A lil old news, but certainly very important one… Image a robot which does not need any human supervision, no manual energy needs, can survive in the wild.. Scary thought, huh! Well, scientists at the University of the West of England (UWE) have designed a robot that does not require batteries or electricity to power itself.

Instead, it generates energy by catching and eating houseflies… If they could only add reproduction and replication ability to it, a next generation of evolution would have kick started..

Click here to read the story.

Perhaps the biggest achievement of man ever was putting man on the moon. I am not talking about the event itself but the abstract idea of the ability of a thinking machine (man) to surprise his future (us) to the grandest extent possible.

Zee Studio channel on the TV is showing the series “From The Earth to The Moon”. The serious is so perfect that it induces a druggy effect on me. When I imagine what was probably going on in the mind of Neil Armstrong, I just can’t express the pleasure I derive from it.

Apollo mission was, of course, a great feat. The moon which we see from here, so beautiful, has been visited by a man. Not just the US, and not even humans, but whole of the earthly life should be proud of it.

I keep on thinking how lucky would be the people who actually witnessed this event. Why was I not born before??? Can such a moment of this scale be witnessed again?? No, man going to MARS would not be a great deal. This is because we have already visited a non-earthly body (moon). Perhaps some other wild abstract idea can arouse same feelings. I think meeting an ALIEN life form could be equally, if not greater, event in earth’s history.

But, I wish I would be the first man to see an Alien….

Hi All,

Artificial Neural Networks have been trained to detect faces. However, is it possible to train a network to pass a judgment on how beautiful a face is? It can rate a face on some scale, say 1-10.

Secondly, can ANNs be trained to distinguish between male, female or not a face??

Wouldn’t it be wonderful to get judgments on your beauty by a software??

Regards,
Paras Chopra

Here is an interesting article titled “Biowar for dummies”. The author claims that with DNA hacking getting easier and cheaper, terrorists can make dangerous-than-nukes bombs without the need of an expert.

This is scary. Something needs to be done to avoid biowars.

This is a true example of biohacking and hopefully researches like this will continue to exploit biology for its complexity and flexibility.

Now to actual story. “”The Physarum polycephalum slime, which naturally shies away from light, controls the robot’s movement so that it too keeps out of light and seeks out dark places in which to hide itself.”"

This essentially means that the slime mold is connected to robotic arm in the same way our minds will be directly connected to internet (if they would still call it so) in future.

Click here to read the story.

What I believed that one gene codes for just one polypeptide seems to be incorrect. I am reading the book Intelligent Bioinformatics where AI techniques relevant for bioinformatics are discussed. In the preliminary chapters, the author says that the belief that one gene can only code for one polypeptide is wrong.

Honestly, I was not aware of this. But, the author says that one gene can produce many different polypeptide chains depending on how many introns are removed and how many exons are shuffled or differently spliced.

Isn’t this cool? The same information can be processed in more than way and having multiple interpretations. It is like same software producing different output depending upon environment. What really amazes me is the fact that cells achieve all this without someone actually telling it how do do it. There is no central authority, yet the cell has conceived extreme examples of complexity. Many more things are waiting to be discovered.

I believe that no matter how clever we become or in how much detail we explore the cell, biology always has something to surprise and amaze us.

I just wish I keep on discovering such facts.

Ultimately, I have decided to release the code for my algorithm Explore. The main objective of this algorithm is to optimize a function or to search for a solution in a finite n dimensional space.

[URL=http://www.paraschopra.com/sourcecode/explore.zip]Click here[/URL] to download source code in Python.

Yesterday, I was thinking that almost all approaches to Artificial Intelligence attempt to simulate parallelization. Be it neural networks, Genetic Algorithms, Searching methods, etc. All attempt to simulate techniques which depend upon parallelization.

This gives me an intuitive feel that the very basic requirement of intelligence is parallel processing. Hundreds of processes taking place simultaneously. In computers we have a serial processor. However, if we have a massive parallel computer (no matter how weak individual processing items may be) we can create artificial intelligence.

In contrast this, our own mind, even though is massively parallel, works in a serial manner. When you are concentrating of one thing, it is very difficult to think of other things. It seems it is some kinda paradox.

Paradox: Natural intelligence arises by simulation of serial computation on parallel hardware. And Artificial Intelligence works by simulating parallel computation on serial hardware.

How can this paradox be resolved?

Presently, I am coding a neural network for protein secondary structure prediction. Initially, I thought of predicting contact maps. But, the I coudn’t find contact map data for proteins. Moreover, parsing the PDB files and then creating a contact map seemed as a daunting task to me.

Therefore, I chose the secondary structure prediction. Well, what I am gonna do is to input a amino acid sequence and then predict secondary structure for each residue. Current methods achieve around 75-80% accuracy, so I will try to achieve atleast 75% accuracy.

My process will be a 3 stage process. First two will involve a neural network and the third one assigns confidence to prediction based on statistical information.

Any ideas, comments or suggestions?

I have an idea of Spam detection using DNA (Hybridization or Sequence Alignment). Like represent each English alphabet by a codon. Lets say we represent ‘a’ by AUG and ‘@’ by AUT. Now we can convert the incoming message into a DNA sequence and search for patterns such as ‘V-i-a-g-r-a’ or ‘C@sin0′ via standard principles of sequence alignment or Hybridization.

There are 3 factors:
1. No of words recognized
2. If hybridization is used, Temperature at which words are recognized
3. How good is a match

Then using the combination of these 3 factors, we can determine if a message should be classified as spam or not.

Any comments, suggestions or questions?