DNA and AI

The human genome sits at roughly 700 MB of data if converted from base pairs to binary data, or, to put another way you have the entire instruction set for the creation of the most adaptable creature we have ever encountered contained in the same space as a CD. This sort of sticks out in the filed of developing AI. Current methods seem fixed on the idea of having enormous databases instructing what the AI should say and do. Is this correct? If the goal of creating AI is to simulate a human like mind, then how can the answer possibly be to use terabytes of information to create it?

It appears that evolution has come up with a clever way of creating intelligence that doesn’t require much information on the actual world the creature is being put into. Our brains can adapt to such a wide range of environments, languages, talents, and situations. Nature has programmed the base information for this incredible machine with a fraction of the 7o0 MB defining the human. Nature and lazy coders like to copy as much as it possibly can in order to be more efficient. It is not by chance that a majority of things in nature have structures defined by recursive functions. A trees branch is self similar to the rest of the tree. Copy and paste. By defining recursive and fractal like functions nature has defined complex structure and solutions with elegant and simple lines of code requiring little storage.

Those familiar with machine learning will be quick to point out the similarities of neural networks and the human brain. In my opinion, no matter how well we define a neural network in code it cannot achieve  the complexities of reality. Reality, despite our best efforts, has a lot of chaos that we normally throw out as noise. But when a computer becomes physical it can utilize the chaotic nature of reality. For instance, using genetic algorithms to configure a 100 Logic Gate FPGA to identify a 1 kHz and 10 kHz tone yielded a chip that worked with some completely bizarre settings. The chip had circuits not even connected to the actual circuit, but when those were changed the chip no longer worked. That’s like your car not working because your neighbor is out of gas… Sort of. Similarly, when the configuration was loaded onto another FPGA it failed. This seems to indicate the randomly evolved circuit was using the EMF interactions of the circuit at a level no one would ever intentionally design for.

We seem obsessed with the idea that we can create a simulation of the brain if we simply scan in the geometry of the human mind and click go. Referring back to chaos, our brain does not exist in a vacuum. Our brain swims in a bath of chemicals and happens to be attached to a rather busy body, but so what? It seems that beyond the obvious chemical interactions of exercise and food even that bacteria in our gut has an effect on our brains. Accurately simulating a human mind would require taking into account all the interactions of the environment, the chemicals in our mind, bacteria, social interactions, physical cues, and those are simply the ones we know about.

Being a lazy programmer myself, I wouldn’t want to program an AI using this approach. I want to know how nature does it. I wish I could just copy the code from DNA, but that will be covered in a future post as being near impossible. A more in depth look into the possible AI revelations based on DNA will come in the next few weeks. I just felt compelled to delve into the fascinatingly limited amount of data the worlds best intelligence is derived from.

 

 

The Cost of Random

Last post discussed the inherent issues associated with evolution and, by extension, human life. The issues don’t stop at the limitations inherent by our basic design. Whenever I hear stories of people who deny the benefits of immunization, or people who spit in the face of science in general I bear in mind one of the core concepts of evolution. Evolution does not have an agenda. It produces truly random changes that do not have any kind of direction. Natural selection culls these random changes into meaningful improvements over time, but when looking at the changes in a single generation evolution creates some interesting results. Obvious examples include the variety of birth defects that occur in the general population. A series of mutations leads inevitably to an occurrence of these. Mutations led to something that does not, in the context of the world as we know it, help the individual perform better  in life based on some metrics. Remember that “better” is always based on some kind of assumption. For evolution, that assumption is the individual’s ability to reproduce, but I know plenty of reproducing people I would not wish to emulate, so each person has a different “better”.

