My editorial on You Tube

This is one of those moments when I need to reassess what the hell I am doing. Scientifically, I mean. Of course, it is good to reassess things existentially, too, every now and then, but for the moment I am limiting myself to science. Simpler and safer than life in general. Anyway, I have a financial scheme in mind, where local crowdfunding platforms serve to support the development of local suppliers in renewable energies. The scheme is based on the observable difference between prices of electricity for small users (higher), and those reserved to industrial scale users (lower). I wonder if small consumers would be ready to pay the normal, relatively higher price in exchange of a package made of: a) electricity and b) shares in the equity of its suppliers.

I have a general, methodological hypothesis in mind, which I have been trying to develop over the last 2 years or so: collective intelligence. I hypothesise that collective behaviour observable in markets can be studied as a manifestation of collective intelligence. The purpose is to go beyond optimization and to define, with scientific rigour, what are the alternative, essentially equiprobable paths of change that a complex market can take. I think such an approach is useful when I am dealing with an economic model with a lot of internal correlation between variables, and that correlation can be so strong that it turns into those variables basically looping on each other. In such a situation, distinguishing independent variables from the dependent ones becomes bloody hard, and methodologically doubtful.

On the grounds of literature, and my own experimentation, I have defined three essential traits of such collective intelligence: a) distinction between structure and instance b) capacity to accumulate experience, and c) capacity to pass between different levels of freedom in social cohesion. I am using an artificial neural network, a multi-layer perceptron, in order to simulate such collectively intelligent behaviour.

The distinction between structure and instance means that we can devise something, make different instances of that something, each different by some small details, and experiment with those different instances in order to devise an even better something. When I make a mechanical clock, I am a clockmaker. When I am able to have a critical look at this clock, make many different versions of it – all based on the same structural connections between mechanical parts, but differing from each other by subtle details – and experiment with those multiple versions, I become a meta-clock-maker, i.e. someone who can advise clockmakers on how to make clocks. The capacity to distinguish between structures and their instances is one of the basic skills we need in life. Autistic people have a big problem in that department, as they are mostly on the instance side. To a severely autistic person, me in a blue jacket, and me in a brown jacket are two completely different people. Schizophrenic people are on the opposite end of the spectrum. To them, everything is one and the same structure, and they cannot cope with instances. Me in a blue jacket and me in a brown jacket are the same as my neighbour in a yellow jumper, and we all are instances of the same alien monster. I know you think I might be overstating, but my grandmother on the father’s side used to suffer from schizophrenia, and it was precisely that: to her, all strong smells were the manifestation of one and the same volatile poison sprayed in the air by THEM, and every person outside a circle of about 19 people closest to her was a member of THEM. Poor Jadwiga.

In economics, the distinction between structure and instance corresponds to the tension between markets and their underpinning institutions. Markets are fluid and changeable, they are like constant experimenting. Institutions give some gravitas and predictability to that experimenting. Institutions are structures, and markets are ritualized manners of multiplying and testing many alternative instances of those structures.

The capacity to accumulate experience means that as we experiment with different instances of different structures, we can store information we collect in the process, and use this information in some meaningful way. My great compatriot, Alfred Korzybski, in his general semantics, used to designate it as ‘the capacity to bind time’. The thing is not as obvious as one could think. A Nobel-prized mathematician, Reinhard Selten, coined up the concept of social games with imperfect recall (Harsanyi, Selten 1988[1]). He argued that as we, collective humans, accumulate and generalize experience about what the hell is going on, from time to time we shake off that big folder, and pick the pages endowed with the most meaning. All the remaining stuff, judged less useful on the moment, is somehow archived in culture, so as it basically stays there, but becomes much harder to access and utilise. The capacity to accumulate experience means largely the way of accumulating experience, and doing that from-time-to-time archiving. We can observe this basic distinction in everyday life. There are things that we learn sort of incrementally. When I learn to play piano – which I wish I was learning right now, cool stuff – I practice, I practice, I practice and… I accumulate learning from all those practices, and one day I give a concert, in a pub. Still, other things, I learn them sort of haphazardly. Relationships are a good example. I am with someone, one day I am mad at her, the other day I see her as the love of my life, then, again, she really gets on my nerves, and then I think I couldn’t live without her etc. Bit of a bumpy road, isn’t it? Yes, there is some incremental learning, but you become aware of it after like 25 years of conjoint life. Earlier on, you just need to suck ass and keep going.

There is an interesting theory in economics, labelled as « semi – martingale » (see for example: Malkiel, Fama 1970[2]). When we observe changes in stock prices, in a capital market, we tend to say they are random, but they are not. You can test it. If the price is really random, it should fan out according to the pattern of normal distribution. This is what we call a full martingale. Any real price you observe actually swings less broadly than normal distribution: this is a semi-martingale. Still, anyone with any experience in investment knows that prediction inside the semi-martingale is always burdened with a s**tload of error. When you observe stock prices over a long time, like 2 or 3 years, you can see a sequence of distinct semi-martingales. From September through December it swings inside one semi-martingale, then the Ghost of Past Christmases shakes it badly, people panic, and later it settles into another semi-martingale, slightly shifted from the preceding one, and here it goes, semi-martingaling for another dozen of weeks etc.

