ID:154193
 
Earlier I said I am going to work on a serious AI library. I decided to make a post on my thoughts and breakthroughs.

A.
I began by sitting down in class and thinking about the nature of animal actions.
All animal actions can be broken down into 3 functions
Consuming
Self Repair (sleeping)
Reproduction
So the prime directive is to live long enough to pass on its genetic code.
B.
As my first test I created a bug. That after 50 turns it would then lay its eggs. But before then it had to find a food source, eat when hungary and rest occationally. So on So on. What I found out is that Natural Selection just happens. The ones that could find the food first were most likely to lay their eggs, thus complete their prime directive.

Next step:
Random Mutations :)


Thats all for now. I shall keep everyone updated.

I'm doing my Masters in CS in the area of basic AI (primarily self-learning knowledge bases). You might want to check into "Genetic Algorithms" for your mutations - good stuff if applied right...
In response to digitalmouse
Ya I think I will.

People may think why are you working with breeding and genetics when you are working with game AI.

In most RPGs the creatures goal is to kill the player. But why?
Cause your human and its not?

When you have creatures who actually have self proclaimed territory it gets intresting.

Here is my dream:
Some players go out and kill this ape creature for good exp. The ape population starts to seriously decrease. The apes are semi intellegent so they decide not to take this anymore. They get a rather large raiding party together and pillage the nearest town.

or

The apes get tired of getting killed so they get together and hide in the trees. There is one ape that can be seen. The normal human party comes up to hunt the ape. Suddenly 30 apes drop down surrounding the humans.
Winbiko wrote:
I began by sitting down in class and thinking about the nature of animal actions.
All animal actions can be broken down into 3 functions
Consuming
Self Repair (sleeping)
Reproduction
So the prime directive is to live long enough to pass on its genetic code.

This seems a bit oversimplified to me; you're missing a few that explain other important animal actions. (Indeed, reproduction is probably the least important thing animals do.)

Behaviors common to most animals and insects:

Foraging: The animal (or herd) searches for food or prey.
Nesting: Animals may seek their own kind to rest with or bring back food. Usually they'll create a habitat for themselves: A den, burrow, lair, hive.... (Sleep is an associated behavior, so I don't count it separately.)
Defense: An animal will be on guard against predators. In communities (like anthills), it's not uncommon for a small group to act as watchers or scouts. In a herd, all the animals are generally alert for danger but act as a unit when attacked.

Note that these behaviors generally break down into subcategories. Foraging would entail not just finding food, but how and where it's eaten. Nesting is more than just making a nest or picking a place to rest; it also includes sleep, sometimes breeding (or caring for a brood), and other nest-related behaviors. Defense would refer to who stands watch and how, but also to the nature of the defensive action taken once danger is sighted.

Lummox JR
In response to Lummox JR
Foraging: The animal (or herd) searches for food or prey.
(this is part of the consuming process)

Nesting: Animals may seek their own kind to rest with or bring back food. Usually they'll create a habitat for themselves: A den, burrow, lair, hive.... (Sleep is an associated behavior, so I don't count it separately.)
(Protected self repair. If a group of animals is hunted by another group, they may began to stay together.)

Defense: An animal will be on guard against predators. In communities (like anthills), it's not uncommon for a small group to act as watchers or scouts. In a herd, all the animals are generally alert for danger but act as a unit when attacked.
(Reproducing, proctecting the queen so she can breed)

Like I said everything can be broken down to those 3.

But an animals ULTIMATE goal is to pass on its own genetic code.
In response to Winbiko
Winbiko wrote:
Here is my dream:
Some players go out and kill this ape creature for good exp. The ape population starts to seriously decrease. The apes are semi intellegent so they decide not to take this anymore. They get a rather large raiding party together and pillage the nearest town.

The apes get tired of getting killed so they get together and hide in the trees. There is one ape that can be seen. The normal human party comes up to hunt the ape. Suddenly 30 apes drop down surrounding the humans.

-drools-

If you make a game with these systems, you are my god.

=V
In response to Vortezz
Why don't you look into the subject and try to tinker with it a little bit?¿ Then you could do it yourself.
In response to Loduwijk
Loduwijk wrote:
Why don't you look into the subject and try to tinker with it a little bit?¿ Then you could do it yourself.

hmmmm hmmm hmm hmm hmm hmmm hmmm hmmm mhmm mmhmmmmm
Memory would need to be kept of way more than billions of GBs. The thinking of the animal can be designed to react to the slightest little things. But always and still always there only things we see.
In response to Green Lime
Bull. The memory requirements might not be bad at all. You are thinking that it would be necessary to record every experience. Thats not the way it works. you only need to discourage some behavoirs and encourage others.

