Sunday, June 6, 2010

Maths and AI

Maths is so complex that most of people don't want to touch with it.But most of our science of computer is related to maths.Most of discovery of AI is related to maths now.But if all of science of AI is related to maths closely?

When we was born from the god, we didn't know what is maths and what is induction, but we can realize Mum is most likely long haired and dad is most likely short haired.We can classify, But we do not know the maths principle of classification. Until now,most of us can do something on schedule, but most of us do not know what is Operational Research.Although we may not know the principle of the maths, but we have intelligence yet.

So if all of our psychology action is according to maths principle?Obviously the answer is no.In fact, some maths principle is comes from our psychology action, for example the induction, not the reverse of that.

So if all of AI discovery is comes from maths?Maybe the answer is no.

Monday, April 26, 2010

Role of Psychology in Artificial Intelligence

Until now I haven't advert psychology,what is the role of psychology in A.I.?

Last article I mentioned that we need all of the data mining process is done automatically.How it can be done automatically? We need psychology. The secret of human intelligence is on the psychology,especially on cognitive psychology. We need artificial intelligence to imitate mental process ,I call it A.I. process.

For artificial intelligence, I consider the role of psychology is more important than artificial intelligence itself. If we are deep into artificial intelligence itself, we will find the complexity of the problem is beyond our imagination, even to uncomputable.But if we centralized into the connection of psychology with A.I., maybe we can find some discovery more easily.

Can A.I. have creativity? Can A.I. have emotion? Can A.I. have different process when it's like a women or reverse?Can A.I. have psychosis? All of the answer is up to the progress of psychology.

But the progress of psychology is lazy, this is the difficulty of artificial intelligence too.

Thursday, April 15, 2010

Data Mining in Artificial Intelligence

Data mining,also known as KDD(Knowledge Discovery in Database) is applied in many subjects nowadays,like commercial application,search engine etc.It partly comes from subjects of Artificial Intelligence ,like Machine Learning,Artificial Neural Networks etc.But can it be used by Artificial Intelligence?

My answer is Surely it can be.My lower article introduce the induction maybe is the key of artificial intelligence,but how to use induction in A.I. is a problem.Data mining maybe is the key of how to use induction.

But maybe someone will say:"Data mining has applied many years,but it's effect in A.I. is not doing well,A.I. progress stagnates yet. A sample is a lot of result of search engine is not applicable."His word is fact.But I will say:"It's not the fault of data mining,it's fault is on the way of how to use data mining in A.I.,and the algorithms of data mining has some space of improving."

We all know the process of data mining includes: problem definition,data collection and data preprocessing,data mining,result interpreting and evaluating.Except that data mining is done by computer automatically,other process need artificial operation or it is fixed process.If we need it be used by A.I.,we need all of the process is done automatically.

But how it can be done automatically,what is the process contains? The answer maybe is a key and the difficulty of artificial intelligence.

Sunday, April 4, 2010

Some Idea from Non-linear Classification of Perceptron Neural Networks

Concerning with perceptron neural networks, non-linear classification is a problem.It base on the mathematical output of all input,It use some kinds of activation function, multilayer perceptron neural networks to solve the problem.But it makes it more complex,in addition,the learning algorithm precision may fall too.The activation function is a mathematical output,in fact,it's a deduction output.Like some other mathematical theorem,it needs so complex logic to prove.

Can deduction logic prove all of problem in reality?Can all assertion proved by logic?Maybe can not.Maths has a lot of axiom,it's output is proved by human experience,i.e. by induction.Maybe the induction is the key to artificial intelligence.The induction needn't complex logic operation,but it's structure needs more careful design.

The Importance of Artificial Intelligence

Last century,our science and technology develop rapidly,including the application of electricity,nuclear power,computer,internet etc.But untill now,the first decade of this century,it seems that the science and technology stagnate.It behave as the slowdown of economics in USA and other country,because the base of USA is on science and technology.

But why,why we can't advance more quickly? Can we advance more quickly than last century?

This is more difficult to answer.For example, if you search the word 'artificial intelligence' in google,you will find 9,550,000 results. If you adhere to read 1000 web pages every day(I think I can't achieve the goal),it will takes more than 26 year to read over the 9,550,000 results.Apparently,if you will research the 'artificial intelligence' ,reading the 9,550,000 result is far from the goal.

In fact,nowadays our science and technology subjects are more like intercrossed.If you need to find some odd thing ,you have to study a lot of subject.

Using search engine is a good way to discovery.But it's not the best way.First,a lot of results of search engine are not related to our discovery goal,you need to get rid of the result,it will takes more time to do it like above sample.Second,the connection of subject can't be use in search engine easily.

My answer is artificial intelligence is the key of developing science and technology.If AI is developed,it will be like the Industrial Revolution in human intelligence.Wish we can find the key of artificial intelligence.