Here’s one way to evolve an artificial intelligence

This picture illustrates an idea for how to evolve an AI system. It is derived from the sensor-brain-actuator-world model.

Machine learning algorithms have been doing some impressive things. Simply by crawling through massive oceans of data and finding correlations, some AI systems are able to make unexpected predictions and reveal insights.

Neural nets and evolutionary algorithms constitute a natural pairing of technologies for designing AI systems. But evolutionary algorithms require selection criteria that can be difficult to design. One solution is to use millions of human observers as a Darwinian fitness force to guide an AI towards an optimal state.

Clarifications

Since there is so much discussion (and confusion) on AI these days I want make some clarifications.

  • This has nothing to do with consciousness or self. This AI is disembodied.
  • The raw data input is (not) curated. It has no added interpretation.
  • Any kind of data can be input. The AI will ignore most of it at first.
  • The AI presents its innards to humans. I am calling these “simulations”.
  • The AI algorithm uses some unspecified form of machine learning.
  • The important innovation here is the ability to generate “simulations”.

Mothering

The humanist in me says we need to act as the collective Mother for our brain children by providing continual reinforcement for good behavior and discouraging bad behavior. As a world view emerges in the AI, and as an implicit code of morals comes into focus, the AI will “mature”. Widely-expressed fears of AI run amok could be partially alleviated by imposing a Mothering filter on the AI as it comes of age.

Can Anything Evolve without Selection?

I suppose it is possible for an AI to arrive at every possible good idea, insight, and judgement just by digesting the constant data spew from humanity. But without an implicit learning process (such as back-propagation and other feedback mechanisms used in training AI), the AI cannot truly learn in an ecosystem of continual feedback.

Abstract Simulations 

Abstraction in Modernist painting is about generalizing the visual world into forms and colors that substitute detail for overall impressions. Art historians have charted the transition form realism to abstraction – a kind of freeing-up and opening-up of vision.

Imagine now a new path leading from abstraction to realism. And it doesn’t just apply to images: it also applies to audible signals, texts, movements, and patterns of behavior.

Imagine an AI that is set up like the illustration above coming alive for the first time. The inner-life of newborn infant is chaotic, formless, and devoid of meaning, with the exception of reactions to a mother’s smile, her scent, and her breasts.

A newborn AI would produce meaningless simulations. As the first few humans log in to give feedback, they will encounter mostly formless blobs. But eventually, some patterns may emerge – with just enough variation for the human judges to start making selective choices: “this blob is more interesting than that blob”.

As the young but continual stream of raw data accumulates, the AI will start to build impressions and common themes, like what Deep Dream does as it collects images and finds common themes and starts riffing on those themes.

http://theghostdiaries.com/10-most-nightmarish-images-from-googles-deepdream/

The important thing about this process is that it can self-correct if it starts to veer in an unproductive direction – initially with the guidance of humans and eventually on its own. It also maintains a memory of bad decisions, and failed experiments – which are all a part of growing up.

Takeaway

If this idea is interesting to you, just Google “evolving AI” and you will find many many links on the subject.

As far as my modest proposal: the takeaway I’d like to leave you with is this:

Every brain on earth builds inner-simulations of the world and plays parts of those simulations constantly as a matter of course. The simple animals have extremely simple models of reality. We humans have insanely complex models – which often get us into trouble. Trial simulations generated by an evolving AI would start pretty dumb, but with more sensory exposure, and human guidance, who knows what would emerge!

It would be irresponsible to launch AI programs without mothering. The evolved brains of most complex mammals naturally expect this. Our AI brain children are naturally derived from a mammalian brain. Mothering will allow us to evolve AI systems that don’t turn into evil monsters.

Having sex with robots to save the planet

Long long ago, there was an accident in a warm puddle. A particular molecule – through some chance interaction with the soup of surrounding molecules – ended up with a copy of itself. Since the surrounding soup was similar to the original, the copy was more likely to replicate itself. And so it did. The rest is history. We call it evolution.

It is possible that similar accidents happened elsewhere around the same time – not just in one single puddle. One could also say that variations of this accident are still happening – only now at a massive scale.

Every act of every living thing can be seen as an elaboration of this original act. Self-replication is the original impetus of all life. We share a common ancestor with amoebas – who replicate asexually. The invention of sexual reproduction boosted genetic creativity. More recently in the scope of Earth’s history, creativity escaped the confines of genetics. We humans are the primary hosts of this creative engine.

Human beings have contrived all of the resulting aspects of survival to an art-form. This includes – not just the act of sex – but also the act of preparing food (cuisine), the act of making sounds and speaking (music and singing), and the act of altering the environment to create new structure (visual art). The abstractions and representations of the world that the brain generates via the body are derivations and deviations from the original acts of survival. It’s a form of self-replication.

The emergence of abstractions, mental models, and representations is increasing in complexity. This is an inevitable one-way blossoming accelerated by the emergence of the animal brain. The human experience is conflicted; we are oriented toward achieving escape velocity from Original Nature, but we also long for Original Nature. How can we resolve this conflict?

The original act of self-replication has powerful repercussions – billions of years after the original accident – it has taken on many forms. It is the reason we humans have strange phenomena like orgasm. And selfies.

Warming

We are at a crossroads in the history of life on Earth. The current era of global warming is almost certainly the result of the overpopulation and hyperactivity of humans, who have released – and continue to release – too much carbon into the atmosphere. One effective solution to global warming would be to reduce the primary agents of the fever…to reduce human population.

