Google

Blog Archive

Recent Comments

2/28/08

Evolution is not slow

I came across a very illuminating and exciting Edge article on the conversation of Richard Dawkins and Craig Venter yesterday.

Richard Dawkins is famous for his gene's eye view of evolution, author of the many books including The Selfish Gene, The God Delusion, and The Extended Phenotype.
Craig Venter decoded the human genome, and more recently created synthetic chromosome from laboratory chemicals, a big step to creating "the first artificial life on earth."

Venter got publicity when he published his complete sequenced genome on the internet. It is conceivable that he might one day send his genome data via email and have the recipient recreate a Craig Venter clone. Very similar to how computer viruses spread through the internet.

The conversation between Dawkins and Venter, moderated by John Brockman and chaired by Hubert Burda, took place in Munich on January 22nd.
A transcript and a complete 1-hour video of the conversation are available.

The event was also widely reported by the German media, see Stern: Dr. Burdas digitales Gipfeltreffen, Spiegel: Craig Venter will Lebewesen e-mailen, and SÜDDEUTSCHE ZEITUNG: Die Zukunft der Selektion

Evolution is not slow.
I should have read the article before my post Why Evolution is Hard to Understand, because it could have cleared up some of my own misunderstanding of evolution. I have underestimated the power of biological evolution, when I said that it is a very slow process, which is true only if we look at the evolution of new species. But evolution also occur within species (so-called micro evolution).

We also tend to consider mostly the visible world, as Venter explains:
"In fact, it was one of the biggest surprises for the scientific community. What we found in the environment. Most people expected just one dominant species. What we found were thousands, tens of thousands, of very closely related organisms, all basically the same linear set of genes, tremendous variation in those genes. But there was not one dominant one. There was this community of related organisms where perhaps none of them had gone extinct, or, if they had, there were literally thousands of ones to replace them. The problem we've had, I think, with looking at evolution, I think it's been overly simplified because we've always been looking at the visible world, not the absolute majority of life on this planet, which is the invisible world."

Evolution in viruses happen very fast. "This is the world of biology that we live in, that we don't see, where evolution takes place on a minute-to-minute basis, not on the speciations of giraffes versus elephants versus kangaroos but the tens of millions of species that constantly are affecting the metabolism of our planet." (Dawkins)

Important role of viruses.
We usually think of viruses as something negative, like computer viruses which delete our files and spy on us, or worse, viruses which bring disease.
But virus is crucial in biology and cultural transfer.
The role of viruses in lateral transfer has now been better understood, which is of utmost importance as a third of our genome is virus, according to Venter.

Dawkins went further by stating that we are a huge society of viruses: "I would like to regard the genomes of the giraffes and kangaroos and humans that you refer to as just another set of viruses in close-knit societies. So the gene pool, I should say, of giraffes, or the gene pool of humans, or the gene pool of kangaroos is a huge society of viruses."

In this connection, Freeman Dyson thinks that evolution is now man-made and open source, referring to our role in changing the environment and the open source nature of material in viruses.

Environmental change.
The conversation confirms that what I called the second (change in environment) and third (change by genetic engineering, cloning, etc) strange loops of reflection are now in full force.
This changes can be good or bad or even destroy all of us.

Both Dawkins and Venter regarded climate change as a big threat. Technology must now find solutions to the effects generated by past technologies which lead to pollution, waste and climate change.
"It's a real-life danger that we're facing now. I've argued that we are 100 percent now dependent on science for survival of our species. In part, science of today has to overcome the scientific breakthroughs of previous years because we've advanced internal combustion engines, because we're so good at burning carbon that we take out of the ground, we did it blindly without any consequences of, that it might totally affect the future of the planet."

" What we're doing with burning oil and coal is we're taking millions of years of compressed biology, we're burning that over the course of years and putting it in the atmosphere."(Venter)

Nobody likes taxes, but carbon taxes seem necessary to deter people from destroying "millions of years of compressed biology". Could it be that high oil prices is a blessing in disguise?