This can be a simple as preventing a certain walk sequence from happening. If the mob wont go near a certain object, whole categories of behavior can be eliminated.
In response to Lummox JR
Lummox JR wrote:
(Indeed, reproduction is probably the least important thing animals do.)

I beg to differ... Reproduction is the most important thing living things do...

In fact...I'll go so far as to say that it's the only thing a living thing truly exists to do...

Everything else in a living thing's life is only done in support of reproduction...

Sure, for the being's life itself, reproduction is very unimportant... But the being's life's purpose is to reproduce... Plain and simple... The life of the creature (including plants) is only there so it can reproduce...

And so on and so on...

I suppose it's a rather sad way of looking at things... (Especially to the religiously minded) Since it means that life is essentially a meaningless loop that only exists to restart itself for the next iteration... But I feel that it's the truth...

We exist to reproduce...




In response to Green Lime
Green Lime wrote:
Memory would need to be kept of way more than billions of GBs...

Not necessarily - if you use a combination of genetic algorithms and software neural networks (both available in C/C++ (basic GAs available in VB too), and probably could be converted to DM-code with a little bit of serious effort), memory requirements go down considerably because you are essentially looping through GA and NN code to 'learn' certain 'behaviors' - better behaviors will have better 'weights' to them (the higher the better for example).

You would then just need to remember behavior patterns associated with certain action/reaction conditions - you should be able to break that down into a few hundred combos.

The mob can then act fairly intellegent through a combination of environment conditions (being eaten, having sex, etc.), and using combinations of these 'act/react combos' you could get fairly complex animal behavior within a few MBs of memory (for storage, not the actual learning algorithms themselves)...

I recently remember seeing some Quake2 bots programmed with custom GA/NN code under 2 MB. It used a "brain.ini" file to store weighted behaviors (how to move, how to shoot, who to shoot, etc.). After a few thousand iterations (about two hours) - these could be called 'generations' depending on who you talk to; in this case I use 'iterations' because any positive genetic code was merged back into the parent; sort of a genetic incest I guess - the bot would start to 'learn' that shooting me was more fun than shooting at the walls or ceiling! :( :)
In response to SuperSaiyanGokuX
Been reading much Freud lately?

Our existence is a huge collection of "meaningless" loops. Everything is a cycle, from the circle of life, to our planet revolving around the sun. Popular theory even suggests that time and the universe itself are repeating cycles.

The lack of meaning is an invention of the human mind. Life does not need meaning to happen. "Meaning" requires a thinking observer and is open to the interpretation of that observer. What is the meaning of this bottle of Member's Mark Natural Spring Water I am drinking? Every person will have their own interpretation, and likely no one will give metaphysical answers.

What about people who choose not to have children? Are they somehow "unnatural"? (I guess that would mean they are supernatural. ;) ) Do creatures that can not breed have no purpose in life? Do mules have any less "meaning" than donkeys?

The meaning of life is a debate for philosophers and has been going on since prehistory. The bottom line is: as thinking beings, if we seek a meaning for our life, all we must do is define it for ourselves and work for it.

In response to Winbiko
Winbiko wrote:
Like I said everything can be broken down to those 3.

Uhhh... but those weren't the one you listed. I did work your thoughts into it, but otherwise I chose a much different categorization.

But an animals ULTIMATE goal is to pass on its own genetic code.

Animals don't have goals except immediate ones, and reproduction ranks pretty low on the list. And nothing, even humans, thinks in terms of passing on its genetic code. When people think about reproduction they think about family, progeny, about raising a future generation, about leaving a part of their personality or some good work behind: This is never cast in terms of raw genetics, unless the particular individual happens to be an ultra-psychotic eugenics fanatic.

I think you're overdarwinizing the issue. Remember that natural selection only works in the case of large populations--it's a statistical process, basically, in which outliers who happen to better fit a niche have a slightly improved chance of breeding and, over time, shifting the average.

In contrast, the functions of an individual life have a lot more to do with sustaining that life. Thus an elderly animal who's past its breeding life will still do most of the same things other animals do. You also can't overlook the fact that although hive insects have many of the same behaviors, most of their population never breeds. (Most hive insects are non-breeding female workers.) So clearly, all things considered, reproduction is the least important biological goal.