The central theoretical question in this economic theory, and a couple of others, spells: do we learn something durable through local shocks? Does a sequence of economic shocks, of whatever type, make a learning path similar to the incremental learning of piano playing? There are strong arguments in favour of both possible answers. If you get your face punched, over and over again, you must be a really dumb asshole not to learn anything from that. Still, there is that phenomenon called systemic homeostasis: many systems, social structures included, tend to fight for stability when shaken, and they are frequently successful. The memory of shocks and revolutions is frequently erased, and they are assumed to have never existed.

The issue of different levels in social cohesion refers to the so-called swarm theory (Stradner et al 2013[3]). This theory studies collective intelligence by reference to animals, which we know are intelligent just collectively. Bees, ants, hornets: all those beasts, when acting individually, as dumb as f**k. Still, when they gang up, they develop amazingly complex patterns of action. That’s not all. Those complex patterns of theirs fall into three categories, applicable to human behaviour as well: static coupling, dynamic correlated coupling, and dynamic random coupling.

When we coordinate by static coupling, we always do things together in the same way. These are recurrent rituals, without much room for change. Many legal rules, and institutions they form the basis of, are examples of static coupling. You want to put some equity-based securities in circulation? Good, you do this, and this, and this. You haven’t done the third this? Sorry, man, but you cannot call it a day yet. When we need to change the structure of what we do, we should somehow loosen that static coupling and try something new. We should dissolve the existing business, which is static coupling, and look for creating something new. When we do so, we can sort of stay in touch with our customary business partners, and after some circling and asking around we form a new business structure, involving people we clearly coordinate with. This is dynamic correlated coupling. Finally, we can decide to sail completely uncharted waters, and take our business concept to China, or to New Zealand, and try to work with completely different people. What we do, in such a case, is emitting some sort of business signal into the environment, and waiting for any response from whoever is interested. This is dynamic random coupling. Attracting random followers to a new You Tube channel is very much an example of the same.

At the level of social cohesion, we can be intelligent in two distinct ways. On the one hand, we can keep the given pattern of collective associations behaviour at the same level, i.e. one of the three I have just mentioned. We keep it ritualized and static, or somehow loose and dynamically correlated, or, finally, we take care of not ritualizing too much and keep it deliberately at the level of random associations. On the other hand, we can shift between different levels of cohesion. We take some institutions, we start experimenting with making them more flexible, at some point we possibly make it as free as possible, and we gain experience, which, in turn, allows us to create new institutions.

When applying the issue of social cohesion in collective intelligence to economic phenomena, we can use a little trick, to be found, for example, in de Vincenzo et al (2018[4]): we assume that quantitative economic variables, which we normally perceive as just numbers, are manifestations of distinct collective decisions. When I have the price of energy, let’s say, €0,17 per kilowatt hour, I consider it as the outcome of collective decision-making. At this point, it is useful to remember the fundamentals of intelligence. We perceive our own, individual decisions as outcomes of our independent thinking. We associate them with the fact of wanting something, and being apprehensive regarding something else etc. Still, neurologically, those decisions are outcomes of some neurons firing in a certain sequence. Same for economic variables, i.e. mostly prices and quantities: they are fruit of interactions between the members of a community. When I buy apples in the local marketplace, I just buy them for a certain price, and, if they look bad, I just don’t buy. This is not any form of purposeful influence upon the market. Still, when 10 000 people like me do the same, sort of ‘buy when price good, don’t when the apple is bruised’, a patterned process emerges. The resulting price of apples is the outcome of that process.

Social cohesion can be viewed as association between collective decisions, not just between individual actions. The resulting methodology is made, roughly speaking, of three steps. Step one: I put all the economic variables in my model over a common denominator (common scale of measurement). Step two: I calculate the relative cohesion between them with the general concept of a fitness function, which I can express, for example, as the Euclidean distance between local values of variables in question. Step three: I calculate the average of those Euclidean distances, and I calculate its reciprocal, like « 1/x ». This reciprocal is the direct measure of cohesion between decisions, i.e. the higher the value of this precise « 1/x », the more cohesion between different processes of economic decision-making.

Now, those of you with a sharp scientific edge could say now: “Wait a minute, doc. How do you know we are talking about different processes of decision making? Who do you know that variable X₁ comes from a different process than variable X₂?”. This is precisely my point. The swarm theory tells me that if I can observe changing a cohesion between those variables, I can reasonably hypothesise that their underlying decision-making processes are distinct. If, on the other hand, their mutual Euclidean distance stays the same, I hypothesise that they come from the same process.

Summing up, here is the general drift: I take an economic model and I formulate three hypotheses as for the occurrence of collective intelligence in that model. Hypothesis #1: different variables of the model come from different processes of collective decision-making.

Hypothesis #2: the economic system underlying the model has the capacity to learn as a collective intelligence, i.e. to durably increase or decrease the mutual cohesion between those processes. Hypothesis #3: collective learning in the presence of economic shocks is different from the instance of learning in the absence of such shocks.

