In last post we have seen how computers evolved in last 5-6 decades to take on more and more functions resembling a human brain. We will continue this discussion in this post.
Now a days a lot of research is being done by computer scientists to make computers think like a human brain. Now computer technology is making much progress towards making computers to think by using technologies related to learning abilities. Human brain learns from experiences. When a toddler accidentally touches a hot object, it realizes what could be hot object and how touching a hot object could be harmful. in fact human brain develops and understands much of the physical world by learning from experiences. Without this learning ability of the human brain, an adult brain could be as good as that of a 6 year old kid.
How a computer scientist will make a thinking computer? Traditionally computers were a merely computation devices. You give some input in form of some instruction to a computer and the computer will do computation and give you output. If you want the computer to think then you will have to create learning libraries and install them with the computer. Whenever a computer encounters some problem and doesn’t know the answer, it will be given ability to search previous experiences which are similar to the present one (though not exactly the same) and infer and find out a close solution and then take it as a learning lesson. Next time, the same problem is posed to the computer, it will not have problem in giving the right answer. this is how a human brain works.
Creating such a functionality in computers has been difficult. it is becoming a reality only now. Techniques like fuzzy logic is used to infer a solution from searching close alternatives already existing in the memory of a computer and finding out the most close alternative out of all alternative solutions. Some searching algorithms used these days include Tabu search, annealing algorithm etc.