our ability into computers, therefore making the computer intelligent as well.

    So my question is, can computers be more intelligent than it’s creators?

  6. Hrafn

    @Eric
    Ah, I see. I haven’t gotten around to reading your essay. Hopefully this weekend. Thanks for the compliment, and feel welcome to drop by anytime :)
    ——–

    @yang yu
    Nice website, I hope in time this web page will be filled with amazing ideas and contributions from talented people around the world.

    Thanks! Hope so too.

    I think before we can define Creativity, we need to first define intelligence. Or rather explain the process in being intelligent.

    But what if intelligence cannot be defined without reference to creativity, and vice versa? Personally, my research has led me to think that creativity is an aspect of intelligence.

    all living things have some sort of intelligence, but some are more capable than others.

    Indeed, our limited understanding and lack of definition of intelligence is not making things easier. It also boils down to the question of when automation stops and intelligence begins.

    I believe Intelligence is a set of principles to allow something, an object, to be ALIVE

    Hehe. Unfortunately there are even some debates surrounding what it is to be alive (see Virus).

    “something” has created a framework flexible enough to sustain not only objects, but also organisms.

    Well, technically an organism is an object. An object that has abilities for sustaining complexity.

    So to summarize the point, the principles of being intelligent it’s ability to “survive” to “live”. The process, or the algorithm of being intelligent is derived from the process of evolution: the need to be ACTIVELY Involved in the environment to satisfy those principles to live.

    Not sure I completely follow what you’re saying here due to the brevity of the note. It’s hard to be certain. But I think you should look up the concept of Autopoiesis — it’s similar to what you’re describing here, and a very interesting theory.

    So my question is, can computers be more intelligent than it’s creators?

    Depends on what you mean by intelligence ;)
    Modern computer programs can learn. Many programs even excel humans in some activities such as Checkers, or flying fighter airplanes.

    But to answer the real question, on whether computers can become as intelligent or more intelligent than human beings:

    We evolved. This entails we are merely biological machines. Machines can be replicated. So granted that we can fully understand the mechanisms of our own intelligences, which we are getting closer to every day, then there’s no reason to think that we can’t create computers that are smarter than we are.

  7. I kind of disagree with your theroy of computers being smarter than humans at checkers and flying fighters :P

    These systems are programed to follow a set of defined actions, either learned or programed. usually these ai will choose the best suitable action that will produce the most favorable outcome based on its learning curve.

    i believe the arechetecture of a ai system should use a network style datastorge. these data are linked based on sequences that are put together by “eqisodic memory”, as well as action – reaction – procedural memory, or similirity categorization links. the workable memory (surface memory) will be limited due to the system archetecture.

    i will continue this when i get back :)

  8. Hi! :D

    Living Article? Huh… Why not “Evolving Article”? I guess it makes much more sense, since it has to do with a seminal idea upon artificial creativity, that is, evolution! :)

    By the way, I guess I could contribute with the following work on evolving two-phase flashing nozzles through evolution strategies:

    http://evonet.lri.fr/CIRCUS2/node.php?node=72

    Regards,

    Marcelo

  9. @Yang Yu (and others):
    It seems your feeling the way the intelligent machine works makes if it’s intelligent or not. However, I disagree. I think it’s the actual behaviour that is being displayed/exposed that makes it intelligent; or intelligent in our human eyes. Maybe a highly intelligent alien race likes to think of us an we do with ants: not so bright at all! Indeed this agrees with you saying there are levels of intelligence.

    In creativity, what it is about here, I believe it’s a vital part of intelligence (and not the other way around). If we like to build artificial creativity, we should build a program/machine that exposes the same creativity as a human does (to build human creativity, that is). No matter how this is implemented (in a computer, by what algorithms, etc.) doesn’t matter. So even if we have a classical algorithm that plays chess more creatively than a human – we have build human-level chess-creativity.

  10. @Marcelo:
    I don’t expect evolutionairy algoritms or genetic algoritms to be the artificial creativity. Rather, those are methods to build AC programs exposing creative behaviour.

  11. Hrafn

    @Marcelo
    Well, it crossed my mind to call it an Evolving Article. But then, the fact that it’s living implies that it’s evolving (and that you better keep an eye on it or it’ll eat you. Much more engaging) ;)

    Thanks for the link!

    @Yang
    I kind of disagree with your theroy of computers being smarter than humans at checkers and flying fighters

    I agree with Eric that it doesn’t matter how it’s implemented. It makes no sense to say that AI systems aren’t better than humans at something if they actually are :)

    Don’t get me wrong, the perceived intelligence level can vary if we’re talking about things such as adaptability, scope or flexibility. That just wasn’t what I was talking about.

    @Eric
    I believe it’s a vital part of intelligence (and not the other way around)

    I think they’re two sides of the same coin.

  12. Hrafn,

    Your blog is awesome. I have always been fascinated with artificial creativity as well. A group at the University of Central Florida in the US has been working on a collaborative art project based on evolutionary computation and neural networks. Here is our link: Picbreeder
    One of the fun things people can do with these images is use them as a canvas and then color them. I have a couple of examples in my site.
    Let me know what you think : )

  13. Hrafn

    Hey Adelein and thanks for the compliments.

    Picbreeder looks good! Frankly, I’m surprised I haven’t come across it before, thanks for letting me know! Some of the top rated pictures you have are awesome. I’ll have to take a closer look when things slow down at school.

Reply to “An Overview of Artificial Creativity”

Please read the Terms of Use before commenting!

Basic HTML allowed (a, blockquote, strong, em)


Other ..

Think Artificial is a proud member of the 9rules blog community.