Lummox JR
In response to Shadowdarke
I would say that the real perpose in a Humans life is knolage, posestion(food,land), and last(to me) booty(lol).
1.Knolage=It is one thing we have been going after all our lives. If it wornt for the search for knolage we would be still in asia picking insects off out mates back.

2.Posetion=My History teacher claims"Everything we have, are, and will do is caused by hunger. I dont agree 100% but at least 50%. The first humans(for less religous) came from asia but there where Humans on the US turf before old Chris Colmbus. They came from asia in search of food. And for you Religios types, why do you think Eve ate the apple? Because she was hungre and it was the best food around and she knew it. But she also knew it had a price, but she didnt care, she ate it anyways because she was hunger. And most animal and gangs live life to defence turf. Most country are mad when getting invaded by other countys, not talking about France(lol). And im sure you dont like it when a sales man forces his way into your house.

3.Booty=Almost everyone has kids. This is the way for imortality. Imagin having someone to take your name and make it famuse. And when i say booty i dont mean for fun but for repurduction resons. How many out there over 21 and looking have not got you some yet?

see.
In response to Lummox JR
bu in a hive or something the breeding is being done, just not by everyone because others have to care for the young. The other worked just take care of the larvi.So they are doing there part in the reperductive area.Look at packs of apes, there is one Alfa male that most the females are impregnated by. Making sure the best geinse are passed on. The males whole young life is training to be the alfa male so they can give there traits to the babys. They get food and stuff to stay alive long enough to reproduse, look at the preying mantuse, the males are killed right after impregnating the female and they know it but dont mind.Why, because in a way they will be back in there first and only born son or daughter.
In response to Lummox JR
Lummox JR wrote:

But an animals ULTIMATE goal is to pass on its own genetic code.

Animals don't have goals except immediate ones, and reproduction ranks pretty low on the list. And nothing, even humans, thinks in terms of passing on its genetic code. When people think about reproduction they think about family, progeny, about raising a future generation, about leaving a part of their personality or some good work behind: This is never cast in terms of raw genetics, unless the particular individual happens to be an ultra-psychotic eugenics fanatic.

True, creatures don't think of reproduction as passing on their genetic code, but that doesn't mean that's not why they are reproducing. Let's say I'm hungry and I see a turkey sandwhich and a pile of twigs on my table. I will choose to eat the turkey sandwhich rather than the twigs. The descision would be made based on the better taste the turkey sandwhich has. The fact that the turkey sandwhich has much needed proteins and carbohydrates and the twig doesn't (or if it does, the enzymes inside me can't break down the twig into proteins and carbohydrates) probably didn't factor into my descision at all. However, the nutritional value of the turkey sandwhich did factor into my genetic makeup. Because years of evolution have taught my body that it needs proteins and carbohydrates my body gives me a pleasurable feeling when I eat a protein or carbohydrate (or sugar, etc. for that matter) (the pleasure being the good taste of a turkey sandwhich). The less nutritional twig gives me an unpleasent sensation (bitterness) when I eat it, so I don't eat it. My thinking was based on good and bad taste, but good and bad taste were based on what is good and bad for me. Although creatures don't reproduce to pass on genetic code, that (in my opinion at least, this is all debatable) is why whatever made them reproduce (want for family, progeny, something to leave behind) was viewed by them as positive.

I think you're overdarwinizing the issue. Remember that natural selection only works in the case of large populations--it's a statistical process, basically, in which outliers who happen to better fit a niche have a slightly improved chance of breeding and, over time, shifting the average.

I don't think that this is being "overdarwinized". natural selection is the idea that creatures better suited for survival survive and breed more than those who aren't, slowly changing the makeup of a population as the inheritors of the genes of the survivors grow in number until the traits of the survivors are basically exclusive to the population. This can also be applied to reproduction itself. Those who are more likely to reproduce (for whatever reason they choose to reproduce) breed more than those who are less likely to reproduce (for whatever reason they choose not to reproduce). Therefore, there will be more of the reproducers heirs. For example, lets pretend that the desire to leave a part of oneself behind is a genetic trait. (I imagine it isn't, but this is just an example. If you want you can replace the desire to leave oneself behind with some other trait more likely to be genetic) The people with this "gene" will probably have children and those that don't have this gene probably won't. Slowly those that don't want to leave something behind will be weeded out of the population until only those who do remain. The reasons we and other animals have for having children don't matter all that much. All that matters in the darwinistic sense is that we do want to have children.

