by Mathijs Romans

Artificial intelligence considered dangerous

December 22nd, 2011 Posted in Science | Comments Off

There are of course a great number of books and movies about the dangers of robots taking over the world. This kind of science fiction is generally not taken seriously or considered relevant for current AI research. This is mostly due to the improbable nature of the scenario’s. The possibility of creating some kind of computer with some kind of elaborate philosophy, one that we can sit next to and have a conservation with, is indeed quite remote.

Yet I consider AI research a possible threat to humanity that we should definitely be aware of. To explain I first have to talk about weak AI and strong AI. Weak AI, also called applied AI, is artificial intelligence applied for a narrow and fixed purpose, e.g. face recognition or data mining. Strong AI is intended to be more broad and general purpose, much like human intelligence. Whether that includes self-awareness is not relevant for this post. Weak AI has made steady progress and is being used today in many places. The development of strong AI has been slow and so far mostly without results.

One important issue of strong AI is whether it can be made sufficiently smart to exceed human intelligence. In fact, it is a stated research goal to investigate this very question. It is obvious that whenever we are able to make an AI that exceeds our capabilities, that AI should also be able to make an AI that exceeds its own. Since our computers are still exponentially improving by the decade, it seems likely we will face this situation, and it can happen in our lifetimes. paperclip, photo by aquapanini on Flickr

There is a final and important ingredient to my point. I used to consider the situation above worrying but not critical, since an AI will not likely have the motivation to improve itself or cause any harm. The well known three laws of robotics and their shortcomings should never come into play, since an AI in principle lacks all the different motivations that we humans have. But I have come across a compelling argument that changes this. The argument is that an AI must always have some purpose, otherwise it will not do anything. And it is in fact quite probable that any purpose, however simple, may end up in a catastrophe.

A simple example was thought up by Nick Bostrom, namely a computer that will maximize its reward by the number of paperclips it can make. We hook it up to a 3D printer and see what happens. I quote Anders Sandberg:

At first nothing seems to happen: the AI zooms through various possibilities. It notices that smarter systems generally can make more paper-clips, so making itself smarter will likely increase the number of paper-clips that will eventually be made. It does so. It considers how it can make paper-clips using the 3D printer, estimating the number of possible paper-clips. It notes that if it could get more raw materials it could make more paper-clips. It hence figures out a plan to manufacture devices that will make it much smarter, prevent interference with its plan, and will turn all of Earth (and later the universe) into paper-clips. It does so.

After giving it some thought I came to the conclusion that is indeed very tricky to give a true strong AI a safe reward function. Therefore I consider strong AI dangerous and since its dangers far exceed the potential benefits any research in this field should be stopped.

Kattentherapie for everyone

March 31st, 2010 Posted in Uncategorized | Comments Off

Nienke has started being a cat behavior therapist. We have made a nice site where you can find how a cat therapist can help you. You can find it at http:://www.kattentherapie.nl.

Single cell sentience

March 14th, 2010 Posted in Science | Comments Off

There is a debate whether thoughts or memories are physically located somewhere in a brain. It is still unclear whether there exist such a thing as a “grandmother braincell” that activates whenever you think of your grandmother. Or more clearly: the activation of that cell really represents thinking about your grandmother, and all relations and associations with that thought are the neural connections of that particular brain cell. I once went to a lecture of physicist De Gennes, who is now doing work on the nature of memory, who talked about the grandmother braincell. Recently there was also news on the discovery of the “Jennifer Aniston brain cell“, a cell which represents thinking about the actress from the TV sitcom Friends.

nervecellOkay, now here’s a thought: suppose it would be true, that each concept can be related to a particular brain cell. Then I can think of a particular concept, namely a brain cell that is related to the concept of itself. Can you imagine? The concept of a brain cell that thinks about itself is a thought, and therefore it must be represented by a brain cell in your brain. Then, indeed, that cell is thinking about itself when it activates. It reminds me a bit of Brouwer’s fixed point theorem.

The difference between time and space

January 4th, 2009 Posted in Science | No Comments »

In this post I would like to explain the way nature seems to choose the most simple and elegant way, but it is often thought to be complex due to the limited minds of us humans. For this I will talk about the relation between space and time, why they are similiar, and why they are different. I hope you will find it interesting.

