There are numerous algorithms or architectures of networks artificial   
   intelligence at present, while each of them only deals with a narrow   
   scope of applications, although many of them involves considerable   
   efforts and complicated methods. The current achievements impress us   
   that true intelligence is far to attain.   
      
   It was the borrowed idea from biology decades ago that contributes to   
   the advent of artificial neural network, which opened the new era   
   although the the model is much simpler than real brain. But since then   
   researchers have been concentrating on more and more complex   
   mathematical methods and rarely pay attention to new developments of   
   brain research. Nowadays it is reported pictures of dream can be   
   recorded; a monkey acts on electronic signals passed on to its brain   
   via an embedded electronic apparatus to its head.   
      
   Although I have not reviewed much in how our brain works. I am sure   
   there are must be limited kinds neurons. A neuron must be capable of   
   simple processing. Its huge population and connections make it   
   powerful.   
      
   Natural intelligence is based on life substances such as protein. But   
   lifeless silicon has showed  encouraging capacity of intelligence.   
   Computer calculates much faster than its human inventors; its vast   
   inventory remembers more accurately and permanently. And recent time   
   has seen the further development of AI.   
      
   Simple structure might be more flexible and efficient, as the   
   interesting discovery that slow learning MLP is equivalent to Taylor   
   expansion and function simulation is very efficiently realised using   
   polynomial[2]. Another example is that simpler polynomial networks are   
   easier to train[3]. It is easy to understand that using small bricks,   
   cements, sand and other basic materials, we can build a house into any   
   shape. The fact is, no matter how complicated a computer application   
   is, finally it is converted into basic digital operations, such as   
   Boolean operations, shifting, that are carried out in ALU of CPU. This   
   is similar to the way of brain.   
      
   I believe that digital network composing of neurons with very basic   
   operations such as Boolean and shift could be trained to compete with   
   human brain. It is time for AI scientists to cooperate with biology   
   researchers worldwide.   
      
   References   
      
   1. How does your brain work?   
   http://www.sciencemuseum.org.uk/on-line/brain/1.asp   
      
   2. Conventional modeling of the multilayer perception using   
   polynomialbasis functions   
   http://ieeexplore.ieee.org/Xplore/login.jsp?url=/iel4/72/4678/00   
   82712.pdf?temp=x   
      
   3. Polynomial Neural Networks   
   http://ulcar.uml.edu/~iag/CS/Polynomial-NN.html   
      
   [ comp.ai is moderated ... your article may take a while to appear. ]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)