I believe willful ignorance to be a part of the equation of life. Society operates like an extremely quick version of evolution. It changes and grows in ways that sometimes produce positive and negative results and has far more fluctuation that actual evolution proper. As a result you end up with a group that shares 99.99% of DNA but has an extremely diverse system of beliefs. The fact that society produces people of extreme beliefs must have benefited the human race at some point. Clearly people denying the benefits of immunization does nothing but harm, but at one point in the history of humanity a similar deviation from the norm led to advances.  Nicolaus Copernicus held completely controversial beliefs, but he held his non-conventional beliefs in real data. He had this little idea called the heliocentric universe that he published before he died. His finding sparked one of the greatest changes in how people thought in history, but at the time I am sure his discoveries caused him nothing but stress, agony, and worry about reprisal.

I see these people as a random solutions to the problem of life, but unlike Copernicus, their experience does not revolve around accurate data. Almost all people feel they do what they believe to be correct. From terrorist to teachers, and wall-street execs to janitors, people often consider their path to be rational and right. I personally cannot rationalize what barbaric things terrorists do, but only a few humans in the world are actually hardwired without emotion, one percent of the general population. So, somehow, someway, the equation of their lives built up to a point where strapping on a vest of explosives and walking into a crowded building seemed rational and right. These people, no matter how insanely violent or murderous, experience the entire range of human emotion. It’s comforting to dehumanize people we consider evil. The idea that Hitler was a cackling psychopath with nothing but the intention of killing in his heart simplifies evil down to something most people can manage. Trying to truly understand how an entire county of people could fall under the will of a single misguided man with almost nothing more than the ability to deliver good speech still baffles me. Putting these thoughts in the context of evolution allows for a somewhat unbiased analysis. Natural selection works hardest during difficult periods of time, drought, famine, meteor strikes, or similar. When drastic changes in the environment happen the surviving population will often be the fringe of the general population. A meteor most likely opened up an environment for mammals to become the dominant beings on the planet. Suddenly being a giant reptile lost its benefits. Post ww1 Germany represents an extremely difficult time for a specific population. One of the go to solutions for societies experiencing times of extreme difficulty involve putting all blame on “the others.” People function well with a solid picture of an antagonist. We like to put a face on the picture of what we consider to be evil. The fringe thought that somehow all of the Jewish people were the source of all of German’s problems suddenly became so appealing because people like Hitler had so much ammo in terms of economic and social issues. In theory almost any minority group could have been singled out as the problem.  Just like evolution, these difficult times led to some with fringe beliefs to perform “better.” Again, “better” in terms of a people suffering from horrid economic conditions looking for something, anything, to believe in that promised a better future.

This is meant to exemplary, clearly the intricacies of such a complicated issue cannot be boiled down to a simple explanation, but such analysis helps in understanding how something like this might happen again. Truly random answers to problems facing evolution and societies lead to creative and innovate solutions, but can often have unexpected costs and limitations. Creative solutions like the scientific theory have grown society beyond what it could have ever accomplished by staying entrenched in the legacy belief structures. Creativity swings in both directions as the leadership of the strong-willed can control even the most intelligent of us all. I take away from this analysis that all assumptions deserve questioning. No matter how clever we think “good” players in the world become the greedy and self-centered will always guise as a positive player as well. It happens now. Corporations and the wealthy pump the world with bad science, manipulated statistics, and publish garbage papers. The evolution for superiority wont end with just the scientific theory, we have a long way to go as a species.

Life as a Program

Putting life in the context of programming, mathematics, and science allows for some interesting insights. One of the more curious revelations that comes from this analysis revolves around genetic algorithms. Genetic algorithms attempt to function like evolution. These algorithms work by creating a random population of solutions to a problem within a certain set of conditions. Consider a program designed to play the game pong. Pong has inputs, outputs, a set of rules, and a way to judge how well a player does based on a score. A genetic algorithm would first guess a population of solutions to the problem of how to win pong based on all the variables available to the program. The first population of solutions would be horrible, just consider the challenge of purely guessing how to play a game. After testing all of the population and giving them a score a certain amount of the better performers would “breed” or cross over. The two solutions would blend together and mix attributes of each of them. Then, a well written program will also mutate the solution, changing a variable slightly from the previous guess. The best performers will continue on to the next generation of solutions. The program will then generate another random population of solutions and the process will continue. Because the best performers “breed”  and carry on into the next generation the populations slowly become better and better at the game as generations continue. Eventually the program will create something that will win a majority of the time, but there exists a problem in this method. The earliest successful solutions steer the whole population in a direction from which there is no coming back from.