Any event can be labeled by simple coordinates: there are three space dimensions, so we need three coordinates x, y, and z to denote where it happened, and there is one time dimension so we need one coordinate t to denote when it happened. This makes it tempting to just see events as points in a 4-dimensional space (which is generally called Minkowski space).

minkowski_a1.pngSo here it is: our event a has got a place in the four-dimensional time-space. We can denote it’s position with a vector 4vec.png. In the illustration I did not draw the third space dimension z, because I cannot draw a 4-D cube, but that is not important for my purpose.  What is important is that in this way, I can represent time and space as absolutely equivalent members of the space-time, and there is also no prefered direction like up or down, of backward and forward in time. But we feel that time and space are not alike in our experience, so something is missing. We should investigate what is the difference, and what is the most elegant way to include it.

In physics, we have learned that it is often instructive to look at symmetries. This is because if the laws of physics are symmetric under a certain transormation, then there exists no measurement that can distinguish between the situation before, and the situation after the transformation, and thus they are equivalent. In this case there are symmetries in the space and time dimensions. If I translate the event a and other events that are caused by it to a different location in space and time, all laws of physics remain the same and the principles of cause and effect work in the same way. This is called translational invariance. Alternatively, we say that the universe is observed by an observer (a physicist, of course), who stands at the point t = x = y= z = 0. Now moving all events is equivalent to moving the observer.

There is another symmetry, which is rotational symmetry. It is equivalent to the observer tilting its head a little and seeing the universe a bit rotated. Let’s look in detail how this works for two events in a two-dimensional space of x and y.

rotate3.png

Here you can see we have rotated two events a and b around the origin, over a certain angle ?. The new coordinates are given by

rotatespaceform.png

Allright, this seems very well. Now lets try the same thing when we mix space and time dimensions. For simplicity, we’ll take a slice of space-time, and rotate in the x and t coordinates.

rotatespacetime.png

I have used an arrow to denote that event a may be the cause of event b, because in the left figure event a happens before event b.  An object may be travelling from a to b. But you see in the right figure, after we have done something to the observer and rotated the universe around him, b surprisingly happens before a! An object travelling from a to b is now going back through time!

Now I don’t completely understand why natures refuses this type of symmetry. but it is certainly strange that one observer finds a to be the cause of b, and another may find the opposite, even though they are watching the same events. So indeed, time and space are different after all.

Now it would make sense to dismiss the idea that space and time are related after all. It would seem simple to just say that time and space do not rotate into each other, and we need other ways of describing a transformation of space-time that affects both space and time. This would be the Galilean transformation. But would this really be the simplest and most elegant way of describing space-time?

In fact, nature was so wise to make rules that are so elegant that humans did not discover them until the 19th century. It works like this: whenever we want to rotate through space and time, we should not rotate over an angle ?, but over an angle i?. And we must replace all time coordinates t with the time coordinate ict:

difference.png

The i is the well known imaginary unit, defined by i2.png The c is the speed of light, which serves as a conversion factor between time and space. That’s it! Now we can write down the rotation equations that we already found in the new system:

rotspacetime.png

And indeed this is a new symmetry of our universe. It corresponds to putting the observer on a moving train and let him observe from there. According to this observer, the time and space coordinates of events are remixed in such a way, that all objects that were previously standing still are now changing their position with more or less the speed of a train. And these two equations describe exactly that. I know for a non-physicist it seems a bit strange with all the i’s in there, but in the end these equations can be used quite easily.

You may object that this solution seems different from the normal idea that when you get on a train, other objects will change their velocities with exactly the speed of your train, and that’s that (Gallilean transformation). And you’re right: the solution is different. And now comes the beautiful conclusion of this story: the above solution is more elegant, it is different, and, if you do it very precisely, you can measure it is also the only correct solution. It is the basis of the theory of general relativity. I conclude that physics is the search for the simplest solution that fits a problem, but sometimes it is very difficult to be smart enough to understand the simple solution. I hope you now have an idea about how this search works, though I admit this story is much too short for a proper explanation. If you like it, please leave a comment!

Animating the living cell

October 5th, 2008 Posted in Science | No Comments »

ATP synthaseI find biochemistry intriguing. The most complicated reactions take place in this messy fluid inside living cells, carefully balancing the laws of entropy and energy to maintain order. Now I came across the channel ndsuvirtualcell in YouTube. This is some kind of educational project, that has produced really nice animations of the most important biochemical processes: photosynthesis, (with a detailed look at Photosystem II and synthesis of ATP), DNA transcriptionmRNA translation and a few others. I don’t really know how these things are discovered, but it looks cool. Do not be deceived by the clean look of these animations though, we can simulate the way some of the smaller proteins actually behave, and it is a whole lot more messy.

