Is Digestion necessary for the Evolution of Predation? | Digital Life

digestion may seem like a pretty boring subject after all we barely even think about this process that permeates all life but as i have recently discovered digestion is an incredibly complex dynamic that plays a far greater role in guiding and influencing the evolution of life than most people realize i am working on an artificial life simulation and i've always wanted to see predation evolve in the ecosystems i'm creating after all it's an incredibly powerful dynamic that brings a lot of complexity and dynamism to an ecosystem and digestion might just be one of the keys to its evolution if you're new to the channel i would highly recommend watching my previous video where i introduced the basic concepts and the difficult task that is ahead of me for the others welcome to the fascinating world of the bibitz [Music] like we saw last time this quest for enabling the evolution of predation is going to be hard not only do we have the difficult task of making predation a viable ecological niche we also have to make the transition to predation possible last time i tried engineering my own species to see if a stable predator prey relation was even possible and we saw that it wasn't this was due to the fact that air before had it too easy and would always evolve after some time to become too hard to catch and kill thanks to the help of the community i understood that herbivores were too good and the solution to give them a harder time was introducing materials and digestion to the simulation initially the babits instantly received the energy of the food they ate as soon as food would enter their mouth the pellets would be converted to energy with an efficiency that depends on the diet gene of the individual to my knowledge digestion was just not an important dynamic to implement the two food types also had the same efficiency curve just because i didn't think it would have much impact what was i even thinking on earth many dynamics and adaptations can be linked to an organism's affinity to a certain kind of food plants are far less energy dense and are usually harder to digest than other forms as such herbivores tend to be slower use more of their time to eat and need to eat a bigger overall mass of feed every day to sustain themselves some animals like the koala specialized heavily to eat eucalyptus sleeves almost exclusively their brain evolved to be as small as possible to maintain basic functions and to use the less amount of energy possible their microbiome is essential to their ability to digest an interesting and potentially disturbing fact is that in order to obtain their initial microbiomes koala newborns need to ingest the fecal secretions of their mothers as their first meal while it's a very interesting example of just how hard being a herbivore can be i don't think i want to go that far on the other end carnivores profit from a food source that is usually far denser in energy and that is proportionally far easier to digest some specialized carnivores can go many weeks between feedings it now seems clear to me that having different statistics across different types of food is the way to go in order to do that i first needed to simulate materials different materials will have different characteristics that will influence how good or bad of a food source it will be like said previously i want the plants to be far less energy dense and be harder to digest but softer and easier to bite off they will also require bibits to be better adapted as the conversion efficiency will very rapidly decrease as the bibit's genetic affinity toward plants decreases potentially reaching the negatives meat should be far more energy dense and be processed easier by the stomach it should also be more forgiving as the efficiency curve would allow bibits with the low genetic infinity toward meat to still be able to process it to an extent now instead of just counting the energy in something and ending that value like i was doing i can track and manage the amounts directly if i have two pellets of the same size one made of plant and the other made of meat they will occupy the same space but have different mass contain different amount of energy and so on this system of material simulation also opens a lot of doors for the future i already see a few more materials that could be added like rocks actual blood fat and so on i would just have to define the characteristics of those new materials and they would integrate nicely with the rest of the systems but i think that their dieting system as well as their vision system will have to be updated before making that a reality and now it's digestion time oh yeah it's all coming together first a bibit needs to swallow something the babits have control over how open their mouth is which lets them control what size of objects will be able to enter their stomach if the object is too big they can attack it to try and bite off a chunk depending on the hardness of the material and if they are strong enough but of course coding is never straightforward especially in artificial life first i add infinitely splitting food which the vivid seemed to have evolved to use as form of locomotion another funny side effect was that instead of an artificial life simulation i now had a functioning nuclear reaction simulation after fixing what i thought was the source of the bug i was left with infinitely splitting pellets then somehow some of them learned to fill their entire stomach from even the tiniest of bites of course we couldn't avoid the classy glitchy physics but hey with hard work and enough time we can always patch things up their stomach can only contain a realistic amount of matter which is proportional to the pivot size smaller bivets have smaller stomachs it just works then they have a neuron that allows them to control the acid level in their stomach which controls the total digestion potential the matter in their stomach is slowly digested depending on the acid level having less acid than the amount of food makes digestion slower but having too much acid compared to the amount of stuff while increasing the speed of digestion decreases the efficiency then each material depending on its respective properties will be digested at a given rate and finally debit depending on its diet affinities will either gain or lose energy proportionally this would mean that a carnivore trying to digest plants would lose energy in doing so considering that plants are less reactive and thus digested slower this could easily be overshadowed by a small amount of meat also being digested at the same time it's all a matter of proportions in fact i've already seen a few carnivores evolved to also swallow plants in order to slow down the digestion of meat in their stomachs and let it accumulate instead of digesting it right away which is pretty interesting also bigger bivets have a bigger