They look nice, those hypotheses. Now, why the hell should anyone bother? I mean what are the practical outcomes of those hypotheses being true or false? In my experimental perceptron, I express the presence of economic shocks by using hyperbolic tangent as neural function of activation, whilst the absence of shocks (or the presence of countercyclical policies) is expressed with a sigmoid function. Those two yield very different processes of learning. Long story short, the sigmoid learns more, i.e. it accumulates more local errors (this more experimental material for learning), and it generates a steady trend towards lower a cohesion between variables (decisions). The hyperbolic tangent accumulates less experiential material (it learns less), and it is quite random in arriving to any tangible change in cohesion. The collective intelligence I mimicked with that perceptron looks like the kind of intelligence, which, when going through shocks, learns only the skill of returning to the initial position after shock: it does not create any lasting type of change. The latter happens only when my perceptron has a device to absorb and alleviate shocks, i.e. the sigmoid neural function.

When I have my perceptron explicitly feeding back that cohesion between variables (i.e. feeding back the fitness function considered as a local error), it learns less and changes less, but not necessarily goes through less shocks. When the perceptron does not care about feeding back the observable distance between variables, there is more learning and more change, but not more shocks. The overall fitness function of my perceptron changes over time The ‘over time’ depends on the kind of neural activation function I use. In the case of hyperbolic tangent, it is brutal change over a short time, eventually coming back to virtually the same point that it started from. In the hyperbolic tangent, the passage between various levels of association, according to the swarm theory, is super quick, but not really productive. In the sigmoid, it is definitely a steady trend of decreasing cohesion.

I want to know what the hell I am doing. I feel I have made a few steps towards that understanding, but getting to know what I am doing proves really hard.

I am consistently delivering good, almost new science to my readers, and love doing it, and I am working on crowdfunding this activity of mine. As we talk business plans, I remind you that you can download, from the library of my blog, the business plan I prepared for my semi-scientific project Befund  (and you can access the French version as well). You can also get a free e-copy of my book ‘Capitalism and Political Power’ You can support my research by donating directly, any amount you consider appropriate, to my PayPal account. You can also consider going to my Patreon page and become my patron. If you decide so, I will be grateful for suggesting me two things that Patreon suggests me to suggest you. Firstly, what kind of reward would you expect in exchange of supporting me? Secondly, what kind of phases would you like to see in the development of my research, and of the corresponding educational tools?

[1] Harsanyi, J. C., & Selten, R. (1988). A general theory of equilibrium selection in games. MIT Press Books, 1.

[2] Malkiel, B. G., & Fama, E. F. (1970). Efficient capital markets: A review of theory and empirical work. The journal of Finance, 25(2), 383-417.

[3] Stradner, J., Thenius, R., Zahadat, P., Hamann, H., Crailsheim, K., & Schmickl, T. (2013). Algorithmic requirements for swarm intelligence in differently coupled collective systems. Chaos, Solitons & Fractals, 50, 100-114.

[4] De Vincenzo, I., Massari, G. F., Giannoccaro, I., Carbone, G., & Grigolini, P. (2018). Mimicking the collective intelligence of human groups as an optimization tool for complex problems. Chaos, Solitons & Fractals, 110, 259-266.

My editorial on You Tube

I keep working on the application of neural networks to simulate the workings of collective intelligence in humans. I am currently macheting my way through the model proposed by de Vincenzo et al in their article entitled ‘Mimicking the collective intelligence of human groups as an optimization tool for complex problems’ (2018[1]). In the spirit of my own research, I am trying to use optimization tools for a slightly different purpose, that is for simulating the way things are done. It usually means that I relax some assumptions which come along with said optimization tools, and I just watch what happens.

Vincenzo et al propose a model of artificial intelligence, which combines a classical perceptron, such as the one I have already discussed on this blog (see « More vigilant than sigmoid », for example) with a component of deep learning based on the observable divergences in decisions. In that model, social agents strive to minimize their divergences and to achieve relative consensus. Mathematically, it means that each decision is characterized by a fitness function, i.e. a function of mathematical distance from other decisions made in the same population.

I take the tensors I have already been working with, namely the input tensor T_I = {LCOE_R, LCOE_NR, K_R, K_NR, I_R, I_NR, P_A;R, P_A;NR, P_B;R, P_B;NR} and the output tensor is T_O = {Q_R/N; Q_NR/N}. Once again, consult « More vigilant than sigmoid » as for the meaning of those variables. In the spirit of the model presented by Vincenzo et al, I assume that each variable in my tensors is a decision. Thus, for example, P_A;R, i.e. the basic price of energy from renewable sources, which small consumers are charged with, is the tangible outcome of a collective decision. Same for the levelized cost of electricity from renewable sources, the LCOE_R, etc. For each i-th variable x_i in T_I and T_O, I calculate its relative fitness to the overall universe of decisions, as the average of itself, and of its Euclidean distances to other decisions. It looks like:

V(x_i) = (1/N)*{x_i + [(x_i – x_i;1)²]^0,5 + [(x_i – x_i;2)²]^0,5 + … + [(x_i – x_i;K)²]^0,5}

…where N is the total number of variables in my tensors, and K = N – 1.

In a next step, I can calculate the average of averages, thus to sum up all the individual V(x_i)’s and divide that total by N. That average V*(x) = (1/N) * [V(x₁) + V(x₂) + … + V(x_N)] is the measure of aggregate divergence between individual variables considered as decisions.

