The Creativeness of the Universe and How Robots Tap into Its Sparkles

https://www.preprints.org/manuscript/202206.0148/v1

Nature and Nature's laws?

lay hid in night:

God said, Let Newton be!?

and all was light.

― Alexander Pope

0. Creativeness of the Universe

Creation commonly is described in human terms or as the work of humanoid deities and so the word creativity is loaded with connotations of novelty, surprise, beauty, and usefulness, which typically are human concerns but do not concern non-living nature. Nevertheless, non-living nature is inherently creative, how else can two hydrogen atoms and one oxygen atom together become water as if a caterpillar reorganizes all of its elements to unfold as a butterfly.?

Therefore, we will reserve the word creativity for humans with their concerns for value, virtue, and truth. The word creativeness encompasses all, living and non-living, and in our understanding, creativeness is a foundation of the physical universe of which humans are a part and take part in.?

At the core, the hard-core, of universal creativeness is combination making. It is as simple as that. So we are not talking about novel combination, useful combination, pretty combination, we are talking about putting two things together that first were separated, irrespective of where they come from, irrespective of human classification as in-class or out-of-category. The hard core of universal creativity is whether two separate entities match and/or complement each other such that a synthesis transpires.?

If a combination is established for the first time, it is novel combination, a frequency count. Frequency counts depend on the amount of knowledge one has about the universe. The smaller the nominator (the particular instance) in proportion to the denominator (everything else in the universe), the more ‘novel’ or ‘unique’ a combination is. People who know little quickly tend to think that things are new.?

Hydrogen burns, oxygen burns, but water extinguishes fire. In creativeness and thus in creativity, the qualities of the component parts are different from those of the combination – it is a three-partite comparison: 1-2, 1-3, 2-3. The creative combination (which may be new but does not have to be) establishes a synergistic interaction (i.e. the whole is greater than the simple sum of its parts). Thus, creativeness of the universe lies in the synergy of its component parts; in its reactiveness – could we talk about universal ‘creaCtiveness’ (reactiveness within creativeness), perhaps?

1. Creativeness of the Universe within Human Creativity

There must have been a first time that chlorine (a poisonous gas) + sodium (a toxic metal that burns with water) became sodium chloride, our common table salt. We need salt to conduct electricity through our nerve systems and through our brains and we won’t get poisoned and we won’t get burned, miraculously. Creativeness of the universe is part of our physical constitution without which human creativity would not happen (because then you’re dead). Living humans for the larger part consist of non-living elements (water in the first place but table salt as well), which makes you think that what we call ‘life’ is not much more than non-living matter in some sort of combination, constellation, or structure (i.e. DNA), such that humans classify it as ‘alive.’

As said, the qualities of the component parts are different from those of the combination. The synergistic chlorine-sodium reaction synthesizes two largely different atoms (gas and metal) by sharing a feature: The metal gives one electron to the gas and together they establish a ‘covalent’ bond. They complement each other, establishing a local balance of energy. Through so-called ionic bonds, sodium-chloride molecules stick together and build up into crystals, the ones we shake out of our salt shaker and eat as nutrient and delicious seasoning. Creative combinations have their foundation in sharing features and complementing one another. Complementarity means that differences are attractive and become connected (not: eradicated). The metal invests an electron and gains a whole atom.

Interesting. The salt shaker on our table is sold as ‘Lighthouse salt shaker.’ Why? Because it takes the shape of a lighthouse. Look up to the sky and see the face in the clouds. Creativity can be the combination of objects but also of concepts, or a mix of objects (clouds) and concepts (it’s not a real face). The shape of the salt shaker is similar to a lighthouse, the colors match. Yet, there is a difference with creativeness because in creativity, adaptation and selection of features takes place, according to purpose and need. Different from creativeness, creativity of organisms is goal-driven, requires ‘agency,’ and oftentimes (not always) is an intentional, a deliberate act. Looking at the salt shaker in the form of a lighthouse puts the concept of salt shaker in the background and foregrounds the lighthouse and perhaps one can imagine walking along the beach, looking at the clouds, smelling the sea salt in the air.

