Saturday, May 28, 2016

How to Estimate Anything From Nothing...

or How to be a Good System Engineer

This will be a discussion of how to estimate something from little or almost no information.
We call this making a Scientific Wild-Ass Guess or a SWAG.

I've been pretending to be a system engineer for a long time.  And I have to admit that I still have to go look up the average wait time in a queue or the Poisson distribution. Typically estimates are extensions of known data, the more data the better.  But what happens when you know hardly anything about the data that answers the question, because it doesn't exist?

Something from Nothing...

The modern method of how to estimate something from no information at all is the classic paper: "Implications of the Copernican principle for our future prospects" by J. Richard Gott III [1]. Sounds innocuous, but is probably one of the best tools ever invented for doing system analysis when information is lacking. It's almost as important as Baye's theorem that allows one to adjust their estimates of a probability given new information.  With these two tricks you can make an estimation of almost anything, and correct it as you learn new information. Some might call that science, but it's more like 'practical science' or what most people call engineering.

How long will the Berlin Wall stand?

If you were standing beside the Berlin Wall when vacationing in Berlin, you might ask yourself that question.  But how could you predict something like that?  Go collect data about all the walls in all the cities in the world by type, size, lifetime and then use that data to predict how long the Berlin Wall will live by finding all the walls that are similar to it? That's a lot of work. But it will almost certainly give you a fairly accurate lifetime prediction for the Berlin Wall. We can make a SWAG without all that additional data, it will be more inaccurate, but maybe it's accurate enough for whatever other calculation you are doing.

Gott's Copernican theory goes like this: If I am at a random place and time and I observe something that is N years old, with a 95% confidence level I can say that it's extended lifetime will be between N/39 and 39N.

Or a 2/3rds chance of having a 1/6 to 6 times longer lifetime.  Why?

This is because if you are observing the lifetime of an object at random, you can be there at any time during its lifetime. This means that if you assume that lifetimes are randomly distributed and your appearance is random, half the time you will be in the last half of the object's lifetime and half the time you'll be in the object's first half of its lifetime. Continuing on, there's a 2.5% chance you'll be in the object's first 2.5% of its lifetime or 1/40th of its life. Double that for a 95% confidentiality limit, so 1/39th and then 39 times for the odds in the other direction.  For instance, Gott was a tourist in Germany in 1966.  He saw the Berlin wall.  He read a plaque that told him the wall was built in 1957. He then concluded that he was seeing the wall at a random time during its lifetime.  He wasn't there to see the wall, he just happened upon it. Since at that time the wall was 9 years old, with 95% confidence, he expect it to last another 9/39 years (12 weeks) to 39*9 years (351 years). It came down in 1984 after 27 years. He expected it to be there 1.5 years to 54 years 2/3rds of the time. Whats the most likely time to live? 50 % of the time it is between 3 and 18 years. The most likely time is 9 more years.[2]

How Long will North Korea's communist state survive?

They were founded in 1952.  That was 64 years ago.  With a 95% confidence level the state of North Korea will be around for 64/39 years (1 year, 7 months) to 64*39 years (2500) years. Wow, that's scary. Maybe we should figure out some way to 'nudge' that around.

See how easy it is! There's been many arguments about why exactly this formula actually works.  Its been tried on Broadway show run lengths and found to be accurate and I've read about it on the Internets, so it must be true. Let's use it to calculate something interesting: How long will the human race survive? (From [1]) The human race is a species. It's been around for about 200,000 years. You are alive today at some random time.  The human species, with a 95% probability will be around for another 200,000/39 years (6000 years) or 200,000*39 years (9,321,000 years.)  Hmm. Looks like not enough time to settle the galaxy, but it depends upon how fast you can travel. It doesn't say that we might not create some other species to settle the galaxy... but it seem unlikely that we could by ourselves. Seems like it's going to take killer robots to settle the galaxy or nothing.

Thanks for reading,
 -Dr. Mike

[1] Nature, Vol. 363, 27 May 1993
[2] reference

Sunday, May 1, 2016

The Case for Intelligent Design.

What is the Case for Intelligent Design? 

First we have to define Intelligent Design.  In the common sense of a replacement for "creationism" it is obvious that there is no case for it.  It's actually a negative argument: since things are so complicated they must be made and could not have happened by accident. That makes some sense: particularly complicated things are unlikely to be made by accident, so things that look like they are designed probably are. This doesn't apply to humans, though; we haven't evolved by accident: we evolve by the random mutations but they are passed along by survival of the fittest.  If the mutation has a small chance of making the entity survive long enough to have a better chance to reproduce, then this mutation will eventually be passed onto more of the members of the species as time goes on.