Now, I imagine two populations: one who actively learns from the observed divergence of decisions, and another one who doesn’t really. The former is represented with a perceptron that feeds back the observable V(x_i)’s into consecutive experimental rounds. Still, it is just feeding that V(x_i) back into the loop, without any a priori ideas about it. The latter is more or less what it already is: it just yields those V(x_i)’s but does not do much about them.

I needed a bit of thinking as for how exactly should that feeding back of fitness function look like. In the algorithm I finally came up with, it looks differently for the input variables on the one hand, and for the output ones. You might remember, from the reading of « More vigilant than sigmoid », that my perceptron, in its basic version, learns by estimating local errors observed in the last round of experimentation, and then adding those local errors to the values of input variables, just to make them roll once again through the neural activation function (sigmoid or hyperbolic tangent), and see what happens.

As I upgrade my perceptron with the estimation of fitness function V(x_i), I ask: who estimates the fitness function? What kind of question is that? Well, a basic one. I have that neural network, right? It is supposed to be intelligent, right? I add a function of intelligence, namely that of estimating the fitness function. Who is doing the estimation: my supposedly intelligent network or some other intelligent entity? If it is an external intelligence, mine, for a start, it just estimates V(x_i), sits on its couch, and watches the perceptron struggling through the meanders of attempts to be intelligent. In such a case, the fitness function is like sweat generated by a body. The body sweats but does not have any way of using the sweat produced.

Now, if the V(x_i) is to be used for learning, the perceptron is precisely the incumbent intelligent structure supposed to use it. I see two basic ways for the perceptron to do that. First of all, the input neuron of my perceptron can capture the local fitness functions on input variables and add them, as additional information, to the previously used values of input variables. Second of all, the second hidden neuron can add the local fitness functions, observed on output variables, to the exponent of the neural activation function.

I explain. I am a perceptron. I start my adventure with two tensors: input T_I = {LCOE_R, LCOE_NR, K_R, K_NR, I_R, I_NR, P_A;R, P_A;NR, P_B;R, P_B;NR} and output T_O = {Q_R/N; Q_NR/N}. The initial values I start with are slightly modified in comparison to what was being processed in « More vigilant than sigmoid ». I assume that the initial market of renewable energies – thus most variables of quantity with ‘R’ in subscript – is quasi inexistent. More specifically, Q_R/N = 0,01 and  Q_NR/N = 0,99 in output variables, whilst in the input tensor I have capital invested in capacity I_R = 0,46 (thus a readiness to go and generate from renewables), and yet the crowdfunding flow K is K_R = 0,01 for renewables and K_NR = 0,09 for non-renewables. If you want, it is a sector of renewable energies which is sort of ready to fire off but hasn’t done anything yet in that department. All in all, I start with: LCOE_R = 0,26; LCOE_NR = 0,48; K_R = 0,01; K_NR = 0,09; I_R = 0,46; I_NR = 0,99; P_A;R = 0,71; P_A;NR = 0,46; P_B;R = 0,20; P_B;NR = 0,37; Q_R/N = 0,01; and Q_NR/N = 0,99.

Being a pure perceptron, I am dumb as f**k. I can learn by pure experimentation. I have ambitions, though, to be smarter, thus to add some deep learning to my repertoire. I estimate the relative mutual fitness of my variables according to the V(x_i) formula given earlier, as arithmetical average of each variable separately and its Euclidean distance to others. With the initial values as given, I observe: V(LCOE_R; t₀) = 0,302691788; V(LCOE_NR; t₀) = 0,310267104; V(K_R; t₀) = 0,410347388; V(K_NR; t₀) = 0,363680721; V(I_R ; t₀) = 0,300647174; V(I_NR ; t₀) = 0,652537097; V(P_A;R ; t₀) = 0,441356844 ; V(P_A;NR ; t₀) = 0,300683099 ; V(P_B;R ; t₀) = 0,316248176 ; V(P_B;NR ; t₀) = 0,293252713 ; V(Q_R/N ; t₀) = 0,410347388 ; and V(Q_NR/N ; t₀) = 0,570485945. All that stuff put together into an overall fitness estimation is like average V*(x; t₀) = 0,389378787.

I ask myself: what happens to that fitness function when as I process information with my two alternative neural functions, the sigmoid or the hyperbolic tangent. I jump to experimental round 1500, thus to t₁₅₀₀, and I watch. With the sigmoid, I have V(LCOE_R; t₁₅₀₀) =  0,359529289 ; V(LCOE_NR; t₁₅₀₀) =  0,367104605; V(K_R; t₁₅₀₀) =  0,467184889; V(K_NR; t₁₅₀₀) = 0,420518222; V(I_R ; t₁₅₀₀) =  0,357484675; V(I_NR ; t₁₅₀₀) =  0,709374598; V(P_A;R ; t₁₅₀₀) =  0,498194345; V(P_A;NR ; t₁₅₀₀) =  0,3575206; V(P_B;R ; t₁₅₀₀) =  0,373085677; V(P_B;NR ; t₁₅₀₀) =  0,350090214; V(Q_R/N ; t₁₅₀₀) =  0,467184889; and V(Q_NR/N ; t₁₅₀₀) = 0,570485945, with average V*(x; t₁₅₀₀) =  0,441479829.

