Reflect on the Science you teach - identify a topic/phenomenon in your subject area that is a complex adaptive system. Which characteristics of a Complex Adaptive System are met by this phenomenon?
Post your thoughts to the forum below, and come back later to read other entries and comment on another teacher's response.
I'm trying to decide if a neural network counts. On the one hand, each neuron unit is following relatively simple rules based on the activations of the other units feeding into it, and there is definitely some emergent behavior that we don't yet understand very well. On the other hand, during the training phase there is higher level coordination going on, in that the activation functions are explicitly tuned based on feedback from the overall network performance on example data. I don't think any of the examples of complex adaptive systems had a centralized system-wide training process at any point.
Wow, that's interesting. I hadn't thought of the training period as system wide directed and organized before. Let me see if I can get an AI researcher's perspective.
I would agree that neural networks are an example of a complex system because each neuron is following a simple set of rules that leads to an emergent pattern or response.
Also, in terms of my curriculum, I would consider thermal energy transfer to be a complex system because the individual molecules are bouncing off of each other depending on the amount of energy they contain and can transfer. This emergent pattern would be less visible, but it would be evident in the overall transfer of heat. I also think of food coloring dropped into water and you can watch the fluid motion of the colored dye.
Another similar example I can think of in my curriculum is the human impact of plastics and other trash as they flow through the oceans. This could be considered a complex pattern because the materials move in response to the currents and the emergent pattern would be the gyres.
Great examples, especially the trash build up in the oceans causing gyres to form. A high school team modeled that for their Supercomputing Challenge project (see http://www.supercomputingchallenge.org/10-11/finalreports/36.pdf)
What is a topic/phenomenon you teach that is a Complex Adaptive System?
A topic/phenomenon I teach that is a Complex Adaptive System would be a simulated Ecosystem like the example in the video. I am also thinking about the science fair we do each year. This would be a benefit to us because there were a lot of bug "funerals" during their different trials. If I can teach them how to code the variables they can conduct their own experiments using the simulations and we won't have to bother the innocent bugs on campus.
I want to know how much we can actually test with this system (doing simulations)? What barriers are there in testable questions my students may have?
The typical example of feedback we learned in psychology had to do with the production of hormones that induce labor. I'm not entirely sure if this qualifies as a complex adaptive system, but the production of oxytocin stimulates contractions that continue labor, and as it goes on there is more oxytocin and stronger/more frequent contractions. Perhaps the agents would be the oxytocin molecules being released by the body?
Currently I am focusing on Earth Systems, especially geological systems. There are many examples of complex adaptive systems including water cycles, rock cycles, and plate tectonics. Earth materials interact with each other with no central control. Some of the interactions react to temperature, and pressure conditions and the characteristics of each kind of material. As a result patterns such as the water cycle (weather and climate patterns) and the configuration of continents ( the occurrence of earthquakes, volcanoes, mountain building and erosion) emerge from these interactions. Sciences of all kinds are filled with examples of complex adaptive systems and not understanding the nature of systems has lead to much misunderstanding and bad decision making.
You have already discovered one of the advantages of using computers to run simulations -- no bug funerals! The only limits or barriers to testable questions are in coming up with questions that work well in agent-based models (like the ecosystems model) and where useful data can be collected by means of data boxes or graphs. Students think of many things that would never have occurred to me, and are usually able to build the code they need. Let us know if you do come up against a barrier, and we will do what we can to help.
If the contractions are stimulating the oxytocin production, which is stimulating more contractions, that would be a textbook example of reinforcing feedback (where more leads to more - sometimes called "positive feedback"). I'm not sure if it meets other criteria to be a complex adaptive system, but perhaps that's because I don't know enough about it. Interesting example!
You've listed some very good examples -- you will be interested in Module 2 (earth science) and we have some other earth science models that have been developed in our after-school program. And we hope you will develop some other models relating to plate tectonics and other geology issues to add to our network.
Responding to Shaw'sae about barriers that exist to posing answerable (testable) questions that students may have.
Yes, it is good that you are thinking about how to address questions that students might come up with. Some student questions will be testable through simulation and others not. In general, running simulations can answer questions about the model posed like this "what is the impact of changing variable on the modeled system?" or "what are the possible outcomes in this modeled ecosystem?" It is important to note that running simulations tell us primarily about the model - they only tell us about the real world phenomenon if the model is a valid representation of the real world. (We will discuss what it means for a model to be valid later in the course.)
I teach physical and geological sciences to 6th & 7th graders, and [am slightly ashamed to admit that I] have not yet incorporated the concept of CAS in my lessons. I think that thermodynamics and molecular motion driven by thermal or concentration gradients would provide appropriate springboards to discuss CAS.
