On Wed, Sep 11, 2013 at 12:55 AM, Russell Standish
<li...@hpcoders.com.au> wrote:
> Should've given this thread a new name yonks ago
>
> On Tue, Sep 10, 2013 at 11:42:32AM +0200, Telmo Menezes wrote:
>> On Tue, Sep 10, 2013 at 1:52 AM, Russell Standish <li...@hpcoders.com.au>
>> wrote:
>> > I suspect the idea is wrong, because it fails to explain the
>> > exponential growth of diversity, seemingly observed by
>> > Palaeontologists such as Michael Benton:
>> >
>> > @Article{Benton01,
>> > author = {Michael J. Benton},
>> > title = {Biodiversity on Land and in the Sea},
>> > journal = {Geological Journal},
>> > year = 2001,
>> > volume = 36,
>> > pages = {211--230}
>> > }
>>
>> Ok, but I guess that depends on how we measure diversity, which is not
>> a trivial matter. From a quick look at this paper, it seems to focus
>> on the number of biological orders/families/genus.
>
> IIRC, it is argued that the fossil data is too poor to do this at the
> species level, and at the phylum level, it consists of a step function
> at the Cambrian explosion.
Ok, but my issue is more with the number of
species/orders/families/genus being an good approximation of actual
complexity. For example, does the fact that there are several species
of big cats represent a significant increase in the complexity of the
earth's ecosystem? Or are these minor variations on the same theme?
>> Suppose we were
>> able to estimate the Kolmogorov complexity of the entire ecosystem, do
>> you figure it would also grow exponentially?
>>
>
> Yes, most likely. In artificial ecosystems, where we have some
> possibility of measuring K complexity, the value is strongly
> correlated with diversity. However, we don't have a good artificial
> evolutionary system exhibiting exponential diversity growth to be sure.
Right. But to have a fair comparison with the biological situation, we
would have to develop an ALife system and then let
botanists/zoologists/... study them as if they were natural systems.
I'm not saying I disagree with you, just that I'm not sure we can be
sure.
>> >> > What is not true is that human beings are more "adapted" than bacteria.
>> >> > That
>> >> > is not true. Because there is no objective and absolute measure of
>> >> > adaptation. It ever depends on the concrete environment, and varies a
>> >> > lot.
>> >>
>> >> Humm... I think ecologists are able to estimate the likelihood of a
>> >> species going extinct. I'd argue that this could be taken as a measure
>> >> of adaption.
>> >>
>> >
>> > That measure is called persistence, and no, it is not really related to
>> > adaption. For an adaption measure, one good possibility is Mark
>> > Bedau's "cumulative evolutionary activity"
>> >
>> > @InProceedings{Bedau-etal98,
>> > author = {Mark A. Bedau and Emile Snyder and Norman H. Packard},
>> > title = {A Classification of Long-Term Evolutionary Dynamics},
>> > crossref = {ALifeVI},
>> > pages={228--237}
>> > }
>>
>> I read this paper some years ago, it's a very nice one.
>> I would say that cumulative evolutionary activity is a metric that
>> applies to the entire evolutionary system as a whole. The article
>> makes it depressingly clear the Holland's Echo does not match the
>> unbounded evolution dynamics found in the fossil record. But maybe I'm
>> missing something.
>>
>
> I'm not entirely sure why Echo is being held up as the great white
> hope :). But, IMHO, it is more significant that Tierra, likewise, does
> not exhibit open-ended evolution.
I really like Tierra. It's one of those ideas I wish I had.
The reason why I take Echo more seriously is maybe it's theoretical
purity. Firstly, it arrises from some deep thought on the possible
abstract building blocks of complex adaptive systems. Secondly, it was
designed to be based as straightforwardly as possible on these same
building blocks. So when it doesn't work we can go back, inspect the
theory and try to figure out the missing ingredient. I feel that
Tierra takes the machine code concept from the van Neumann
computational model and is forced to contort it to achieve the goal of
creating a complex adaptive system. One example is how it has to deal
with brittleness -- the vast majority of mutation on regular machine
code lead to unviable individuals. But again, both models are probably
quite equivalent.
> And even with Geb (Alistair
> Channon's system), I have expressed my doubts about what is being claimed.
I wasn't aware of Geb, thanks for that. I am skeptical that the
fundamental problem can be solved by changing representations (neural
networks instead of machine code or whatever). It seems to me you
might agree. In any case, I have to study his work in more detail.
>> In the previous discussion I was arguing that persistence could be
>> intuitively taken as a fitness measure of some specific population or
>> species, and I still feel that's the case. If you want to estimate the
>> biological fitness of an individual, you could determine an analogous
>> probability of the individual producing x viable offsprings before
>> dying.
>>
>> I think.
>
>
> Yes - that is commonly done for fitness. In fact, usually it is the
> number of grandchildren that is used. But fitness is not a measure
> of adaption, as I understand the term to mean.
I agree. I would even say that fitness is just a useful fiction.
> Bedau's idea of
> comparing that figure against a neutrally shuffled model (ie the
> amount of fitness over above what you'd expect from chance) seems more
> like what adaption should represent.
Agreed.
Telmo.
>>
>> Telmo.
>>
>> >
>> > --
>> >
>> > ----------------------------------------------------------------------------
>> > Prof Russell Standish Phone 0425 253119 (mobile)
>> > Principal, High Performance Coders
>> > Visiting Professor of Mathematics hpco...@hpcoders.com.au
>> > University of New South Wales http://www.hpcoders.com.au
>> > ----------------------------------------------------------------------------
>> >
>> > --
>> > You received this message because you are subscribed to the Google Groups
>> > "Everything List" group.
>> > To unsubscribe from this group and stop receiving emails from it, send an
>> > email to everything-list+unsubscr...@googlegroups.com.
>> > To post to this group, send email to everything-list@googlegroups.com.
>> > Visit this group at http://groups.google.com/group/everything-list.
>> > For more options, visit https://groups.google.com/groups/opt_out.
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "Everything List" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to everything-list+unsubscr...@googlegroups.com.
>> To post to this group, send email to everything-list@googlegroups.com.
>> Visit this group at http://groups.google.com/group/everything-list.
>> For more options, visit https://groups.google.com/groups/opt_out.
>
> --
>
> ----------------------------------------------------------------------------
> Prof Russell Standish Phone 0425 253119 (mobile)
> Principal, High Performance Coders
> Visiting Professor of Mathematics hpco...@hpcoders.com.au
> University of New South Wales http://www.hpcoders.com.au
> ----------------------------------------------------------------------------
>
> --
> You received this message because you are subscribed to the Google Groups
> "Everything List" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to everything-list+unsubscr...@googlegroups.com.
> To post to this group, send email to everything-list@googlegroups.com.
> Visit this group at http://groups.google.com/group/everything-list.
> For more options, visit https://groups.google.com/groups/opt_out.
--
You received this message because you are subscribed to the Google Groups
"Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to everything-list+unsubscr...@googlegroups.com.
To post to this group, send email to everything-list@googlegroups.com.
Visit this group at http://groups.google.com/group/everything-list.
For more options, visit https://groups.google.com/groups/opt_out.
