Nature Runs on Information
A look at the biomimicry design principle — nature runs on information, and its applications
Machine learning algorithms have become quite ubiquitous in solving complex, data-driven problems. They have wide-ranging applications, from agriculture and advertising, to recognizing handwriting and analyzing financial markets. Unlike ordinary computer programs, where a set of instructions are fed into the computer, here the machine identifies patterns in the information provided, and makes predictions or decisions for new data.
Deep learning is a subfield of machine learning that is inspired by how the brain works. It is able to learn from unlabeled and unstructured data, and can perform human-like functions such as speech recognition and object detection. Here, deep learning mimics one of the most obvious examples of the biomimicry design principle — nature runs on information.
Biomimicry is the process of using nature’s millennia of experience to solve our problems without harming the environment. There is a huge focus on sustainability and humility. To make the design process simpler, the main facets of nature’s designs have been broken down into ten design principles, which is not an exhaustive list, but rather serves as a starting point to consider possible solutions from.
These principles help in designing objects or services that function efficiently, consume less resources, and produce better outcomes. But unlike the rather selfish approach we usually take with design, a biomimetic approach results in sustainable solutions that work with the environment, rather than destroy it. It also provides us with a newfound respect for nature, who has had years to perfect her solutions with the aid of evolution. It is definitely conceivable that the answers to all our problems exist out there. It is up to us to go and get our hands dirty to find the best one.
One of nature’s design principles is the fact that nature runs on information — a simple statement, but one that has widespread implications on how nature functions. Information is received from the environment, and communication can occur within the body of an individual organism or among a group of organisms, and is used to respond to that information in an appropriate manner. This allows the organisms to survive environmental changes.
One such example is observed in the slime mold Physarum polycephalum, a unicellular organism. Generally, when we think of complex behaviors, we also think of complex organisms. But in this case, a simple organism has the ability to learn from its environment and past behaviour, and can even communicate its learnings to other slime molds. This communication happens when two slime molds fuse together. The exact method of information transmission is unknown, but a recent study suggests that stimuli in the surroundings can trigger a signaling molecule, most likely calcium, to propagate through the body of the slime mold. This creates a positive feedback loop, as the signaling molecule increases the local contraction amplitude and also generates additional cytoplasmic flows, which helps it move even further into the network.
It has been observed that the slime mold can be trained to cross a bridge with salt, something it usually avoids. When this slime mold was fused with another and then later separated, the other slime mold was also able to cross the same bridge, clear evidence that the information had been communicated successfully. This also explains how slime mold is able to find the shortest path between stimuli, despite not having a central organizing body like a nervous system.
This strategy has been used by scientists to create networks that balance efficiency with robustness, i.e. networks that minimize energy while still being able to recover from sudden changes. When a model was made representing the cities in Japan, and a railway network was self-organized through this method, the results were very similar to the actual network in use, although in this case it was achieved without a central organising body, making the results even more impressive. Such a model would be useful to improve different self-organized networks like remote sensor arrays, mobile ad hoc networks, and wireless mesh networks.
In another example, one that exemplifies the age-old adage ‘you are what you eat’, Malacosteus niger, a type of dragonfish, had gained the ability to sense far-red light from the bacteria it had consumed, thus allowing it to hunt at night. Far-red light refers to the light in the visible spectrum whose wavelength is the longest. Not all organisms can see this light, but M. niger can sense it because of two sets of bioluminescent organs, where one gives off blue-green light and the other expels far-red light. While most organisms at this depth can sense the former, only dragonfish sense the latter. There are two functions of this strategy — it can communicate with other dragonfish, and it can see prey that other organisms can’t. Thus the light allows it to sense its environment, and using this information, dragonfish feed and communicate. They received this ability because of a certain pigment present in their retina that can absorb far-red light. This pigment is from the chlorophyll of bacteria that also have this function, which the dragonfish indirectly consume by eating copepods that feed on it. Thus, this is an example of how the smooth functioning of nature depends on its ability to learn from its environment and act accordingly.
There are a lot of real-life applications that have been devised through biomimicry that also exemplify this principle. One has been inspired by the slime mold mentioned earlier, and is an algorithm that is used to predict the formation of cosmic webs that connect galaxies, potentially allowing us to map the entire universe.
Another useful application is the creation of bandages for burn victims that are able to recognize disease-causing bacteria which usually escape detection by humans. The way this works is by the production of fluorescent signals when such bacteria are present on the surface of the bandage. This invention was inspired by the way our body is alerted to the onslaught of bacteria that break through cell membranes. In such cases, the bacteria release chemicals that inform the body of its presence. Similarly, the capsules containing fluorescent dye present on the surface of the bandage react with the toxins that the bacteria release, and UV light can be used to check for leakage of the dye from the capsules.
The term Black Swan is used to refer to events that have a huge impact on society but are rarely known to happen. Predicting these events are tricky, but doing so could save a lot of resources, as it would help us to be prepared for the worst. Such a prediction model was created by a team of researchers from Stanford using information collected from different ecosystems — marine plankton, intertidal mollusks, and deciduous forests. It was observed that even though these ecosystems were so varied, changes in species level remained similar, and it was possible to accurately predict disastrous events that are unlikely to occur. It was ascertained that universal laws do control species dynamics, and the model developed from the high quality of data supplied could then be used in situations with less data.
It is now clear that nature uses data to decide how to react to its surroundings in an efficient manner. If you think about it, this is also apparent in our daily interactions with our environment. If it is cold, we put on a sweater. If our stomach growls, we eat. If we yawn, we know it is time for us to go to bed. However, even a doctrine as simple as this can have a revolutionizing impact when integrated with our designs and plans. The value that high quality data can add to our applications is not only limited to machine learning algorithms, but can be used in all aspects of our life. Students of biomimicry can incorporate and benefit from this design principle in two ways — one is by understanding the process by which nature receives and transmits information in order to imitate it, and the other is by extracting information from nature to make decisions. Information makes us smarter, and should be used to make our solutions smarter as well.
[1]: Mechanism of signal propagation in P. polycephalum https://www.pnas.org/content/114/20/5136
[2]: Information Gained, Stored, and Transferred Without Brain, Mary Hoff, Ask Nature https://asknature.org/strategy/brainless-slime-molds-both-learn-and-teach/
[3]: Rules for Biologically Inspired Adaptive Network Design https://science.sciencemag.org/content/327/5964/439.abstract
[4]: Bacteria Help Sense Far‑red Light, Ask Nature https://asknature.org/strategy/bacteria-help-sense-far-red-light/
[5]: Revealing the Dark Threads of the Cosmic Web https://iopscience.iop.org/article/10.3847/2041-8213/ab700c/meta
[6]: Prototype Development of the Intelligent Hydrogel Wound Dressing and Its Efficacy in the Detection of Model Pathogenic Wound Biofilms https://pubs.acs.org/doi/abs/10.1021/acsami.5b07372
[7]: Forecasting unprecedented ecological fluctuations https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1008021
