Reuters has a nice little blurb about having a robot learn to fly by using genetic algorithms [here]. Many people do not realize that certain machine learning techniques which imitate nature are capable of solving problems that would be hard or impossible for a human programmer to solve. Genetic algorithms, artificial neural networks and swarm intelligence algorithms are a few of the techniques that we have borrowed from nature. Interestingly enough, I had a conversation with a programmer friend of mine last night who was not familiar with the state-of-the-art in machine learning and was trying to make the age-old argument to me that only an intelligent being could create the complex things around us. This argument is commonly known as the "intelligent watchmaker" argument. Research is showing at an increasing frequency that nature is very capable of solving complex problems without the intervention of any outside intelligence. |