Schizophrenia

September 3rd, 2008 Posted in Science | No Comments »

Schizophrenia is a very complex disease or diseases. People who have schizophrenia have recurring episodes, psychoses, which are characterized by delusions, often paranoid or megalomanic, and hallucinations. Thoughts can be disorganized and a sufferer from psychosis has difficulty determining what is real and what is not. The disorder usually appears in the late teens, early twenties. People with schizophrenia suffer from symptoms such as hearing internal voices not heard by others, or believing that other people are reading their minds, controlling their thoughts, or plotting to harm them. These symptoms are often terrifying for them.

Prevalence of schizophreniaThere are things I’ve learned about schizophrenia I find surprising and interesting. Schizophrenia is not at all a rare thing: it is estimated that 0.5-1% of all people suffer from it. In the figure on the right you see that the prevalence of schizophrenia is actually quite high compared to other well-known diseases such as alzheimer and multiple sclerosis. Yet many people don’t know what it is, or have misconceptions about it. This is surprising, and I wonder why that is. In my opinion, the reason may be that after two centuries of study, we still can’t give a satisfactory answer to the questions: What is schizophrenia? What causes it?

There is no clinical test that can prove somebody has schizophrenia, the diagnosis is based just on symptoms. Therefore, it is not even known if we are talking about one disease, or multiple diseases with overlapping symptoms. Some have suggested that people with schizophrenia just have a peculiar personality and different ideas about the world and reality, and some theorize that the diagnosis only shows a broken relation between the individual and society, not necessarily an inherently sick person. Though I think this cannot give a complete picture, we must keep an open mind about these things, which are part of so called anti-psychiatry. The situation is made more complex by the fact that delusional thoughts are by definition impossible to classify as such by the person who has them. There are many people who will disagree with their diagnosis, even outside of a psychotic crisis.

Also, after all so much time and research, we still have no full understanding of the causes. Some risk factors have been identified. The strongest is a genetic one, roughly 25% of cases have a direct relative with the same disease, or a similar one such as bipolar disorder. Problems in the development of a child may be influential, especially around birth. Later, stress and social exclusion are related to the development of schizophrenia, and a stressful event often precedes onset of the disease. Here, distinguishing between cause and effect becomes confusing, and it has proven difficult to answer some important questions:  Must a person have a physical vulnerability that can be triggered to develop the disease? Is it pre-destined, and will a stress event just advance the onset? Similary, there are very little reliable predictors for predicting the prognosis of a first-episode patient. A person may improve, but often the disease becomes chronic.

The fact that the nature of schizophrenia is difficult to comprehend, that a patient seems to be very “different” in many ways, and that it can cause such great distress, may make people prefer to classify it as a “remote” problem. This increases stigma, already a serious issue.

ChlorpromazineThe psychological effects of psychosis are devastating. Fortunately there are strong drugs, called antipsychotic drugs, that have turned out to be effective for many. The first such drug was chlorpromazine, which was accidently discovered in the 1950’s and has since been described as the greatest breakthrough in psychiatric medicine. This and other drugs have very serious side-effects which makes it very difficult to stay on treatment for years. Newer drugs have been developed, in the mid 90’s we saw the arrival of the so called atypical antipsychotics. These were hailed as a terrific breakthrough by the big pharmaceutical companies, but now it turns out that they’re not a fantastic improvement upon the old ones. The market for these drugs is enormeous, but we don’t seem to be making much progress.

This is mostly due to the fact that we are still not sure how these drugs work. They all block dopamine-receptors in the brain, which accounts for their effectiveness, but there are confusing issues. The chemical effects of the drug takes place within minutes, but the strongest treatment effects are visible only after two weeks. Maybe they work by breaking some self-enforcing loop, slowly diminishing the psychotic effects. Maybe you can compare this to treating somebody who is angry by a sedative: this will in no way cure the patient, but diminish his symptoms. Indeed, indifference is a side-effect of many antipsychotics. Part of the problem of developing new drugs is that since we have no clinical test for schizophrenia, we don’t know exactly what we are trying to cure. Animal testing is an important part of any drug development, but in this case it is not even clear how to give animals schizophrenia, and how to measure whether they have been cured. Nevertheless there is an impressive amount of research going on. Some research points out that our current drugs do not improve the long-term prognosis. Prognosis in developing countries where these drugs are generally not available may in fact be better than in developed countries, though it is very difficult to draw conclusions from this. Psychic therapy is also an important part of an effective treatment.