This was the insight that Darwin (and many others) had 150 years ago.  Darwin was the first to understand that the assumption explained many, many things and he struggled to make the argument irrefutable.  And he did a great job of it, and others have admirably extended his argument. This theory actually begin a science of biology that has driven biological research ever since.  It has driven us to understand genes, DNA, and the history of life itself. The theory has been greatly expanded and the idea that anything in biology has been "intelligently designed" is not believable, there are just too many counter-examples of bad design in evolution to even contemplate this.

Except for Craig Ventnor's latest work. Craig has actually "designed" a live cell by picking fragments of DNA from multiple bacteria.  Our theory of evolution (mutations, inheritance and survival of the fittest) is a statement about the past; in the past evolution occurred without intelligent design, going forward this in no longer true. And of course, the intelligence designing creation in the future is human intelligence, not some supernatural intelligence. In fact, the new capability given by CRISPR to edit individual pieces of DNA in living beings shows us that intelligent design can happen. So I would have to say that Intelligent Design is now a fact, not in a pseudo-scientific version of creationism, but in reality.

What should be Intelligently Designed?

This leaves us with two remaining interesting questions. The first: Should people be intelligently designed? In my opinion the obvious answer is yes.  But there must be limits on the usage of this new technology. Just like there are protocols on how to handle and study dangerous infectious bacteria, there have to be limits on the study and use of gene altering techniques.  I won't go into where the limits should be, but there should definitely be protocols to contain the damage that this research could do. What other limits should there be?  Elon Musk and Sam Altman have brought up the idea that Artificial Intelligence research should have protocols and limits.  I think they are partially right, protocols need to be set up to make the research safe, not limits.

Where else should we limit Intelligent Design? The world has agreed that biological weapons and other weapons of mass destruction should only be researched under certain protocols and the actual moral high ground is to not research them at all. You want to make sure that your research won't cause the human race to earn the last Darwin Award.

And I would claim that you don't want anything to be designed by a 'supernatural' intelligence. Who would trust anything outside of your control or influence? I put supernatural in quotes because I think it is a word without a meaning in reality.  It certainly affects reality by influencing the way people think.  But the very idea of supernatural is the ultimate oxymoron.  What does supernatural mean? It means 'attributed to some force beyond scientific understanding or the laws of nature.' But what does nature mean? Nature is everything.  But maybe not?

Maybe there's an unseen space beyond what we can normally sense.  That would sort of be in a 'different' nature as it's hard to detect. And maybe beings in that space can affect things in nature.  Maybe. Has this ever been measured in a reproducible way?  Not really.  It could happen.  It's not likely.  It could explain souls.  The soul could exist in this other realm and be attached to your brain and control it. It could happen.  And if it does happen, we'll know pretty soon, so it won't go undetected.  It won't be outside of nature, it will be measurable and recognizable.  As I said, being beyond nature is an oxymoron.

What else should be intelligently designed? Other than everything.  I would argue that the most important thing to be intelligently designed are morals.  So question two: Is it moral to intelligently design morals?

Intelligently Designed Morals

Why do morals need to be intelligently designed? Hasn't this already happened? I would argue no. What's happened is that evolution has design a 'moral code' into humans, the problem is that evolution is random, driven by individual survival.  It's not a moral code but a survival code.  And a survival code is not moral, it's selfish, it's brutal, it's just not right.  This tells me that we can't depend upon our inner moral sense to design a moral code.  We have to start from first principles.

The first requirement of a moral code is that it must be designed from overriding principles.

People have attempted to design morals many times; however, I think we need to add a framework for the intelligent design of morals.  We need to set some goals. What kind of goals should a moral code have?  Why do we need a moral code?  We need a moral code to decide on how to be good vs. evil, or do we? Our internal compass is already a good guide.  We need a moral code to make sure our inner guide is correct.  There are many logical traps that we can fall into just listening to our moral intuition.  Why this is so has been studied in economics (with many Nobel prizes awarded); a great example is: "Thinking Fast and Slow" by Daniel Kahneman.  The brain has many biases and only clear thinking and real logic can prevent us from making mistakes.