And so, converting that original act of replication into works of art is not just creative and exciting: it may be necessary. Humans must transcend the Earthly act of self-replication in order to preserve the health of the planet.

The future of sex will be…let’s just say…interesting. Every cell in our body contains the blueprint of a desire to replicate. Nature and society are structured around the elaborate machinery that has emerged to ensure self-replication – of human bodies and culture. This desire has made its mark on every aspect of society – even if we don’t recognize it as such. We cannot escape it. And so we need to virtualize it, because self-replication of human beings (physically) has become a threat to the planet that sustains us. It’s our duty to Mother Earth.

I am a living organism and so I have to contend with this crazy desire to replicate. Note: I am childless. I have never replicated my genes and have no intention to do so at this stage in my life. But I am passionate about replicating ideas, art, words, and software.

Now, what about the title of this blog post? Will people eventually start having sex with robots? It will certainly be more subtle than that. In fact, it has been said that by the time we get to that point, WE will be the robots.

Is this the kind of future I want? Strangely, yes. Because I will have long returned to the Earth – my molecules will have been handed down through generations of living things. I will be a part of Earth’s physiology. My tribe will be bigger than humanity.

One of my molecules may even end up in a warm puddle somewhere.

Virtual Reality is Biologically Inevitable

Musicophilia_front_coverI am reading Oliver Sacks’ Musicopholia. He discusses several patients he has seen who suddenly become obsessed with music, or suddenly start hearing music in their head as a result of a brain injury. These are called “musical hallucinations”. He describes temporal-lobe epilepsy patients who have musical hallucinations just before a seizure. Fascinating stuff.

It reminds me of how our brains are in the habit of “playing” things – not just music, but scenarios, stories, past experiences, and experiences we wish we could have.

virtual_reality_helmetThe term “Virtual Reality” is usually accompanied by high-tech images of people with clunky things stuck on their heads.

But there is another way to understand virtual reality: it is an inevitable fact of biological evolution on Earth.

What? Virtual reality is more than just a technological innovation?  A gimmick? Yes. Absolutely. Virtual reality has its roots in the early formation of life on Planet Earth.

Years ago I read Daniel Dennett’s book Kinds of Minds. I remember looking at diagrams of how animals form internal representations of the external world.

089a

Dennett shows how the evolution of nervous systems gave way to brains and ultimately consciousness. And along the journey, internal representations became increasingly sophisticated and better at predicting the outcomes of potential actions.

Throughout the history of biological evolution, animal brains became increasingly complex and adaptive to the complexity of the environment (which itself became more complex because of the brains of other animals…and so on). From genetic adaptation … to consciousness: all animals build internal representations of the word in order to function within it. This might be considered the very basis – the original impetus – of intelligence.

images

There was an amazing discovery that I leaned about from reading On Intelligence: The neocortex at the top of the human brain sends information DOWN, as well as information being sent UP from the senses. In other words, while the senses are passing information from the ears, eyes, fingers, etc. up to the higher levels of the brain, the higher levels of the brain are also sending down “expectations” of what might be coming up.

To put it another way. The brain is constantly projecting an internal virtual reality and checking to see if this matches up with the signals coming in.

Picture1

If everything matches up, and if the colliding signals are in agreement, the brain interprets this to be “business as usual”. But if any differences are detected, then various neural networks kick into action, and attempt to process this difference. This may seem strange if you are used to thinking of the brain as a passive recipient of information from the senses.

But consider this: it would be very inefficient to try to soak up the entire gamut of high-resolution reality as it floods-in through the senses. Who has time for that? It is more efficient to run an internal virtual reality based on expectation in parallel with actual reality and only jump into action with something doesn’t match up. Apparently, the six layers of neocortex described in the book are in the business of doing just that. And the higher the cortical layer, the more abstract the processing.

Think about it: the higher-region of the brain is projecting as much virtual reality down toward the senses as the senses are sending signals up to the higher regions of the brain.

Thus: the brain is a virtual reality engine.

And the collective of all animal brains have a major impact on the environment. It’s a feedback loop. The biosphere is a gigantic feedback loop of internal representations, which constantly change reality and subsequently adapt to it.

This massive cross-projection of multiple virtual realities within the biosphere started even before there were animals with brains. One could say that biological evolution has always been in the business of mapping reality into various internal representations – stored in the genes of organisms – as well as in the extended phenotypes that adorn the environment. Human brains are just the most sophisticated version of the self-reflection that emerges from the fabric of the biosphere.

Screen Shot 2015-04-02 at 12.21.09 PMSo, consider the musical hallucinations that Oliver Sacks describes. Consider the unfortunate individuals who fall victim to schizophrenia. Consider the anxiety of playing out the evening’s events before your first date. These are internal virtual realities gone awry.

 

Sony-virtual-reality-headsetWhen I see images of people with big-ass chunks of technology stuck on their faces, I wonder whats going on in the scope of the big picture – in terms of the evolution of brains. Is our internal virtual reality not sufficient enough? Is technological virtual reality just a continuation of the human instinct to tell stories, paint pictures, make movies, and games?

Perhaps the evolution of virtual reality is just that: a continuation of something that we have been doing since we became human: extending our inner-virtual reality with more and more artificial layers on the outside.

Humans are not content with plain old “natural” virtual reality. We have to take it to extremes. And given that we are not content with reality as it is (both internal and external), I guess it’s inevitable.