An example would be the crystallization of salt during the formation of evaporites. The atoms are leaderless (although I did have quite the laugh envisioning a single sodium atom leading the charge complete with sword), the formation of crystal lattice structures constitute an emergent pattern, NaCl crystals form based on atoms following simple rules (self-organizing), the system is adaptive and can demonstrate chaotic behavior in the presences of alternative reactants, temperature and pH fluxuations, water currents, etc). Crystallization is also based on stochasticity.
I am thinking about our recent Chemistry unit (this was my first experience teaching chemistry), would any chemical reaction be considered Complex Adaptive Systems? There are individual parts (atoms) and they behave in specific ways (have simple rules that they follow). They definitely react but I don't feel that they adapt. The reactions are leaderless as well.
One topic that comes to mind is population ecology and how animals/plants interact. Specifically, I am thinking of wolf, rabbit, and grass populations and how the size of one affects the size of the other. This phenomena meets the following characteristics of a Complex Adaptive System: many parts, no leader, and the system adapts to change.
A complex adaptive system I teach would be energy pyramids-how many producers are necessary to provide energy to primary consumers, and how many primary consumers are necessary to provide energy to secondary consumers, and so on. Each of the levels of the pyramids include various species which act in certain ways and provide energy for the next level above, and include interactions with predator and prey and amounts of each level necessary for a healthy ecosystem.
I am a middle school technology teacher. I wanted to introduce coding into my standards and am working closely with our science teachers to make it more meaningful to the students. I am currently working with my 7th Grade MP4 students. As 6th graders, they had worked on the introductory Module 1 and are now working on Module 3, Ecosystems. They just finished studying the ecosystem in Yellowstone and the reintroduction of wolves to balance it. Many of them have created models showing this complex adaptive system. Other students have chosen different ecosystems. We live close to the Atlantic Ocean, so one of my student groups studied the impact of humans on the dolphin population. Another group studied the Lionfish as an invasive species. Students in the past have studied Manatees and human interactions, deer populations, humans poaching on elephants, and fish dying due to lack of oxygen. These are complex adaptive systems because they have no leader, follow simple rules, and are unpredictable.
Like many of the others I think ecology and the interactions of energy through the Earth would be a complex adaptive system. But I also see my chemistry unit in which we are looking at the interactions of atoms (and their electrons) with one another. The students see both the interaction of the individual, looking at how electrons behave within an atom. To the whole how the atom interacts with other atoms in a chemical reaction.
My first thought was about the different spheres of the Earth, biosphere, atmosphere, mesosphere, lithosphere, etc and how they all interact and influence one another while really working separately following the aforementioned rules. I see this also as how the Universe unfolded from the Big Bang til now creating patterns of simple to complex from development of stars and galaxies to the development and evolution of life. Kind of excited right now at all of the possibilities and this idea as the undercurrent.
I'm looking forward to using computer science in my classroom. I was thinking about the layers of the Earth as a complex system: biosphere, atmosphere, hydrosphere, lithosphere. They all adapt and work off of each other.
I believe the life cycle of a star could be an example of CAS. The atoms and molecules in a nebula have no central control, but through their simple interactions, patterns emerge and the system self-organizes into a star. Each atom or molecule follows a simple set of rules, and reacts predictably to changes in ambient conditions (e.g. temperature, gravity, pressure). Due to the scale of their interactions to each other and their environment, simple particles and groups can evolve into complex systems such as stars.
In reply to jsimpson:
I'm new to CS and have not studied CAS, so please excuse if this sounds ridiculous: I think that the adaptivity of the system refers to the sum of the parts and not the individual agents themselves. In the case of chemical reactions, each agent (atom/molecule) follows the set a rules (reactivity, due to electron configuration), and the emergent system adapts to changes (e.g. ambient temperature, pH, salinity, etc).
I'm thinking of a model we played with the students that was able to mimic predator/prey/habitat resources. As one went up the other went down and students could collect data and graph and look at limiting factors in an ecosystem. I believe this would be considered an adaptive system and I think the activity would transfer over nicely into as a computational simulation. Hopefully I'll learn enough with this training to be able to do it.
Currently, I teach Biology as well as Anatomy/Physiology. The first thing that came to mind was the human body itself. After all, a CAS is defined as "a system made up of many individual parts or agents." These agents or parts would be our cells, tissues, organs, organ systems. All of these individual parts then, do follow simple rules themselves using various methods, i.e. hormone regulation, cell signaling. Patterns can be found throughout the entire body - take the pulse for example. Most importantly, if the system is altered in any way, there is a reaction (positive and negative feedback loops).
Wow! How science and computers have evolved and even merged since I was in school. What a way to incorporate student's knowledge and skills with computers to create a hands on application of more abstract concepts.
Populations - growth and interactions with the environment.