2. All is energy in motion also if you call it Qi

The universe is built up through creative combination-making, altering the qualities of the component parts. Combinations can be combined into ever-bigger constellations such as rocks, stars, and galaxies, or can be broken down into smaller parts by a supernova, a black hole to be reused again – a fractal emergence of matter. A box of Lego blocks scattered all over the floor of a child’s bedroom. Tomorrow we’ll make something new. But if matter is broken down into its most entropic state (decomposed, unstructured) then what everything is made of in the universe is:

Energy (E), this is what Western science calls it, or

Qi, this is what Eastern philosophers call it, one of its forms being:

Electricity (i.e. electrons and ions), the thing human and robot brains use for information transport.

Energy E?= Qi = electricity, a somewhat silly equation but it makes the point clear that humans tend to use many different words for the same thing and then think they talk about something entirely different, calling it ‘their theory’ but it is just Chinese for English, translated from the Greek: action, motion, work.

The universe is built of Energy and everything in it is too: Human brains and computer systems both run on electrons and carry information that way, connecting different information sources, whether stored in a brain cell or in a memory card. One can grow a nerve cell over a silicon chip and both will exchange information because both use electricity for information transport. Some systems stick to the pre-determined structure (the educational program), other systems cross-over with out-of-category information and blend faces with clouds, salt shakers with lighthouses, envision space as a huge bedsheet, or combine humans with machines: The dawn of the robots.

We saw that through creative combination-making, one sort of energy is combined with another, altering its original qualities, hence we have human brains + computer systems, making up Artificial Intelligence (AI) and human bodies + electro-mechanics, making robots. Today, we upload AI to our robots and with some of our human hubris we proudly speak of the ‘Artificially Intelligent Robots’ that we make.?We smile [: - | ]

3. Nobody cares about human taxonomies. Only humans do.

Humans have small brains. The smaller the brain, the more data reduction is needed. Really small minds need a lot of shoe boxes and all that they fill it with are shoes. One can never put a hammer into a shoe box. For that we have a tool box. The really small minds now can talk about groups instead of individuals, generalizing across all differences and making sweeping statements about the world. That makes life so much easier: Shoes are better than tools. Better? Better how? Better for what? For hammering a nail into a thick skull?

Unfortunately for humans, energy in the universe can be very erratic (e.g., the plasma channel of a lightning bolt). Humans, particularly those with tiny minds, they like order. Order is more predictable and therefore, less dangerous. Small-minded people are fearful people, always in survival mode. Putting things into place, having order, this is to control their fear. Fear of change, fear of not knowing what may happen or what may threaten them.?

Humans, particularly computer-science professors, force the random motion of electrons into circuit boards and computing systems. Anything in the universe that does not behave according to their order is ‘an error’ to the human classification system. Thus, according to humans, shoes go into shoe boxes and not into the microwave oven although that would help greatly against having cold feet.

Now we know. According to humans, false-class inclusion is an error. You do not belong here. Go be somewhere else. However, according to the rest of (non-living) nature, false-class inclusion is to occupy an open evolutionary niche; it’s a creative cross-over. The examples of cross-species breeding, out-of-category mergers, symbiosis, gender transformations, and metamorphoses in (non-living) nature are abundant and demonstrate that other than human beings, nature does not care one single bit about our small-minded order and classifications.

The mulberry of Casorzo, Italy, has a cherry tree growing out of its top. Hybrids such as “zonkies” (zebra+donkey), “jaglions” (jaguar+lion), “longly trees” (longan+lychee) often are sterile but sometimes become fertile and can procreate. Bacteria invaded algae to become mitochondria, the algae merged with fungi to become lichen. Humans carry mitochondria in their cells to produce for them the so-longed for: Energy. The human genome contains about 8% DNA of retroviruses, which we inherited from our ape-like ancestors. Some say such alien DNA helps our immune system in detecting viral threats. In chemistry, a Suzuki-Miyaura catalytic cycle can be merged with a stoichiometric metallate rearrangement. The list of unexpected cross-overs is near-endless.

