speculations on extraterrestrial life...

There is a view that existence of life in other planets is almost a certainty. It is also very likely that two life forms in different planets end up communicating. Science fiction is full of stories where civilizations collaborate forming galactic federations. The characters in such novels end up being remarkable human like. I often wonder if it is possible for "human" like life showing up on other planets. It does seem rather reasonable to expect that "intelligent" species on other planets look dramatically different from us. Is it also reasonable to expect them to have eyes, ears, hands, legs, brains etc?

I'm of the view that it is indeed very reasonable to expect them to possess the faculties that humans and lot of other life forms on earth possess. My reasoning goes something like this. The big bang throws out a huge amount of energy. A fair part of it ended up being matter. These elementary particles ended up producing "elements". I think it is reasonable to expect other planets to have "matter" that is largely like the "matter" on earth. Solid, Liquid and Gaseous. The laws of Physics would ensure that Chemistry works much like it does on earth. An extension of Chemistry to Biology and origin of life on some planets is also inevitable and can be expected to occur very similar to how it did on earth. So, it is reasonable to expect that the laws of biology are also universal. Natural selection, I think, is a fundamental law of nature. I think it is very very reasonable to expect any intelligent life on other planets to be subject to the laws of natural selection. In short, I think the laws of physics, chemistry and biology ought to be universal.

Light exists everywhere. Life everywhere will evolve sight. Eyes are inevitable. Such a reasoning would almost extend to the sense faculties that we possess.

Locomotion would add survival value. Life everywhere would evolve legs. It seems reasonable to expect hands to follow. Bipedal life would evolve.

Life would evolve very much like it did on earth. Most intelligent life on earth does have hands, legs, brains etc. Is there any reason to expect things to be different on another planet? I think not.

My speculation is that if and when we find life elsewhere, it is going to be boringly similar to life on earth. Maybe pink plants and green men. That's about it.

BTW, blogging is apparently good for your mental health. Some more reason to keep typing.

Genetic algorithms in the Design of Comparators - Part 2

This blog follows up on my previous blog on the use of Genetic Algorithms in the Design of Comparators. The basic mutation methods used earlier used only single gene mutations. The next generation is just derived from one "parent" with some "genes" changed. It is very easy to modify the process such that the new gene for any iteration is derived by combining the changes that were made in the previous generation that results in the two best "children".

Let C = F (X1, X2...Xn) be the cost function of the "comparator" where X1, X2 ...Xn are the genes. If the changes in genes Xi and Xj result in the two best comparators in an iteration, the evaluated cost function will be of the form

Ci = F (X1, X2, ..., Xi, ...Xn)

Cj = F (X1, X2, ..., Xj, ...Xn)

In the previous method, I would have just used (X1, X2, ..., Xi, ...Xn) as the base gene for the next iteration. I could very easily change that to use to two best genes and use (X1, X2, ..., Xi, .., Xj, ...Xn). It also makes sense to evaluate the "base" in each iteration since it is new.

One can expect the effect of the two genes to "superpose" and result in a larger improvement than each individually. This need not happen in all iterations though. One might expect such a process to produce improvements in the design faster than the previous method. I don't think there is any reason to expect dramatically different "results" though.

Extensions that pick the top three best or more genes are also possible. I live on earth and not a para universe. Hence my preference for two.

A variant of the idea is to to pick up all "mutations" that result in an improvement from the "base" or "parent" gene in each iteration. Improvement is defined as the "cost function" evaluating to a value higher for the "child" than the "parent". Superposition principle would for the most part still apply and one can expect the next "gene" to be better than the previous one. I did notice that most of the "children" quickly end up being "not better" than the parent and very few changes actually improve the "cost function". I'm not sure if these methods are any better than the method that just selects the best in each iteration. It appears that "vanilla" natural selection performs just as efficiently as any "intelligent" variant of the idea. Maybe Intelligence has a lot less practical value than you think.

Another aspect of the two methods described here that make them different from the older method is that these methods can result in designs that are temporarily worse than the previous iteration. There is a possibility that the "mutations" that are being combined in one generation result in a worse cost function in the next. This is in contrast to the first method where the cost function can never reduce from one iteration to the next. The possibility of a temporary decline is a price one needs to pay for the possibility of faster increase in the cost function. Intelligence, invariably implies fallibility. The theory of natural selection, apparently, does not allow species to get worse temporarily. That might disqualify these methods from falling into the category of Genetic Algorithms. Maybe the biologists have some thing to think about. Can intelligent species temporarily decline?

Who watches the Watchmen!

I read this quote in Dan Brown's Digital Fortress. Quis Custodiet ipsos custodes? A lot of us have been thinking about the question that the quote asks. The Hindu carries on its front page the story behind why much of India (granted it is urban, middle class, cable tv watching India) is on the topic now. Before we go any further, there is a petition you can sign to support the cause. There is a group on facebook that you can join as well.

Back to the question. It is the public that has perform the role of watching over the watchmen. Some would argue that the pressure and prejudice so generated is unfair to the accused. I guess that's a good reason to ensure that the trial is speedy. The people who are getting involved would be much happier if they did not need to get involved in the first place. It is heartening to note that people don't shy away from this responsibility when needed though.

The other arms of the government, namely the judiciary and legislature also need to be part of the solution. The public involvement certainly makes the legislature interested. I believe that the judiciary would have done the right thing (albeit slowly) in any case, including pointing out that the investigating "executive" has acted to favour another wing of its own.

Such cases almost always bring the question of "Trial by Media" and its fairness into question. I think media has been a tool the public has used to further its cause throughout history. I have not heard many people complain about the role of the print media in the freedom struggle. Why do we do so in such cases?

Does the media keep the people interested in these cases to further its own cause? Is public opinion being manipulated? The responsibility to present the facts truthfully lies with the media. It is hard to believe that so many people are acting in concert to whip up public sentiment. I also believe that the public needs to get their facts from multiple sources to ensure that they are not being manipulated.

Actually all I wanted to say was Jail the b@$!@^d ! NOW! It would have been easier to do that. But then hyper-intellectual BS is bliss. In case you do not notice, I did not say Hang the b@$!@^d! NOW! Aah! That feels so good!