Here Venter offers hope that biotechnology, which he calls "the only nanotechnology that works", could reverse produce fossil fuels from carbon dioxide. Power plants and cement factories could be used as abundant sources of carbon dioxide for this process. The product could then be consumed or pumped into oil wells and coal beds.
Is this too futuristic? Venter does not think so.

2/18/08

Genetic Algorithm for 3D Traveling Salesman

My last post might give the idea that I am against genetic algorithm (and the related genetic programming, memetic algorithm, evolutionary computation), which is far from true, I do like the genetic family of algorithms. I was only saying that they are not suitable for representing biological evolution.

Here is a recent application of genetic algorithm to solved a 3D traveling salesman problem. You can watch a movie of the algorithm as it is run.

More about the subject, Visualisation of Genetic Algorithms for the Traveling Salesman Problem by Johannes Sarg in Java and Traveling Salesman Problem
Using Genetic Algorithms
in Microsoft.Net are a sample of sites giving you source codes of the algorithm.
The above algorithms are quite straightforward and intended for introduction. More advanced algorithm are available in research reports.

2/15/08

Why Is Evolution Hard To Understand?

A friend of mine, pondering on the Darwin Day Celebration asks why evolution is so much misunderstood compared to, say Einstein's relativity theory, which is generally acknowledged to be hard to understand.

Jacques Monod once said (as quoted by Dawkins), "Anther curious aspect of the theory of evolution is that, everybody thinks he understands it". This is true, but many thinks evolution only in its much simplified form "Our great great grandparents were apes".
Compared this to the almost-zero platitude "Everything is relative".
Both sentences are intuitive, easy to understand and very misleading.

When it comes to the harder parts of Einstein's tensor manipulations and differential geometry, we rather leave them to experts to verify, and take them for granted.
It is different with evolution theory, for the reasons I am outlining below.

There are similarities between evolution theory and relativity theory, but they also have very significant differences.

Relativity is rather remote from us, we can visualize space travel like in the twins paradox, but only in our imagination. Evolution, on the other hand, concerns us very much, we are in the middle of it. Inevitably, our cultural baggage from our upbringing interferes. Some of us consider themselves created with a soul and a self, and find it difficult to accept evolution, not for lack of understanding, but because we don't want to understand.

Another difficulty is the time span involved in evolution. We can watch evolution in our life times in cases like the evolution of anti-biotic resistance, and some evolution in animals and plants, but the really big changes are beyond our life times.

Relativity is mostly described in the language of Mathematics. Evolution, as a biological process is much more complex than physical sciences.
Mathematics is also useful in biology, but there are many biological processes which cannot be described succinctly using Mathematics. Instead one uses algorithms to describe its processes. Unfortunately, the algorithms are also very complicated, much more so than computer programs, and some of so-called genetic algorithms are not faithful representations of biological evolution (more of this below), as they make too many simplifying assumptions.

Some people have difficulties understanding the difference between a random algorithm and a non-deterministic algorithm. Evolution is non-deterministic and it employs random variables as its components (e.g. mutations), but evolution is not a random process. Non-deterministic search algorithms (e.g. Tabu search, simulated annealing, ant colony algorithm) for example, employ random elements, but they can, under certain conditions, be guaranteed to find an optimal solution. I.e. a non-deterministic algorithm may have a deterministic outcome.

Evolution is technically defined as a change in the frequencies of the genes found in natural populations.
Evolution is not about the origin of life or the origin of the universe.

The characteristics of evolution is that it must have a copying operator (replication, reproduction, inheritance) which makes imperfect copies giving rise to variations, which together with other types of variations form a population from which selection is made by one method or the other.

Natural selection is the most important selection mechanism, but it is not the only one. Discoveries of other selection methods (including artificial selection) do not invalidate evolution. Natural selection exists side by side with other selection mechanisms.

Fitness is often used as a measure for selection. This is an unrealistic simplification, reinforced by the concepts of the genetic algorithm family. Genetic algorithm can be useful, for example in solving optimization problems, where there is a fitness function defined externally.
In biological evolution, it is more likely that there are multiple fitness functions, and the fitness measures do not have to be defined a priori.