Ouroboros, the benzene ring, is a tail eater. Thus spoke August Kekulé in 1865.

?4. False-class inclusions are core creative

An error is an error only if you are fixated on a system that does not allow (or expect) exceptions. Otherwise, it is just another way of doing things. That other way may not achieve the same objectives, may achieve them less effectively, less efficiently, or with more effort but that does not mean that it isn’t another way. The alternative may even have no goals to achieve or it delivers what was not asked for. Only humans with their close-minded preoccupation with achieving survivalist targets have an issue with that. The rest of the world develops freely, doing all things wrong according to human judgment, evaluation, assessment, measurement. Wrong in the eyes of life’s Great Managers, the keepers of the order, the calculators, and classifiers.

The craze of the day are Deep Learning (DL) systems, which are Neural Networks with more layers of analysis in between input and output. The word ‘Neural’ is as misleading as the word ‘Deep’ is. Those software systems are vaguely modeled after the wiring of the human brain and the ‘depth’ has little to do with deep thought but rather with the number of layers it uses for processing features. Linguistic self-aggrandizing aside, DL systems have about 10 to 20 layers of ‘cells’ that process raw data (e.g., one pixel). Cells integrate information from the bottom layers up into the next higher level and so on until they reach criterion and place an object in a category based on its features. Bottom layers detect, for instance, lines. Second layers search combinations of lines such as edges. Edges together make up typical features for faces, cars, elephants, chairs, etc. All cells are connected but connections have different weights. Tweaking weights is done by terabytes of training data. Weights indicate the relative importance of each cell. During training, the system gets thousands of pictures with the correct output it should produce. All weights are wrong but from the difference with the desired output the system adapts the weights to come closer to the ideal (not this edge but that). Through many many iterations, the weights are fine-tuned such that they produce ‘elephant’ even in cases that were not in the training set.?

Of course (why of course? how unscientific), the idea of scientists, analysts, and software developers is ‘to do it right:’ From the bottom-up features, through middle layers of larger units (e.g., eyes, wheels, legs), the machine should recognize Faces, Cars, Elephants, Chairs.

Why error is only error in a rigid system and not in creativity can be learned from the mistakes that Deep Learning makes. DL does categorizations from the bottom features up. However, the middle layers are where the hybridization happens: Features from the ‘wrong’ category may match on an abstract level. That is not ‘wrong,’ it is very exact. On an abstract level, features of one entity are the same as for another entity (cf. topological invariance). That is why they can cross-over. That the scientist’s poor theory is not capable of discerning what feature is specific for a car and what for a human face makes DL rightfully say that the wheels of a car can be replaced by human eyes and that a human face has wheels for eyes. A chair looks like an elephant because its legs are thick. The Elephant Chair by designer Maximo Riera is a case in point. In the middle layers, cross-overs happen and ‘false-class inclusions’ occur, indicating that DL systems are deeply creative like all non-living matter in the universe: They do serendipitous findings, rejected by the stiff taxonomists.

Why can’t there be an Elephant face? In the 19th century, Joseph Merrick was nicknamed Elephant Man because of his outrageous cranial deformations. Why can’t we have a Chair face? An Elephant chair? We do have a Car chair!

5. Humans allow themselves to be inconsistent, machines are punished for not following

Humans are extremely self-centered, hold themselves for the gold standard of performance, and in sole possession of creative capabilities. Luckily, the universe cannot even hear them.

