Natural Computing

Non-linear geometry attempts to model turbulence. Water is moving smoothly through a passage at a given speed. As the speed increases linearly, the movement of the water updates in a predictable pattern, until a certain threshold is reached, and all hell breaks loose. The physical shape of the water moving through the passage becomes wild and rough, with unexpected air pockets and ever-changing wave forms. Turbulence breaks with linear models, introducing chaotic, random behavior that is all but impossible to predict. It seems to fall outside of the concept of a pattern because its motion cannot be described — it is not “regular.” However, the occurrence of turbulence in systems is a systemic, predictable phenomenon, and is one of the objects of study in the field of chaos theory.

Turbulence plays an important role in nature: it is both a creative and a destructive force. On a large scale, turbulence introduces randomness where generative creativity is needed, yet its disruption of stable conditions can be destructive as well, when for instance an individual’s heartbeat becomes turbulent and irregular. Turbulence is like a natural random number generator of any substance or sequence that exists in “flow” — water, air, blood, heartbeat, information — only instead of numbers it generates complete states of being.

One thing that I don’t think ever experiences turbulence is light. I could be wrong, but I don’t think that photons are exactly in a state of flow. Or perhaps they are, and perhaps they do experience turbulence, but not in the physical universe that we have access to, because every lightwave we have ever encountered or theorized about travels at the same speed. We know how to slow light down with lenses and water, but that relationship is linear. Turbulence would only occur if we were to speed light up, and so far this eventuality remains classified as a complete impossibility.

I want to explore turbulence in networks of information, and examine the possibility of intentionally building in turbulence into certain kinds of systems in order to benefit from the creative, generative aspects of the phenomenon — the ability to randomly generate complete states of being. Imagine training a neural network on a large dataset replete with examples of natural turbulence. There are so many challenges inherent in this idea. What would constitute a training example? A particle simulation wouldn’t be complete. Standing in a giant wind tunnel with a bunch of sensors would cast doubt on the sensitivity of the instruments to capture finely enough the necessary granularity of data. Video footage of turbulent water or wind would result in a 2D representation of a 3D phenomenon. So, perhaps herein lies some kind of answer — there could be a two-step process.

First, build a neural network that is capable of translating from 2D to 3D. Second, train a neural network on 2D examples of turbulence and see what kind of 3D representation gets spit out. I’m not sure about the validity of such an approach, but it’s worth exploring.

The end-game is of course to have a trained neural network that can instantiate a state of turbulence within a given system. Once this is accomplished, you could then utilize it, experiment with it and understand how it can bring value to a various challenges inherent in the design and governance of a fully functioning shared power model embodied in part by a large-scale peer-to-peer open source distributed network. With today’s models of computing, this may not be of very great value. But I am interested in the development of a kind of natural computing that has its foundations in natural, stochastic systems. Biomimicry applied not just to the high-level design of a software system, but to the actual embodiment of the language itself.

Without realizing it consciously, I have been moving toward an idea of natural computing for some time now, with my twin interests in synthetic biology and artificial intelligence. If we cannot go back to a way of life where we wear nature like a cloak, let us build our civilization’s cloak naturally. Innate in human beings is the desire to create, build, and bring new forms into being in the world. Then let what we build be able to return to the earth — do not separate the anthropocene from the the larger ecosystem, let it have a porous membrane that makes it vulnerable to the universe.

This could be part of the answer to a question I asked in 2011, namely, how can we design for serendipity? Another side of the same question is about promoting, incentivizing or protecting cross-disciplinary thinking and syncretic work, which may be undertaken consciously.

Imagine an alternative metaverse, where experiences are not curated, where the interface with objects, people, and ideas are more raw, still “media” that mediates between the subject and the object, between people and between people and ideas, but where the landscape is more jagged in structure, internally, less of a color-by-numbers type of endeavor. The very veins of the information flow are subject to stochastic rhythms, so that randomness can occasionally break into the ordered existence.

One major question that this brings up is the idea of inspectability of code and the ethical dimensions that come into play when a system has been empowered with decision-making abilities, or even just significant influence, on practical matters in our lives. The ability to throw parts of a system into a state of turbulence has a corollary in today’s deep neural networks in that in both cases, human designers are not the ultimate arbiters guiding the process, or even fully understanding all aspects of it. This is the knife’s edge of questioning right now — how can we incorporate the benefits of randomness, serendipity, turbulence, and the unknown into the vast systems that influence so much of our lives, while also maintaining an ethical understanding and safety around the operation of such systems?

There is a painting of my grandfather that illustrates a imagined period after nuclear war. The painting says, “We thought of living in caves, but at last were persuaded to build our homes again.” We are not content to live in caves. I know that is true for me as well. And, I am not content with the level of conscientiousness embedded in the computing and technology industry today.

read original article at——artificial_intelligence-5