I think this is a very crucial aspect of evolution: fitness is adaption to the environment. In the beginning the environment is considered as external and fixed or beyond our control, but as we evolve, we also change the environment, we change the definition of fitness. Evolution does not occur in isolation, as in idealistic genetic algorithms, the species co-evolved together and with their environment.
This is an example of a strange loop or what Barendregt called reflection. It is a very potent recursion in Logic, Mathematics, in Lambda Calculus, in mindfulness, in nature and in life.
This is the second strange loop in biological evolution. The first is the loop gene -> host ->gene by reproduction. The third strange loop is gene -> host -> gene by genetic engineering and cloning.

It should be noted that variation can arise from many sources: mutation, migration, genetic drift, recombination, etc. It was recently reported that viruses could form genetic material from which to build new genes.

Cloning and genetic engineering would bring new mechanisms for replication and variation, enriching and not invalidating the framework of the evolution.

It is true that the detail algorithm of biological is not complete, and it is not clear whether it ever will, but it does not make evolution theory less valid.
By analogy, Newton's gravitation theory remained valid, discoveries of electricity-magnetic forces and other forces do not change its validity. Only general relativity makes corrections to Newton's theory, in the sense of limiting its domain of validity.
Again, should we have complete theories of black holes and black matter, they would become refinements of our existing physical theories.
Advances in molecular biology (e.g. miRNA) will most likely refine or augment evolution theory.

I have alluded several times above the misconceptions arising by using the model of genetic algorithms. To be fair, they were inspired by, and not meant as models for biological evolution. They are useful in themselves. They can help, to some degree, in visualizing natural evolution, as in the interplay of mutation and crossover rates, how it sometimes lead to blind alleys, and how building blocks are formed.
As models of natural evolution, they are too crude, and misleading.

I think it would be very useful to have a new algorithmic model of evolution, which has a high degree of faithfulness to biological evolution. Such an algorithm could be used to demonstrate how complexity can emerged from simple things. This would save us from unproductive discussions about irreducible complexity put forward by creationists.

In summary, we still have a long way to go to make evolution understood by "the man on the street". There are many misunderstandings about evolution. Darwin Celebration Day should be celebrated by our efforts to communicate evolution. It is hoped that there will be less people who do not understand evolution.

“I think everyone is realizing that we need to be doing a great deal more. We just haven't made the effort to communicate evolution to people in terms they can understand. Evolution is exciting.”
Judy Diamond

2/12/08

Darwin Day Celebration

The Darwin Day Celebration is an international effort that is now an official program of the Institute for Humanist Studies. On or near Feb. 12, people around the world celebrate science and humanity on the anniversary of Charles Darwin's birth.

For Release

This Feb. 12 is the 199th anniversary of Charles Darwin's birth.

Hundreds of groups across the United States and the globe will celebrate the date as "Darwin Day" in honor of the discoveries and life of the man who famously described biological evolution via natural selection.

"Darwin Day promotes understanding of evolution and the scientific method," said Matt Cherry, executive director of the Institute for Humanist Studies which runs The Darwin Day Celebration. "This celebration expresses gratitude for the enormous benefit that scientific knowledge has contributed to the advancement of humanity."

Next year will mark both the 200th anniversary of Darwin's birth and the 150th anniversary of the 1859 publication of Darwin's "The Origin of Species", which presented the scientific theory that populations evolve over generations through natural selection.

The theory of evolution was controversial in Darwin's time and remains controversial in the United States today.

Recent Gallup polls show that 43 percent of Americans reject the theory of evolution and instead believe that "God created human beings pretty much in their present form at one time within the last 10,000 years or so." And at least four 2008 presidential candidates have said they do not believe the theory of evolution.

The Darwin Day Celebration started with one event in 1995. Last year there were more than 850 Darwin Day events world-wide. Darwin Day festivities can include debates, lectures, essay contests, film festivals, museum exhibits, art shows and even an "Evolution Banquet" with "Primordial Soup" followed by a "Darwin Fish Fry."