Evolution functions just like these algorithms, and suffer from the same core issue. Consider the skeletal structure for vertebrates. Almost all vertebrates share the most basic design for a skeleton. Before the concept of a skeleton existed evolution guessed at millions upon millions of solutions for the problem of  creating a structure for a creature. There were probably solutions using different materials, asymmetrical patterns, and an overwhelming majority were probably as terrible as a computer that is guessing how to play pong. Eventually, slowly, randomly, the symmetrical calcium based structure we now recognize in creatures ranging from blue whales to mice came into existence. It works incredibly well, and has many advantages, but is it the best design there will ever be? I would guess no. Somewhere in this universe exists a better form and function of skeletal structure than ours. A solution heavily depends on the problem of course, so for this it is assumed that the alien world has similar conditions to earth. Unfortunately, humanity will never get a shiny new skeletal design. Our skeleton represents a local maximum in our design, and the mystery alien skeleton of perfect design represents a global maximum.

How can I possibly say we will never have a skeleton made of metal with super healing abilities?  Because evolution doesn’t work well in reverse and rarely makes dramatic changes. Barring genetic manipulation, our species will continue to have a skull, a single heart, a brain, two eyes, and so on. Change happens slowly and in order to reach a global maximum evolution would have to go backwards for a period of time. Our design would need to get worse before it got better, and the basic function of evolution doesn’t support getting worse at survival. Whatever limitations stem from having our design will continue forever. Some other planet with a different set of conditions will produce a completely different set of species from us. Forget the Star Trek humanoid assumption we have for a creature, these sorts of begins would be fundamentally unique and would not have the same limitations we have. I have always had assumptions about evolution leading to better and better outcomes, but reality shows that evolution leads to the best solution given a set of assumptions, and at its core cannot overcome the flaws in the basic design.

Our lives have a certain similarity to evolution and genetic algorithms. Ignoring for a moment human nature and instinct it is possible to look at life through the lens of evolution. In childhood I essentially guessed how to be a person. I had a limited set of assumptions about the world, I had inputs into my life including, parents, education, siblings, and other factors. There were outputs in the form of experiences, scars, pains, joy, and happiness. So my brain continued evolving to solve the problem of living in this world based on my random inputs. Unfortunately my methods suffer from the same issues as evolution. My earliest solutions shaped my mind into something that will most likely not change. Language represents this concept most clearly. I could learn another language, but it will get harder for me to learn a language before it gets easier. I would have to put in considerable effort, go against my nature, experience stress and a lack of comfort, and no matter how hard I tried I would never dissolve the English portion of my brain and the assumptions that come from that.

Essentially, I believe that because of the similarity between life and evolution it would make sense that I am living at a local maximum for what I can become. I have a certain set of core assumptions, some of which will be wrong, that limit what I can accomplish. This makes me question a lot. What would a global maximum version of myself be? What could I possibly do to reach that? How difficult might it be to reach this? I have one thing over evolution, I can go against what feels right in my mind. Evolution shows that change is slow, uncomfortable, stressful, and often gets worse before it gets better. In order to try and get closer to my global maximum I must question my assumptions, especially more most core assumptions about how I solved the problem of living. This can be done by making small, uncomfortable changes and seeing the results. One such uncomfortable change is the creation of this blog, and the continued analysis of minor changes in life. I have always enjoyed writing, but I feel a great deal of stress doing so with the intent of a human being actually reading it.

So I start this with the intent of analyzing life in the context evolution with the goal of finding TheGlobalMax.