Hmm, interesting. Working my way through intelligent cognition with a sigmoid, after 1500 rounds of experimentation, I have somehow decreased the mutual fitness of decisions I make through individual variables. Those V(x_i)’s have changed. Now, let’s see what it gives when I do the same with the hyperbolic tangent: V(LCOE_R; t₁₅₀₀) =   0,347752478; V(LCOE_NR; t₁₅₀₀) =  0,317803169; V(K_R; t₁₅₀₀) =   0,496752021; V(K_NR; t₁₅₀₀) = 0,436752021; V(I_R ; t₁₅₀₀) =  0,312040791; V(I_NR ; t₁₅₀₀) =  0,575690006; V(P_A;R ; t₁₅₀₀) =  0,411438698; V(P_A;NR ; t₁₅₀₀) =  0,312052766; V(P_B;R ; t₁₅₀₀) = 0,370346458; V(P_B;NR ; t₁₅₀₀) = 0,319435252; V(Q_R/N ; t₁₅₀₀) =  0,496752021; and V(Q_NR/N ; t₁₅₀₀) = 0,570485945, with average V*(x; t₁₅₀₀) =0,413941802.

Well, it is becoming more and more interesting. Being a dumb perceptron, I can, nevertheless, create two different states of mutual fitness between my decisions, depending on the kind of neural function I use. I want to have a bird’s eye view on the whole thing. How can a perceptron have a bird’s eye view of anything? Simple: it rents a drone. How can a perceptron rent a drone? Well, how smart do you have to be to rent a drone? Anyway, it gives something like the graph below:

Wow! So this is what I do, as a perceptron, and what I haven’t been aware so far? Amazing. When I think in sigmoid, I sort of consistently increase the relative distance between my decisions, i.e. I decrease their mutual fitness. The sigmoid, that function which sorts of calms down any local disturbance, leads to making a decision-making process like less coherent, more prone to embracing a little chaos. The hyperbolic tangent thinking is different. It occasionally sort of stretches across a broader spectrum of fitness in decisions, but as soon as it does so, it seems being afraid of its own actions, and returns to the initial level of V*(x). Please, note that as a perceptron, I am almost alive, and I produce slightly different outcomes in each instance of myself. The point is that in the line corresponding to hyperbolic tangent, the comb-like pattern of small oscillations can stretch and move from instance to instance. Still, it keeps the general form of a comb.

OK, so this is what I do, and now I ask myself: how can I possibly learn on that thing I have just become aware I do? As a perceptron, endowed with this precise logical structure, I can do one thing with information: I can arithmetically add it to my input. Still, having some ambitions for evolving, I attempt to change my logical structure, and I risk myself into incorporating somehow the observable V(x_i) into my neural activation function. Thus, the first thing I do with that new learning is to top the values of input variables with local fitness functions observed in the previous round of experimenting. I am doing it already with local errors observed in outcome variables, so why not doubling the dose of learning? Anyway, it goes like: x_i(t₀) = x_i(t-1) + e(x_i; t_-1) + V(x_i; t_-1). It looks interesting, but I am still using just a fraction of information about myself, i.e. just that about input variables. Here is where I start being really ambitious. In the equation of the sigmoid function, I change s = 1 / [1 + exp(∑x_i*W_i)] into s = 1 / [1 + exp(∑x_i*W_i + V(T_o)], where V(T_o) stands for local fitness functions observed in output  variables. I do the same by analogy in my version based on hyperbolic tangent. The th = [exp(2*∑x_i*w_i)-1] / [exp(2*∑x_i*w_i) + 1] turns into th = {exp[2*∑x_i*w_i + V(T_o)] -1} / {exp[2*∑x_i*w_i + V(T_o)] + 1}. I do what I know how to do, i.e. adding information from fresh observation, and I apply it to change the structure of my neural function.

All those ambitious changes in myself, put together, change my pattern of learing as shown in the graph below:

When I think sigmoid, the fact of feeding back my own fitness function does not change much. It makes the learning curve a bit steeper in the early experimental rounds, and makes it asymptotic to a little lower threshold in the last rounds, as compared to learning without feedback on V(x_i). Yet, it is the same old sigmoid, with just its sleeves ironed. On the other hand, the hyperbolic tangent thinking changes significantly. What used to look like a comb, without feedback, now looks much more aggressive, like a plough on steroids. There is something like a complex cycle of learning on the internal cohesion of decisions made. Generally, feeding back the observable V(x_i) increases the finally achieved cohesion in decisions, and, in the same time, it reduces the cumulative error gathered by the perceptron. With that type of feedback, the cumulative error of the sigmoid, which normally hits around 2,2 in this case, falls to like 0,8. With hyperbolic tangent, cumulative errors which used to be 0,6 ÷ 0,8 without feedback, fall to 0,1 ÷ 0,4 with feedback on V(x_i).

The (provisional) piece of wisdom I can have as my takeaway is twofold. Firstly, whatever I do, a large chunk of perceptual learning leads to a bit less cohesion in my decisions. As I learn by experience, I allow myself more divergence in decisions. Secondly, looping on that divergence, and including it explicitly in my pattern of learning leads to relatively more cohesion at the end of the day. Still, more cohesion has a price – less learning.