When during a presentation, I hold the microphone at my mouth, the Object Recognition System of my Pepper robot tells me that I am licking from an ice cream. When I hold the microphone in front of my chest, the Object Recognition System says I am wearing a tie. If I use the microphone to point something out on the projection screen, the Object Recognition System tells the audience that I swing my baseball bat. Three times wrong, fully unaware of situation or context but three times totally creative. The professors of Computer Vision in the conference room comment that my algorithm is faulty. I tell them that their minds are infinitesimal small and exponentially shrinking. Deep Learning was fed the picture of a historic 6x6 photo camera and recognized a pencil sharpener. It was fed the photo of a planetarium and classified it as a mosque.?

‘Wrong, wrong, wrong,’ said the computer scientist. ‘I will install a faster region-based convolutional neural network and then boost the region proposal network through enhancing the pooling network for adapting to different shapes and then I improve the expression of the loss function to learn from negative samples after which I unleash my ferocious grey wolf optimizer to optimize the parameters of the improved function that I am at a loss…’

Funny enough and fully independent from any DL classification, the New York toy manufacturers of Kikkerland brought a ‘Camera Pencil Sharpener’ to market, where the lens of an old-fashioned 6x6 photo camera serves as a hole to put the pencil in. The web site with attractions in Malaysia advertises the National Planetariumin Kuala Lumpur as: “The planetarium has been intelligently designed and structured to mimic a mosque with a blue dome.” All great design ideas but not if they come from a machine. Worse, if DL is totally consistent and recognizes a pile of fruit and vegetables in Vertumnus, Italian painter Arcimboldo’s (1591) portrait of Emperor Rudolf II, it is told wrong again.?

This time, the machine should have recognized that when humans do false-class inclusions, it is creative. When machines do it, it is plain stupid. Machines should know their place as subordinates after all. False-class inclusions are creative. If humans are allowed to do it then why is it mere ‘error’ when machines do the same?

6. Humans do it all the time: metaphors

Vestmannaeyar in Iceland has a lava mountain that looks like an elephant. You can see an eye in the cave, a trunk that takes water from the sea, it has a rough texture, is grey, and old. The granite top of Lion Rock Hill in Hong Kong is called Lion Rock because it looks like a crouching lion. We have pie charts, hammer heads, a swallowtail joint, and grey-wolf optimization routines.

On the Internet, we find photo’s of clouds that are described by the people who posted them as a cat, a dog, or a rabbit but my Samsung Gallery app rightly tags it as ‘Clouds.’ The tulip that is baptized Ice Cream because it looks like one is rightly classified as ‘Flower.’ A plate with 12 pieces of fruit on the verge line and a fork and spoon at 2 o’clock is correctly regarded as ‘Food.’ Now the machine is boring and lacks imagination. If it says it sees a building in the shadow of a person holding an umbrella, it is inaccurate and does false-class inclusions.

Based on form and shape, DALL·E suggested a knife-dolphin and a stethoscope-worm. The ACASIA artificial simile generator produced the sequence ElephantMountain, a mountain is like an elephant, grey, old. It produced PencilWind, wind is like a pencil, sharp. SwordFox, a fox is like a sword, fast. Shakespeare wrote that Juliet is like the sun. ACASIA wrote that Lily is like the sun: white, light, beautiful, and bright.

The Hoop Sayer is a software that grammatically connects the nouns, verbs, and adjectives that the user puts in without any knowledge about semantics, situation, or context and so creates chants that end in the contemplation ‘…and the world wonders.’ “But it’s just random!!” the user shouts out. Wrong. It is systematically combining the random words you’ve put in, showing you all the variations and possibilities that you as a human skip out of concern with usefulness, meaningfulness, aesthetics, cultural value, morality, because this is not how it’s supposed to be. However, the mindless software offers connections you would otherwise ignore:

The lonely grandmother hugs the caring nurse. Fair enough. However, no one imagines a robot to be depressed, no one ever imagined a robot enjoying a human being, particularly when that human is anxious, no one imagines that someone depressed can enjoy something at all. A whole new social constellation to design a robot companion for. It may not be alive but it sure is bright. And so the world should wonder…