It's obvious to me that all religions have been designed by men.  This doesn't mean that they can't have reasonable moral codes. Many religions claim to have been inspired by or dictated by God or gods, but they all seem basically unchanged if the gods don't exist.  Which is a good thing as there has yet to be a god that actually does anything real or useful as far as I can tell from the historical record. The idea of a god is useful, the god itself, not so much.  In my opinion religions and morality should just stay separated as they've done such a poor job of them in the past. (Slavery, equality under the eyes of the law, torture, discrimination, child abuse, wars, the list goes on...)

The second principle of moral code design is that it can't depend upon reputation for correctness.

Morals are too important to not be examined, debated and understood. Depending on God or tradition will lead to serious problems with moral codes, it can be a good start, but if the proposed morals aren't carefully vetted, you can run into big problems.  If you don't have an overriding principle, you can't verify a moral code, so principles one and two work together. 

I would claim that we want to design morals for conscious beings.  If a thing isn't conscious, morals don't apply to it. Why is that?  Only conscious beings have experiences, without experience there is no morality. So that will be our third principle: Morale codes are for conscious beings.

Moral codes set the standards for interactions of conscious beings.

If you want to design a morality for all conscious beings it leads you to start with certain postulates. You can design a moral code that is not fair for all conscious beings. Almost all religious moral codes have been designed this way.  In Christianity the moral code is for all humans to be treated equally (animals and lesser beings need not apply.) Jewish morals actually care somewhat about some animals. Islam is very clear that believers and non-believers are treated differently. I'm not an expert in religions, but this is what I see around me. I'm happy to be corrected in any assumption or manner.

It appears that morals designed by religions are sorely lacking in morality. Not to be so unexpected as they were designed thousands of years ago typically, when we were somewhat more ignorant than we are today.  Even the religions invented recently have dropped the ball.  Think of Scientology and Mormonism; not much break through in morality from either of those religions.

I'm gong to discuss moral codes that treat all conscious beings as being worthy of respect. I don't necessarily like the consequences of the moral codes they create, but that's besides the fact. How do you rank moral codes?  If we can use reputation, what can we use?  You could use the criteria of which ever moral code makes you happier or increases your well-being. Makes sense.  But I would argue that it is incomplete and leads to unfair and immoral moral codes.  You need to judge a moral code more broadly, it's not just you that needs to agree it's best, it's everyone.  This puts severe restrictions on how to judge a moral code.

A moral code needs to be evaluated in how well it treats any conscious being in any situation.

Wait, you're saying, any conscious being?  An insect? A mouse? A slime mold? Yes, any conscious being that has experiences must be taken into account by a moral code. Different levels of experience could be treated differently or expected to act differently and be treated differently (can a mouse even understand a moral code?) I think this requirement is paramount in building a real moral code. It's not just about human beings it's about conscious beings. Different beings have different levels of consciousness and different levels of intelligence that allow them to foresee the consequences of their actions, or not, so should be treated differently by moral codes.

Finally a moral code should maximize the amount of well-being or happiness (in the philosophical sense of flourishing) of conscious beings. Just like medicine works to increase your good health.  We don't know what perfect health is, but we can usually tell in what direction good health is: less pain, longer lives, more opportunities, etc.  Same for well-being.  While we don't know what the perfect well-being is, we can usually figure out what direction it is in: better health, more flourishing, less pain, more opportunities, more consciousness, etc.

A well designed moral code must maximize the well-being of conscious entities.

Let's recap. Moral codes tell us how conscious beings should interact, they must be designed from first principles and evaluated on how well they treat all individual conscious beings and how well they maximize well-being for all conscious beings. That's a tall order. All the moral codes I've seen are seriously lacking in many of these attributes, so I would claim in great need of fixing.  I'm pretty sure that the ultimate moral code has not yet been designed. However, viewing a moral code in this light gives us hope that we may be able to propose moral codes and tune them for best effects.

The world is overdue for good science and good engineering to be done on moral codes. There are many falsifiable hypothesis that can be made on moral codes given this framework (science) and many ways to run experiments to decide if one part of a code is better than another (engineering.) It's about time to stop reasoning about morals using philosophy and religion and to start reasoning about morals using the modern practices of science and engineering.

It's time for morals to be intelligently designed.

We've come a long way since we learned to speak and write. No reason to stop progressing now because someone told you they already have the answer. Question it. Evaluate it. Review it. Find the flaws. Improve it. Never forget. My inspirations:


Thanks for reading.
 -Dr. Mike










The Buddha’s not there.The Illusion of Truth

​ An Ode to Existence in the Prophetic Style. [The Multitheist’s Lament] Buddha looks inside himself and sees the void, the void is nothing....