I am consistently delivering good, almost new science to my readers, and love doing it, and I am working on crowdfunding this activity of mine. As we talk business plans, I remind you that you can download, from the library of my blog, the business plan I prepared for my semi-scientific project Befund  (and you can access the French version as well). You can also get a free e-copy of my book ‘Capitalism and Political Power’ You can support my research by donating directly, any amount you consider appropriate, to my PayPal account. You can also consider going to my Patreon page and become my patron. If you decide so, I will be grateful for suggesting me two things that Patreon suggests me to suggest you. Firstly, what kind of reward would you expect in exchange of supporting me? Secondly, what kind of phases would you like to see in the development of my research, and of the corresponding educational tools?

[1] De Vincenzo, I., Massari, G. F., Giannoccaro, I., Carbone, G., & Grigolini, P. (2018). Mimicking the collective intelligence of human groups as an optimization tool for complex problems. Chaos, Solitons & Fractals, 110, 259-266.

My editorial on You Tube

I keep working on the application of neural networks as simulators of collective intelligence. The particular field of research I am diving into is the sector of energy, its shift towards renewable energies, and the financial scheme I invented some time ago, which I called EneFin. As for that last one, you can consult « The essential business concept seems to hold », in order to grasp the outline.

I continue developing the line of research I described in my last update in French: « De la misère, quoi ». There are observable differences in the prices of energy according to the size of the buyer. In many countries – practically in all the countries of Europe – there are two, distinct price brackets. One, which I further designated as P_B, is reserved to contracts with big consumers of energy (factories, office buildings etc.) and it is clearly lower. Another one, further called P_A, is applied to small buyers, mainly households and really small businesses.

As an economist, I have that intuitive thought in the presence of price forks: that differential in prices is some kind of value. If it is value, why not giving it some financial spin? I came up with the idea of the EneFin contract. People buy energy from a local supplier, in the amount Q, who sources it from renewables (water, wind etc.), and they pay the price P_A, thus generating a financial flow equal to Q*P_A. That flow buys two things: energy priced at P_B, and participatory titles in the capital of their supplier, for the differential Q*(P_A – P_B). I imagine some kind of crowdfunding platform, which could channel the amount of capital K = Q*(P_A – P_B).

That K remains in some sort of fluid relationship to I, or capital invested in the productive capacity of energy suppliers. Fluid relationship means that each of those capital balances can date other capital balances, no hard feelings held. As we talk (OK, I talk) about prices of energy and capital invested in capacity, it is worth referring to LCOE, or Levelized Cost Of Electricity. The LCOE is essentially the marginal cost of energy, and a no-go-below limit for energy prices.

I want to simulate the possible process of introducing that general financial concept, namely K = Q*(P_A – P_B), into the market of energy, in order to promote the development of diversified networks, made of local suppliers in renewable energy.

Here comes my slightly obsessive methodological idea: use artificial intelligence in order to simulate the process. In classical economic method, I make a model, I take empirical data, I regress some of it on another some of it, and I come up with coefficients of regression, and they tell me how the thing should work if we were living in a perfect world. Artificial intelligence opens a different perspective. I can assume that my model is a logical structure, which keeps experimenting with itself and we don’t the hell know where exactly that experimentation leads. I want to use neural networks in order to represent the exact way that social structures can possibly experiment with that K = Q*(P_A – P_B) thing. Instead of optimizing, I want to see that way that possible optimization can occur.

I have that simple neural network, which I already referred to in « The point of doing manually what the loop is supposed to do » and which is basically quite dumb, as it does not do any abstraction. Still, it nicely experiments with logical structures. I am sketching its logical structure in the picture below. I distinguish four layers of neurons: input, hidden 1, hidden 2, and output. When I say ‘layers’, it is a bit of grand language. For the moment, I am working with one single neuron in each layer. It is more of a synaptic chain.

Anyway, the input neuron feeds data into the chain. In the first round of experimentation, it feeds the source data in. In consecutive rounds of learning through experimentation, that first neuron assesses and feeds back local errors, measured as discrepancies between the output of the output neuron, and the expected values of output variables. The input neuron is like the first step in a chain of perception, in a nervous system: it receives and notices the raw external information.

The hidden layers – or the hidden neurons in the chain – modify the input data. The first hidden neuron generates quasi-random weights, which the second hidden neuron attributes to the input variables. Just as in a nervous system, the input stimuli are assessed as for their relative importance. In the original algorithm of perceptron, which I used to design this network, those two functions, i.e. generating the random weights and attributing them to input variables, were fused in one equation. Still, my fundamental intent is to use neural networks to simulate collective intelligence, and intuitively guess those two functions are somehow distinct. Pondering the importance of things is one action and using that ponderation for practical purposes is another. It is like scientist debating about the way to run a policy, and the government having the actual thing done. These are two separate paths of action.

