artificial intelligence -- Artificial intelligence , By studying human intelligence , Understanding human intelligence （ see , hear , say , write , smell , Ability to think, etc ） The essence of , Produce machines with human intelligence .
big data – High density , Large volume , Multiple dimensions , High value data .
The relationship between AI and big data
1, The value of big data
* Decisions without data support are often inaccurate .
In a stable income group , About three to five adults are speculating , Statistics show that 95% Our personal investment can't win the market in the end ,50-70% Frequent short-term transactions and even loss of money . But my friends around me have not low education at the same time , Investment performance is not better than this figure . Then why do they want to speculate , On the one hand, I am quite confident in my ability to speculate in stocks , On the other hand, see the example of who makes money around , Make them feel easy to make money . But just look at the statistics , We'll come to the opposite conclusion .
* Significance of large amount of data
2005 year , I've never done machine translation before Google We got Franz, an expert in machine translation . Oko , A year later, the best machine translation system in the world was built , stay NIST In the annual review results of ,Google Of BLEU score 51%（ The accuracy of manual translation is only 50%-60%）, Second place ahead 5%, However, translation based on grammar rules SYSTRAN Only for 10.79%. The secret of Oko is ： Method or two years ago , But it uses thousands or even thousands of times more data than other research institutions , Train a six element model . generally speaking , To estimate N Conditional probabilities of metamodels , Enough data ,N The bigger , More data ,N Generally not more than 3. If you use two or three times more data , Machine translation will work better , But tens of thousands of times more , The accumulation of quantitative change leads to qualitative change , Achieve good results .
* The importance of big data
In healthcare , Genetic defects are related to many diseases , To find out the relationship ：
To find out the mechanism of a certain gene by experiment （ It could be a long process , Maybe we need to start with the study of flies ）, And the physical changes brought about by its defects , And then figure out what kind of disease this change might cause , Or under what circumstances . Scientists have studied for decades , It's hard to find the causal relationship between smoking and many diseases .
The statistical method of data is the opposite of these methods , Starting from data , Finding the statistical correlation between gene defects and diseases , And then analyze the internal causes of this correlation .
Take diabetes for example ,
A The representative gene is defective ,B Diabetes .# Number of representative samples .
Big data is very useful , If there are more , More complete , More comprehensive data , We can find unexpected surprises out of it . AI's intelligence is contained in big data .
2, AI enables all walks of life
AI with the development of big data , Develop intelligent application incisively and vividly , Widely used in all walks of life . Including smart home , Smart Finance , Intelligent customer service , Intelligent manufacturing , Intelligent medicine , Intelligent art creation and other fields .
- Smart home ：Echo loudspeaker box , Intelligent cooking robot , Intelligent security and so on . Embedding AI technology , Just by touching , gesture , speech recognition , Face recognition can realize various functions .
- Smart Finance ： By tapping the financial market , Risk change , Rules of user data , Make personalized service for users . Including intelligent investment consulting industry , insurance industry , Credit industry, etc .
- Intelligent customer service
： Through speech recognition technology , Natural language processing technology , Speech synthesis technology for human-computer interaction , For customer service , From business consulting , Professional questions and answers , Business handling , Complaint feedback can replace manual customer service , So as to reduce customer service pressure , Reduce enterprise operation cost , Improve customer satisfaction .
- Other applications ：