For information, visit: www.DarwinDay.org

Course available (free):
SCH100: Evolution, Creationism and the Nature of Science (Cornerstone) Massimo Pigliucci, Ph.D.: Evolution, Creationism and the Nature of Science will use the contemporary example of the evolution-creation controversy to provoke critical thinking about the nature and function of science as a method for understanding the world we all share. Intelligent Design and other forms of creationism will be explained and rebutted, highlighting the characteristics of empirical vs. anti-scientific and pseudoscientific thinking.

Darwin quotes:
-----------------------------------------------------------------------------------------------------------------------------------

I cannot persuade myself that a beneficent and omnipotent God would have designedly created parasitic wasps with the express intention of their feeding within the living bodies of Caterpillars—Charles Darwin

-----------------------------------------------------------------------------------------------------------------------------------

I have called this principle, by which each slight variation, if useful, is preserved, by the term of Natural Selection— Charles Darwin

-----------------------------------------------------------------------------------------------------------------------------------

Barendregt's Cover-Up Model of the Mind

Henk Barendregt (homepage ) is widely known for his work in Lambda Calculus and Type Theory, less known is his writings on Insight Meditation.

He used reflection without interference to observe his own mind, and the Cover-Up Model is the outcome of such experience. Reflection without interference is practiced in Insight or Vipassana Meditation.

He discovered three important characteristics of the Mind confirming what others have also experienced:

  1. it is constantly fluctuating
  2. it is unbearable
  3. it is not under our control

Buddhists recognize these as the three characteristics of all things. The first is impermanence (anicca). The second is dukkha (sometimes is translated as suffering), which Barendregt prefers to think as akin to the emptiness of extentialism or nausea. The third states that the mind is not under our control, in fact there is no central control as in the Cartesian doctrine. Since self is the illusion of such central control, the third characteristic is usually called anatta no-self.

The Cover-Up model says that we will always try to avoid the nausea by covering it up. Cover-up can take the form of feelings and thoughts: positive thinking, pleasure seeking, distractions (talking, watching TV, eating, etc), mysticism and many others. Meditation which only makes us relaxed or happy but does not lead insight is also a form of Cover-Up.

Cover-up does not handle the nausea directly, it just makes it less visible (for a while). The nausea appears hidden when we cover-up, we become ignorant of it.
However Cover-Up does not last forever, and we will have to constantly make ourselves busy to do it.
The analogue is when we sit for a long time, and feel uncomfortable, we change our posture, until we feel uncomfortable again.
In contrast to Cover-up, the real way to cope with nausea directly is the path of purification through mindfulness. There is a nice picture of this practice in a poster by Barendregt.

The explanation of why mindfulness works leads us to the Abhidhamma model
of the ancient Buddhist tradition (Tipitaka). According to the Abhidhamma, the stream of consciousness is discrete, basically serial but with parallel sub-branches.

Barendregt claims that the Cover-up model can be translated in terms of the Abhidhamma model.

Links:

2/5/08

Free Classic Computer Books from ACM

The Association of Computing Machinery (ACM) has made some very important classical computer science books available.
If you are not an ACM member, you can register for free.

The books include classical landmarks, some written by Turing award winners.

Books include EF Codd on Relational Database, Hoare & Jones Essays in Computer Science, Goldberg's SmallTalk, Kernighan & Plauger's Element of Programming Style, Dijkstra's Essays and Structured Programming, Brinch Hansen's Operating System and Concurrent Programming, Aho & Ullman Parsing, Minsky's Computation, Papert's Mindstrom, von Neumann's Computer and the Brain, Allen LISP Anatomy, Iverson's APL, Yourdon's Software Engineering and IBM on IBM System/360.

It is a real wealth of great work, and even now, more than 20 years later, they are still important to read.

Computer Science is a fast moving field, but there are basic ideas which do not change much with time. They are the invariants in the process of change.

Related: