h+ Magazine

What’s holding artificial life back from open-ended evolution?

Viewing 5 posts - 1 through 5 (of 5 total)
  • Author
    Posts
  • #29153
    Peter
    Member

    Evolutionary artificial life systems have demonstrated many exciting behaviors. However, there is a general consensus that these systems are missing some element of the consistent evolutionary innovation that we see in nature.

    [See the full post at: What’s holding artificial life back from open-ended evolution?]

    #29180
    Peter
    Participant

    I would like to point out two things.

    First in evolutionary algorithms population usually have very little
    influence on environment if any at all. While in nature such influence is
    very strong even in simple cases.
    Most extreme example of this are of course humans as high life standard in
    first world countries basically “disabled” natural selection in those regions.

    The second thing is that in nature evolution also reaches dead ends,
    there are not as solid as in computer algorithms however that can be
    due to number of variables that natural evolution deals with, that was
    even briefly mentioned in this text.

    #29182
    Marcos
    Participant

    Answer this:

    1) How the definition of complexity limits the individual’s potential of achieving it? Why it isn’t there from the start and what the individual must do in order to achieve it? Not to limit this to “Genetic” Algorithms (since you already noted its dire limitations) nor sound too much like Agent Smith but, “‘we’ may lack the proper language to describe it”.

    2) How the dichotomy between individual and environment limits the shifts you describe in #5? Perhaps this may even solve the individuality conundrum you pose.

    #29192
    Peter
    Participant

    @ Marcos
    As it seems that article author is not responding to this allow me
    to address first issue that you have risen.

    In posted abstract complexity was defined as

    amount of environmental information that they (organisms) can incorporate into their genomes

    In this case when looking from point of individual, like you desired in
    your question, there is given complexity from the start (conception) and
    it will remain unchanged* till the end (death).

    You see genetics don’t care to much about individual lives and changes,
    including rise of complexity level, happen during generation change and
    are usually analyzed from the point of population.

    However if we won’t limit ourselves to genetic level there is some hope
    for individuals :).
    We can now define complexity as

    amount of environmental information that organisms can incorporate and
    pass to others

    This definition means that we have both genetic and cognitive information as
    in case of humans (and most animals with larger brains) learned experience
    can be taught to others.

    So you can rise you complexity by learning from experience and from others,
    this ability of rising one complexity within single lifetime was serious
    evolutionary advantage as you can see that most animals you will encounter
    have some learning process besides instinct inherited via genes.

    Well that was long post, I hope you found it at least bit interesting 🙂

    *It’s not really true as there are mutations and other genomic operators that
    change your chromosomes, however from point of our definition of complexity
    they don’t count.

    #29410
    Marcos
    Participant

    @ Peter

    Hi, sorry for the delay. It seems the email notification is not working anymore here either =(

    Yes that was very interesting, thank you. I’m pretty sure if those models accounted for what you say, the ‘individual’ complexity there would soar way higher.

    Another way to approach the issue I’ve risen is to literally blur the lines between the concepts of “individual” vis a vis “environment”.

    For example, all organisms incorporate food (however one defines it, e.g. photons) and assimilate it into themselves. How that influences evolution itself? What about an even more literal interpretation of this, perhaps so literal as to make it hide in plain view, say tool-making? Arguably what first differentiated higher primates from the rest.

    Much like your example, this is hardly hard-codable into inheritable languages. Maybe when you say “genetics don’t care” is a hint that it may be impossible to care (maybe evolution would have figured it out by now), at least in a direct (“hard-coded”) way. Thus brains are needed. 🙂

Viewing 5 posts - 1 through 5 (of 5 total)
  • You must be logged in to reply to this topic.