The latest science comes from genetic research. Many methods have become available to study which genes are contributing to a vulnerability for schizophrenia, and what is their function. The newest results add to the already complex picture we have: there are many small defects in many places that all contribute a little bit. Though this may explain why schizophrenia has never been elimated by evolution, it means that finding a cure will still be a very challenging task.

Why we need gasoline tax, and how to lower fuel consumption

June 15th, 2008 Posted in Science, politics | 1 Comment »

Prices are high

So the truckers are protesting against high gas-prices in several countries. The truckers are unhappy and demand compensation from their governments, and to increase the pressure, their actions are directed at the general public.

This is all wrong. Everybody always wants compensation from the government, be it truckers, farmers, fishermen or busdrivers. This makes no sense, as the government is not the cause of their trouble. The cause is the price of crude oil, which has gone up from 40 dollar to an incredible 135 dollar per barrel in four years. Ironically, as the fuel tax is a fixed price per liter, this means the relative taxation has gone down from 72.8% in 2004 to 57.6% at this moment.

So before we try to find a solution, we must ask: why has the price of oil increased so much? Quite simple: the production of oil is stalling while the demand is increasing. It has been theorised by Mr Hubbert in 1956 that the pattern of oil production follows a certain curve: at first production rises due to the discovery of new oil-fields, than it stalls as the discovery rate equals the rate at which the oil-fields run out, finally the production decreases as the discovery of new oil-fields cannot keep up. The US oil production followed this curve nicely with a peak in 1970, as predicted. The moment that world production will peak is called peak oil. When this will happen or wether it has already happened is of course a matter of debate, but it is likely to be around 2005-2010, which means about now.

I hear crazy things about who is to blame for high gasoline prices: taxes, speculation, accidents, or an OPEC strategy. But the fact of the matter is that is simply about supply and demand. The elasticity of oil is very low, meaning we have very few alternatives at the moment, and this accounts for the soaring prices. This also means that high gasoline prices are here to stay and will in fact keep on rising no matter what.

The only solution is to reduce our consumption. This is important in several ways. First, it is better for our environment. Second, it stop the constant flow of oil money to abroad, money which we could better spend here to improve our own economies. Third, the money now gives support to undemocratic and isolationist powers and we must stop that. Governments in countries such as Saudi Arabia, Russia, Iran and Venezuela are basically legitimized by a large budget surplus, not by their actions. Their citizens will not complain as long as there is money to spend or it is too late.

And the best way to reduce consumptions is by having a fuel tax. Thus I am in favor of increasing the fuel tax (and perhaps reduce some other tax in return). The best idea I’ve heard so far is to set up some sort of buyer’s cartel of countries that raise the fuel tax, thereby forcing the oil-producing countries to lower their prices and keeping the money mostly in our own pockets.

Lower your own fuel consumption

I have been driving a car to work in the past 2 months and have picked up some strategies for lowering my fuel consumption. They are not rigorously tested, but I did go down from 16 km per liter to 19 km per liter, which is quite good for a volkswagen Fox.

I have tried to put these in order of importance:

  1. Avoid driving. Obviously, but matters a lot.
  2. Buy a car with a small and efficient engine. Why do you need a car that goes 200 km/h if you are only allowed to go 120 km/h? If you need a big engine for quick manoevering, you just are not very good at driving.
  3. Buy a car with low air-friction.
  4. Buy a light-weight car. Matters especially for city-driving. This and the last two points are indicated by the car’s official fuel consumption.
  5. Drive in the highest possible gear. As long as your rev doesn’t go below 1500 there is never a problem. You can go to fourth gear when you go 50 km/h, to fifth when you go 70 km/h. When you accelerate from a green light go to second gear as soon as possible. Don’t accelerate in a too-high gear though, you’ll increase consumption of you press the gas-pedal too far. Accelarating quickly to go to a higher gear is good, surprisingly.
  6. Drive slowly. Do not go over the speed limit. Drive 100 km/h where you can go 120 km/h. Calculate for yourself what the time difference is. This point combined with the previous shows that the best way to drive a car is to go 70 km/h (or less) in fifth gear.
  7. Anticipate situations or other traffic. Can be very important. It’s a bit of a sport: try to never break in a traffic jam or elsewhere. Think ahead. Do not accelerate if you have to decelerate later on.
  8. Slow down without breaking. There is some confusion on whether to drive in neutral (clutch pressed) or to use the engine for breaking. The engine in modern cars completely shuts off fuel flow if you use your engine for breaking, if you drive in neutral the engine needs a little bit of gas to keep running. But: breaking with your engine gives large friction losses inside the engine. The rule of thumb is: slowing down in neutral is better, unless this means you will have to apply the breaks later on. In that case, using the engine for breaking is preferable. It may even be theoretically be possible to lower your fuel consumption if you keep accelating and slowing down while driving, though this is better suited for hybrid cars. Do not turn of the engine in your car as you will lose pressure on the breaking fluid the second time you use the breaks, as I can testify.
  9. Avoid air-conditioning. Also avoid other appliances if you don’t need them: rear window heating, lights etc.
  10. Keep your tires inflated to the right pressure.
  11. Keep your windows closed. Use the ventilation system instead.
  12. Lower the weight of your car. Do not bring unnecessary stuff.