Whatever. What the second hidden neuron produces is a compound piece of information: the summation of input variables multiplied by random weights. The output neuron transforms this compound data through a neural function. I prepared two versions of this network, with two distinct neural functions: the sigmoid, and the hyperbolic tangent. As I found out, the way they work is very different, just as the results they produce. Once the output neuron generates the transformed data – the neural output – the input neuron measures the discrepancy between the original, expected values of output variables, and the values generated by the output neuron. The exact way of computing that discrepancy is made of two operations: calculating the local derivative of the neural function, and multiplying that derivative by the residual difference ‘original expected output value minus output value generated by the output neuron’. The so calculated discrepancy is considered as a local error, and is being fed back into the input neuron as an addition to the value of each input variable.

Before I go into describing the application I made of that perceptron, as regards my idea for financial scheme, I want to delve into the mechanism of learning triggered through repeated looping of that logical structure. The input neuron measures the arithmetical difference between the output of the network in the preceding round of experimentation, and that difference is being multiplied by the local derivative of said output. Derivative functions, in their deepest, Newtonian sense, are magnitudes of change in something else, i.e. in their base function. In the Newtonian perspective, everything that happens can be seen either as change (derivative) in something else, or as an integral (an aggregate that changes its shape) of still something else. When I multiply the local deviation from expected values by the local derivative of the estimated value, I assume this deviation is as important as the local magnitude of change in its estimation. The faster things happen, the more important they are, so do say. My perceptron learns by assessing the magnitude of local changes it induces in its own estimations of reality.

I took that general logical structure of the perceptron, and I applied it to my core problem, i.e. the possible adoption of the new financial scheme to the market of energy. Here comes sort of an originality in my approach. The basic way of using neural networks is to give them a substantial set of real data as learning material, make them learn on that data, and then make them optimize a hypothetical set of data. Here you have those 20 old cars, take them into pieces and try to put them back together, observe all the anomalies you have thus created, and then make me a new car on the grounds of that learning. I adopted a different approach. My focus is to study the process of learning in itself. I took just one set of actual input values, exogenous to my perceptron, something like an initial situation. I ran 5000 rounds of learning in the perceptron, on the basis of that initial set of values, and I observed how is learning taking place.

My initial set of data is made of two tensors: input T_I and output T_O.

The thing I am the most focused on is the relative abundance of energy supplied from renewable sources. I express the ‘abundance’ part mathematically as the coefficient of energy consumed per capita, or Q/N. The relative bend towards renewables, or towards the non-renewables is apprehended as the distinction between renewable energy Q_R/N consumed per capita, and the non-renewable one, the Q_NR/N, possibly consumed by some other capita. Hence, my output tensor is T_O = {Q_R/N; Q_NR/N}.

I hypothesise that T_O is being generated by input made of prices, costs, and capital outlays. I split my price fork P_A – P_B (price for the big ones minus price for the small ones) into renewables and non-renewables, namely into: P_A;R, P_A;NR, P_B;R, and P_B;NR. I mirror the distinction in prices with that in the cost of energy, and so I call LCOE_R and LCOE_NR. I want to create a financial scheme that generates a crowdfunded stream of capital K, to finance new productive capacities, and I want it to finance renewable energies, and I call K_R. Still, some other people, like my compatriots in Poland, might be so attached to fossils they might be willing to crowdfund new installations based on non-renewables. Thus, I need to take into account a K_NR in the game. When I say capital, and I say LCOE, I sort of feel compelled to say aggregate investment in productive capacity, in renewables, and in non-renewables, and I call it, respectively, I_R and I_NR. All in all, my input tensor spells T_I = {LCOE_R, LCOE_NR, K_R, K_NR, I_R, I_NR, P_A;R, P_A;NR, P_B;R, P_B;NR}.

The next step is scale and measurement. The neural functions I use in my perceptron like having their input standardized. Their tastes in standardization differ a little. The sigmoid likes it nicely spread between 0 and 1, whilst the hyperbolic tangent, the more reckless of the two, tolerates (-1) ≥ x ≥ 1. I chose to standardize the input data between 0 and 1, so as to make it fit into both. My initial thought was to aim for an energy market with great abundance of renewable energy, and a relatively declining supply of non-renewables. I generally trust my intuition, only I like to leverage it with a bit of chaos, every now and then, and so I ran some pseudo-random strings of values and I chose an output tensor made of T_O = {Q_R/N = 0,95; Q_NR/N = 0,48}.

That state of output is supposed to be somehow logically connected to the state of input. I imagined a market, where the relative abundance in the consumption of, respectively, renewable energies and non-renewable ones is mostly driven by growing demand for the former, and a declining demand for the latter. Thus, I imagined relatively high a small-user price for renewable energy and a large fork between that P_A;R and the P_B;R. As for non-renewables, the fork in prices is more restrained (than in the market of renewables), and its top value is relatively lower. The non-renewable power installations are almost fed up with investment I_NR, whilst the renewables could still do with more capital I_R in productive assets. The LCOE_NR of non-renewables is relatively high, although not very: yes, you need to pay for the fuel itself, but you have economies of scale. As for the LCOE_R for renewables, it is pretty low, which actually reflects the present situation in the market.

The last part of my input tensor regards the crowdfunded capital K. I assumed two different, initial situations. Firstly, it is virtually no crowdfunding, thus a very low K. Secondly, some crowdfunding is already alive and kicking, and it is sort of slightly above the half of what people expect in the industry.

Once again, I applied those qualitative assumptions to a set of pseudo-random values between 0 and 1. Here comes the result, in the table below.