In contrast, the functions of an individual life have a lot more to do with sustaining that life. Thus an elderly animal who's past its breeding life will still do most of the same things other animals do. You also can't overlook the fact that although hive insects have many of the same behaviors, most of their population never breeds. (Most hive insects are non-breeding female workers.) So clearly, all things considered, reproduction is the least important biological goal.

True, many creatures who'll never breed or are past the age of breeding live. However, there lives are often at least somewhat important to the survival of those who do breed. The non-breeding female workers in most hive insects are responsible for feeding the breeders and caring for the young. This way the young are born and mature into adult hood so that some number of them can become breeders and reproduce themselves one day.

If reproduction weren't important to each and every specie, then they wouldn't reproduce in large enough quantities to survive for however long they've been around. When something is unimportant it doesn't happen, or happens on a small scale. If reproduction didn't happen on a huge scale then the population of an area would dwindle and dwindle until nothing lived there anymore. Instead life blossoms and spreads to all corners of the globe. There are humans living on every continent (in large numbers on 6 of the 7). If we as a specie don't care about reproduction, why do we do it so much?
In response to Lummox JR
I think the confusion here is that Winibiko and LummoxJR are using one term to mean two different things. Win seems to be taking a macroscopic, species wide view of "goal", Lummox a microlevel, individual view. Both are correct. If an individual organism can be said to have a "goal" it is generally serlf-preservation behaviors. But Winibiko is correct in pointing out that a species "goal" is the propogation of that species. This doesn't mean that every individual consciously decides this, but rather as the product of millions of years of evolution and natural selection, behaviours and drives leading to effective procreation have been reinforce (members not having these would be less likely to have there genes added to the pool). Example: I like to read. But I also like females and food. What I decide to do depends on my mood and plans. But what I desire is also largely determined by my heredity and thus part of a larger picture as well- the general genetic traits of all humans. I may have significant variation as an individual, but, as Lummox pointed out, statistically the species behaves in certain ways.

Make any sense?

-James
In response to Luap
Luap wrote:
If reproduction weren't important to each and every specie, then they wouldn't reproduce in large enough quantities to survive for however long they've been around. When something is unimportant it doesn't happen, or happens on a small scale. If reproduction didn't happen on a huge scale then the population of an area would dwindle and dwindle until nothing lived there anymore. Instead life blossoms and spreads to all corners of the globe. There are humans living on every continent (in large numbers on 6 of the 7). If we as a specie don't care about reproduction, why do we do it so much?

Jmuprph's comment about macro- vs. micromanagement apply. In essense I'm talking about individuals, not collectives. There are a few individual behaviors meant to serve a collective good, especially in hives, but on an individual level breeding takes a low rung on the ladder.

Lummox JR
In response to Jmurph
You hit the nail on the head there. The collective behavior of a colony, hive, flock, group, etc. is vastly different from individual behavior and ultimately follows statistical trends. There are many times that it's desirable to control that behavior.
But, since Winbiko was talking initially about AI coding, that would refer to the individuals within such a group. Thus my categorizations were based on the individual, not the collective, since collective behaviors have little or no bearing on individual behaviors.

I think in both cases it's the goal of both the individual and the collective to survive. A collective survives by establishing new blood, growing, expanding its territory, and occasionally by starting new collectives. It takes a long view, where reproduction has to be a big part of the process because it's the constant renewal of a group's numbers. An individual survives by very different mechanisms, none of which have to do with reproduction. (Reproductive activity as an instinct or an intrinsic good encourages individuals to behave in a manner suited to this aspect of the collective, however.)

What's bizarre is that you can take the collective approach and dial it down to cells: Our cells make up a collective. They divide to replenish dead cells or grow.

Lummox JR
In response to Lummox JR
Lummox JR wrote:
You hit the nail on the head there. The collective behavior of a colony, hive, flock, group, etc. is vastly different from individual behavior and ultimately follows statistical trends. There are many times that it's desirable to control that behavior.

I know you qualify this a bit in your later paragraphs, but I wanted to highlight this and respond by saying that collective/hive behavior can be strikingly similar to individual behavior. Insects in particular demonstrate this well, almost to the point that individual hives/colonies seem to have thier own "personality."

I believe it's similar to the way all of our many thought patterns come together to form a unique personality. The individual patterns interact with, conflict with, and support each other in millions of different ways, the end result being what we perceive to be intelligence. This is why a single ant seems dim-witted, but a colony of ants seems much less so.

All of this is making me want to dive back into my all-time favorite book "Goedel, Escher, Bach," for some more fun reading on similar topics. There is actually an imaginary dialogue in there with a colony of ants consiousness. Quite delightful stuff!
Page: 1 2