This is where I attempt to pull together all the different ideas from the first lecture into one post. This is mainly for myself, but you never know who else will find it interesting…
Reading back on my notes it all seems a bit disjointed, but let’s see how this goes. This is just my understanding of what we heard so it might be wrong or misguided in places, but as we were told yesterday if you never make any mistake you will never learn… I’m sure someone will set me straight.
I guess the place to start is with what is the goal of CSIM and that is: How do we introduce technology into society in a way that it works for everyone? One good example of this not being fulfilled is mobile phones. They may be very powerful and have hundreds of features, but very often these are buried in so many menus and submenus that unless you read the manual cover to cover you will not be able to use it to its full potential. This may seem a bit of a trivial example, but one familiar to many people.
One proposal to solve this is to build technology that is sentient and therefore understands our needs and can fulfill them for us without the need for special expertise on how to interact with these machines. This would in turn make it accessible to everyone.
However, in order for us to build sentient machines we first need to understand what it means to be sentient. Back in 200AD the philosopher Plotinus came up with this question: “And we, who are we anyhow?” This question is still not understood and has remained unanswered to date. Now, how does building sentient machines help us to answer Plotinus question?
I’m going to park that for a minute (mainly because my notes from the lecture park that for a while) and talk about the two views on how we can achieve the goal of building sentient machines. The two views being:
- Comptutational
- Biological
The computational view is mainly being championed by a guy called Ray Kurzweil. He believes that there are 6 epochs of evolution that humanity has (we’re currently in Epoch 4) and will go through. This is best explained by the below image:
Ray Kurzweil's 6 Epochs of Evolution
As far as I can gather from the little reading I’ve done on this, he essentially believes that evolution into the next epoch will be driven by computational advances which will enable us to master the methods of biological evolution and use them to evolve ourselves in conjunction with technology.
The biological view centres around the brain and the idea that in order to “replace” a physical, biological brain, you first need to fully understand it. This could be through building a physical model, as was done many years ago when DNA’s molecular structure was discovered, or perhaps some other way we haven’t yet thought of.
The
SPECS (Synthetic Perceptive, Emotive and Cognitive Systems) research group at UPF is working towards the goal of building “robots” to insert into society so that they can co-exist with and help humans to improve their welfare and quality of life. There are 5 aspects in which welfare can be improved by sentient machines:
- Physical environments
- Help with natural/man made disasters: use of sentient machines to understand what needs to be done and execute it. When the Japan earthquake struck, there was a need to go into the Fukushima nuclear plant and take a radioactivity reading. Simple as that might sound, there was no technology “sophisticated enough” to go in, take the reading and come out. Something that would be a very simple task to a human but physically impossible due to the environment.
- Economic environments
- Production and manufacturing capabilities: We have lost most of our manufacturing capabilities in Europe due to pressures from much cheaper goods produced outside of Europe. These in turn will eventually increase in price as well. We can turn to technology in order to produce goods cheaply throughout the world
- Urban environments
- Transportation systems: Sensors in roads, traffic lights, etc… to optimise traffic flow
- Social environments
- Aging Populations: “Robots” to help care for and augment capabilities of elderly people
- Private environments
- Human life and capabilities: Extend and augment these through technology
One of the goals of SPECS is to solve and address these issues in a practical sense through the use of sentient machines as an enabling technology for sustainable welfare.
Taking a huge step back to the beginning, in order to understand sentience and how that came about, we need to think about the 3 BIG questions:
- What is the origin of matter and the universe?
- What is the origin of life?
- What is the origin of consciousness?
I’m going to concentrate on the question of consciousness as that is what we discussed in class. One way to try to understand consciousness could be to build machines of different levels of complexity (from microbes, to insects, to vertebrates, etc…) in order to approximate the different levels of complexity found in nature. This could in theory allow us to understand how evolution arrived at consciousness for humans. Since evolution conserves organizational principles (50% of DNA shared between all living organisms,) if we can understand the design principles of biological systems we should be able to re-create them.
This creates the following positive feedback loop:
- If we can Understand nature by
- Building machines that mimic it
- this will allow us to Repair and Augment human capabilities
- which will in turn give us more insight and Understanding into nature
- and so on and so forth…
This all leads us to the fact that we need to understand what is a mind in order to be able to build one. The problem is that the “mind” is a label that we humans have invented and placed on a particular phenomenum that we experience. However, that label could be inexact and might not be identical to the actual phenomenum (or could in fact be a collection of phenomena.) The fact is that the concept to describe mind has been invented at some point. However, just because we universally use a particular concept, does not make it “true.” How can we know what aspects of reality these concepts actually capture?
There is a guy called Julian Jaynes that came up with the following hypothesis through studying ancient texts. According to him, 3000 years ago humans lived in a schizophrenic state where their behaviours resulted from instructions by “external voices” which they called Gods. This was due to the fact that they had bicameral brains, in which cognitive functions were split between one half which commanded (the external voices or gods) and a half that followed those commands. Studying texts from before 1000BC he found that there were no literary references to cognitive processes or evidence of self awareness, which meant that humans felt like they were at the mercy of the gods and merely pawns without free will. All their actions were controlled by these external voices. However, from 1000BC onwards, humans start to rely on instrospection and consciousness and “Gods” become external actors that can influence human destiny but are not directly controlling human’s wills.
Still, the brain is a story teller and will come up with reasons why its “owner” has made a particular decision, when in fact it could all be down to instinct, so maybe we haven’t come so far into self awareness yet.
Well, there you go. Quite disjointed and probably full of gaps but this is what I absorbed from the first lecture. It’s become clear that I have a lot of reading and thinking to do in the next few days and I’m really looking forward to it!
I leave you with some relevant material from this first lecture:
Homework
Read the paper: “Unskilled and Unaware of It: How Difficulties in Recognizing One’s Own Incompetence Lead to Inflated Self-Assessments” by Justin Kruger and David Dunning, find another matching paper (to support or refute it) and write a 1 page essay on both.
Reading list
- Plotinus
- The Singularity is near – Ray Kurzweil
- Do Androids Dream of Electronic Sheep – Philip K Dick
- In fact, anything by Philip K Dick (which I was planning on doing anyway before I started this Masters and now it seems too much like homework…)
- The Serpent and the Rainbow – Wade Davis
- The Ego Tunnel – Thomas Metzinger
- The Theory of Affordances – James J. Gibson
- The Illusion of Conscious Will – Daniel M. Wegner
- The Origin of Consciousness in the Breakdown of the Bicameral Mind – Julian Jaynes
Links