The worms are coming

June 4th, 2008 Posted in Fun | 1 Comment »

Found a wormToday, I cleaned the worm-bin. What? Yes, I have worms. Keeping worms is not a very common hobby, but it is actually quite interesting. To begin with, I have thousands of pets, which is unique. Also, I recycle my garbage and use the vermicompost to fertilize the plants on our balcony. It forces me to think about garbage as valuable stuff - I now feel uneasy throwing away those nice cauliflower leaves in a normal bin. But most importantly, vermicomposting takes you to the level of very small animals, a category completely ignored by most people.

If you talk about “protecting nature”, most people think about protecting wolfs or panda bears or giant whales or perhaps some rare little plant that has very peculiar needs. Great amounts of effort are put into the protection of certain types of birds, and farmers are supposed to go around bird nests when they mow their fields. But what is nature? It is about all levels of life, plants, animals, funghi and bacteria, eating each other or avoiding to get eaten, and using all these special strategies for that, so we can make nice documentaries about them. If you think about it, most of the life actually happens at the levels of smaller organisms. Unfortunately, insects and worms have an extremely low “aaibaarheidsfactor” (Dutch word invented by Rudy Kousbroek, qualification of how much people like a kind of animal).

The worm-bin can give you nature on your own balcony. There is a whole zoo of animals in there: woodlice, mites, centipedes, snails, spiders and many things that I can’t even see. Keeping worms lets me think at the level of the worm: potato peel, mmm! Banana peel, not so good because of the pesticides. Hey, there is a staple on this teabag, you can’t eat that!

Worms going into their new fresh garbage.Keeping worms is quite easy. I read a book called “Worms eat my garbage” by Mary Appelhof which explains it all. I have about 0.5 kg worms. A worm-bin does not smell if properly set-up, because there is no anaerobic decay. It does not take much time, except once every few months when the worms need to be separated from the ready compost. There are perhaps easier methods than hand-sorting, as I did, but I am still only learning…

Nanda and Milan on their way to India!

May 15th, 2008 Posted in Fun | 1 Comment »

Nanda en Milan vertrekken!Incredibly, Nanda and Milan have just started pedaling for six months. Their bikes will take them through Germany, Switzerland, Italy, Turkey and India (with a little help from boats and planes). They are thoroughly prepared and have done “bicycle holidays” before, and I think they will have a great time. The most fascinating thing about this kind or thing is the fact that everything you need is in your luggage, and your transportation is your bike. Everything two people need for six months is in these pictures.

Nanda en Milan vertrekken - deel 2!

When asked for their feelings, they replied “excited” and “tense”. My own feeling? “slightly jealous”. You can read about their adventures on their website.

Baby computer

April 28th, 2008 Posted in Fun | 4 Comments »

The Asus EEE pcHave you ever wondered why it is everybody is Baby computeralways trying to improve things, but the actual progress made is not so clear, really? Take a computer for example: they get better every year, cheaper, smaller etc. but what have we really accomplished in the past 5 years? Not so much, perhaps.

Three Asus EEE pc’s

So I bought the Asus EEE pc. It cost me 329 euro, it runs on a neat 0.6 GHz processor, has 512 MB memory and weighs less than a kilogram. And it’s quick! It boots in about 25 seconds and does everything I want (i.e. it runs linux). Now that’s progress!