Table 1 – The initial values for learning in the perceptron

The way the perceptron works means that it generates and feeds back local errors in each round of experimentation. Logically, over the 5000 rounds of experimentation, each input variable gathers those local errors, like a snowball rolling downhill. I take the values of input variables from the last, i.e. the 5000^th round: they have the initial values, from the table above, and, on the top of them, there is cumulative error from the 5000 experiments. How to standardize them, so as to make them comparable with the initial ones? I observe: all those final output values have the same cumulative error in them, across all the T_I input tensor. I choose a simple method for standardization. As the initial values were standardized over the interval between 0 and 1, I standardize the outcoming values over the interval 0 ≥ x ≥ (1 + cumulative error).

I observe the unfolding of cumulative error along the path of learning, made of 5000 steps. There is a peculiarity in each of the neural functions used: the sigmoid, and the hyperbolic tangent. The sigmoid learns in a slightly Hitchcockian way. Initially, local errors just rocket up. It is as if that sigmoid was initially yelling: ‘F******k! What a ride!’. Then, the value of errors drops very sharply, down to something akin to a vanishing tremor, and starts hovering lazily over some implicit asymptote. Hyperbolic tangent learns differently. It seems to do all it can to minimize local errors whenever it is possible. Obviously, it is not always possible. Every now and then, that hyperbolic tangent produces an explosively high value of local error, like a sudden earthquake, just to go back into forced calm right after. You can observe those two radically different ways of learning in the two graphs below.

Two ways of learning – the sigmoidal one and the hyper-tangential one – bring interestingly different results, just as differentiated are the results of learning depending on the initial assumptions as for crowdfunded capital K. Tables 2 – 5, further below, list the results I got. A bit of additional explanation will not hurt. For every version of learning, i.e. sigmoid vs hyperbolic tangent, and K = 0,01 vs K ≈ 0,5, I ran 5 instances of 5000 rounds of learning in my perceptron. This is the meaning of the word ‘Instance’ in those tables. One instance is like a tensor of learning: one happening of 5000 consecutive experiments. The values of output variables remain constant all the time: T_O = {Q_R/N = 0,95; Q_NR/N = 0,48}. The perceptron sweats in order to come up with some interesting combination of input variables, given this precise tensor of output.

Table 2 – Outcomes of learning with the sigmoid, no initial crowdfunding

Table 3 – Outcomes of learning with the sigmoid, with substantial initial crowdfunding

Table 4 – Outcomes of learning with the hyperbolic tangent, no initial crowdfunding

Table 5 – Outcomes of learning with the hyperbolic tangent, substantial initial crowdfunding

The cumulative error, the first numerical line in each table, is something like memory. It is a numerical expression of how much experience has the perceptron accumulated in the given instance of learning. Generally, the sigmoid neural function accumulates more memory, as compared to the hyper-tangential one. Interesting. The way of processing information affects the amount of experiential data stored in the process. If you use the links I gave earlier, you will see different logical structures in those two functions. The sigmoid generally smoothes out anything it receives as input. It puts the incoming, compound data in the negative exponent of the Euler’s constant e = 2,72, and then it puts the resulting value as part of the denominator of 1. The sigmoid is like a bumper: it absorbs shocks. The hyperbolic tangent is different. It sort of exposes small discrepancies in input. In human terms, the hyper-tangential function is more vigilant than the sigmoid. As it can be observed in this precise case, absorbing shocks leads to more accumulated experience than vigilantly reacting to observable change.

The difference in cumulative error, observable in the sigmoid-based perceptron vs that based on hyperbolic tangent is particularly sharp in the case of a market with substantial initial crowdfunding K. In 3 instances on 5, in that scenario, the hyper-tangential perceptron yields a negative cumulative error. It can be interpreted as the removal of some memory, implicitly contained in the initial values of input variables. When the initial K is assumed to be 0,01, the difference in accumulated memory, observable between the two neural functions, significantly shrinks. It looks as if K ≥ 0,5 was some kind of disturbance that the vigilant hyperbolic tangent attempts to eliminate. That impression of disturbance created by K ≥ 0,5 is even reinforced as I synthetically compare all the four sets of outcomes, i.e. tables 2 – 5. The case of learning with the hyperbolic tangent, and with substantial initial crowdfunding looks radically different from everything else. The discrepancy between alternative instances seems to be the greatest in this case, and the incidentally negative values in the input tensor suggest some kind of deep shakeoff. Negative prices and/or negative costs mean that someone external is paying for the ride, probably the taxpayers, in the form of some fiscal stimulation.

I am consistently delivering good, almost new science to my readers, and love doing it, and I am working on crowdfunding this activity of mine. As we talk business plans, I remind you that you can download, from the library of my blog, the business plan I prepared for my semi-scientific project Befund  (and you can access the French version as well). You can also get a free e-copy of my book ‘Capitalism and Political Power’ You can support my research by donating directly, any amount you consider appropriate, to my PayPal account. You can also consider going to my Patreon page and become my patron. If you decide so, I will be grateful for suggesting me two things that Patreon suggests me to suggest you. Firstly, what kind of reward would you expect in exchange of supporting me? Secondly, what kind of phases would you like to see in the development of my research, and of the corresponding educational tools?