stomach and will be able to digest more matter in parallel as a result every bibits will run on what we'll call an energy budget depending on their food of choice their affinity toward it their size their metabolism speed and other factors they'll have a certain amount of energy per second that they can use assuming they're always full if they use less than this limit they'll accumulate the surplus in their energy reserves up to a certain amount if they use more they'll tap into their energy reserves and if it's empty their health potentially even dying of exhaustion as a result depending on their size metabolism and food affinities they'll have different lifestyles that are more viable another thing to consider and that i'll need to keep on balancing in the future is the efficiency curve of the materials basically if you want a viable evolutionary path to carnivory omnivory and scavenging seem to be intermediaries of choice so the efficiency curves for both plant and meat must be adjusted so that an omnivore diet is also viable just like many of you commented in the last video if we go back to the simulation i presented in the last video the herbivores managed to out-compete their predators over time what we wanted to create with this update was basically to bar off a certain portion of the genetic domain to arbiverse leaving a safe haven for predators that will have an easier time catching preys that run on a tighter energy budget now with tests i see that this is indeed the case herbivores now have a limit after which they can't keep getting better they do evolve in response to being chased and do get better at avoiding their predators but only to a certain extent additionally i followed on the suggestion many of you made and added a more robust aging system that ensures that the pivots that are becoming too old start to lose strength speed and viability they'll make very good praise for their potential predators i still haven't managed to make the predator niche completely viable but this seems much better than before now they end up eventually dying from other sources than their praise being too good sometimes it's just that they keep eating all their own babies or that they eat so many herbivores that they drive themselves to extinction i'll call that progress so what's next it's abdominally clear what you all want i did a quick calculation and approximately a third of the comments on the last video pointed to plant evolution i absolutely agree that this is something that needs to be done at some point in fact i already had it in my roadmap for some time but i admit that it makes so much sense in the context of nerfing herbivores and encouraging multiple niches to form i also think that it will make the simulation a lot more dynamic and interesting however a lot needs to be done before that's possible as an example for plant evolution to be interesting it would be nice to have environmental variation that would influence how different types of plants would grow across the map and for that to make sense we need bigger simulations that allow for more variation also babits would have to be able to differentiate between the different kinds of plants and different types of plant parts which they are not yet capable of doing because their vision system is extremely lacking in order to give them a new vision system as well as allow plants to evolve too i first need to finally implement a new modular evolutive algorithm that i've been designing and working on for a long time but always put off implementing in the bibits because of the amount of work that it will require well if implementing plant evolution is a new prior already i guess it's finally time to implement the biome algorithm in debits basically i want to make it so that their bodies genes and capacities can evolve more freely compared to now at the moment the bipeds have a fixed amount of genes brain inputs and brain outputs in fact a bibit as they are at the moment will never be able to evolve new body ports new sensors or new capacities this is what i want to address and revolutionize with the biome algorithm i am also planning on implementing this new algorithm through dots and ecs a new framework by unity that is extremely promising in terms of performance and efficiency which will allow us to run far bigger simulations with more potential to then implement more features like environmental variations more complex systems and much more obviously this will not be easy and will probably take some time however in the meantime i'll try making other types of videos related to the project that will keep you distracted i'll show you interesting simulations people manage to evolve maybe some educational videos about life and evolution using the bibits as a tool and other stuff like that we'll see in the meantime i've uploaded the most recent and stable version of the simulation to itch.io where you can download it and play it for free i would encourage you to play with the parameters and try your end at making an environment where predators can exist if you join the subreddit i'm sure i and others will be able to help you out on where to start there is also the wiki which is pretty bare bone at the moment but quickly growing and that offers a lot of insights and information that might not be obvious since my last video many have asked me how they can contribute to the project i decided to write a post about that and build a trader board with many very tasks that you can take on in order to help out it has anything from meme posting to giving me video ids i'll leave the link in the description for the interested as always it might not seem like much but subscribing is one of the best thing you can do to support the channel it really helps me see that what i'm doing is appreciated it helps you see the video as soon as they come out and it lets the youtube algorithm know that it should show that stuff to more people i'll call that a win-win-win if you want to go the extra mile in supporting the project patreon is the best way to do so at all tiers you'll get access to the discord server where i and other patreons hang out and for most tiers you'll get access to the alpha releases of the project allowing you to experience the new features before they are ready for major releases and the fun side reward is that you'll see your name appear in the end of video simulation that's coming shortly so again thank you for supporting the project and following its development i really hope to see you all next time [Music] you

Channel: The Bibites: Digital Life

Views: 228,385

Keywords: AI, Evolution, Genetic Algorithm, Ecosystem, Simulation, Artificial Life, ALife, Neural Networks, Algorithms, Natural Selection, C#, Unity, NEAT, gamedev, game development, Life, Digital Life, Science, Biology, Genetics, devlog, digestion, materials

Id: kNWb7e8FZDo

Length: 19min 30sec (1170 seconds)

Published: Wed Mar 16 2022