News classification
Contact us
- Add: No. 9, North Fourth Ring Road, Haidian District, Beijing. It mainly includes face recognition, living detection, ID card recognition, bank card recognition, business card recognition, license plate recognition, OCR recognition, and intelligent recognition technology.
- Tel: 13146317170 廖经理
- Fax:
- Email: 398017534@qq.com
Artificial intelligence analysis
Artificial intelligence analysis
Artificial intelligence is divided into several levels. First of all, the underlying layer is large data cloud computing. Because you have large amount of data, you need to put it in the cloud to deal with it. Big data, cloud computing, GPU/FPGA and other hardware acceleration, new form of neural network chips and other computing providers. In the technical layer, it is to do various algorithms, such as machine learning, deep learning, and strengthening learning. Application layer is the application of all kinds of applications, such as intelligent advertisement, intelligent diagnosis, automatic writing, identity identification, intelligent investment advisor, intelligent assistant, self driving vehicle, robot and so on.
What is big data, the importance of listening to other people's data every day, and what is big data? From the data definition, if it is used on all meter disposal of structured data is not big big data, the concept of big data should be said that it is the source of diversity, its structure is heterogeneous, is unstructured data, it is not only a large amount of the data. It is out of order, according to information theory, entropy, information quantity is very large, this is the big data. Big data which is the most important correlation and causality, a lot of people including scientists, some fantastic, very vague about big data mining related to magic expression, this is wrong, just discover the correlation is not enough, but also to analyze causality, A launched B or B launched A, or AB are introduced. You only use data analysis to figure out that they are related, and there is a certain vague and unsure connection between them is not enough.
For example, A and B, you can discover that A and B have some kind of correlation, which is not enough. This vaguely ambiguous correlation is in the key business scenario, and you can't use it for reference. I need to make a profit in the stock market, only the correlation is not to use the stock trading algorithm to make arbitrage. In artificial intelligence data analysis calculation there are many algorithms, I want to say is that some algorithm is useful in a specific category in many algorithms, I talk about the algorithm, my background is the calculation theory of logic background, I emphasized on any one industry, to analyze the excavation inside the pain point from the perspective of the origin and theoretical logic. If you use machine learning or neural network, you can calculate the resolution bias, which is bias, if not means that your algorithm is not informed of certain black box algorithm, the algorithm is useful you but you have no way to prove that you are right as long as the algorithm, Bayesian statistics can be calculate the resolution bias. What is the criteria of science, there is also a Bayesian hierarchical Bayesian, now popular is the deep learning neural network inside multilayer Bayesian network to dividends, belongs to the multilayer, and with Bayesian networks can be used to explore the data behind, so Bayesian network can make some deep learning can not do. For example, a wide range of infectious diseases such as the spread of SARS nodes to explore, for example like SARS, bird flu, if starting from Beijing, Wuhan, Zhengzhou, after Ji'nan, but some of the city's communication node from the infectious disease statistics chart and the data inside is not, as long as this situation with hierarchical Bayesian the network, can discover the spread of infectious diseases hiding nodes, we can discover hidden relations between nodes, but also can discover hidden nodes behind a layer of nodes, based on the statistical data of infectious diseases, as long as a way to discover the relationship between nodes and implicit, other deep learning machine learning approach regardless of all.
Just as we know the probability map, the industry now in the natural language understanding research inside machine learning to use the best, is that it can use a large amount of data to Machine Translation, but only the application of the traditional Machine Translation, this way the statistical significance on the traditional to find corresponding relation on a wide range of data, it is not enough. The latest research in the study is to use the probability graph to do the natural language understanding and machine learning, which can achieve better translation results.
There has been a famous debate recently that deep learning is the overthrow of everything. Is said to have deep learning what are capable, there is another problem, alchemy good or chemistry, if not quite understand the definite conclusion that it can not prove it, the result is qualitative, so it is a kind of alchemy alchemy, behind every element is how to response. They remind the repercussions of the chemical law is clear, this is science. What is the name of science, science only conclude that the specification is certainty, whether with certainty, if you find a rule, before the miscellaneous no chapter of this situation, with some certainty, as long as the law is in the affirmative, what you find is really a scientific law, if not sure it is not a science. Can data science be established? Everyone in the world are now speculation, speculation, big data artificial intelligence data science, if the criteria data science is only in this way was, not sure whether or not the correct way to determine if it is not a science, he is only a statistical result of statistics in science speaking, all of the scholars believe that the statistics is not scientific, because it is random.
We are now talking about small data and zero data. Now many companies say the key to artificial intelligence is whether it can have big data.
What is big data, the importance of listening to other people's data every day, and what is big data? From the data definition, if it is used on all meter disposal of structured data is not big big data, the concept of big data should be said that it is the source of diversity, its structure is heterogeneous, is unstructured data, it is not only a large amount of the data. It is out of order, according to information theory, entropy, information quantity is very large, this is the big data. Big data which is the most important correlation and causality, a lot of people including scientists, some fantastic, very vague about big data mining related to magic expression, this is wrong, just discover the correlation is not enough, but also to analyze causality, A launched B or B launched A, or AB are introduced. You only use data analysis to figure out that they are related, and there is a certain vague and unsure connection between them is not enough.
For example, A and B, you can discover that A and B have some kind of correlation, which is not enough. This vaguely ambiguous correlation is in the key business scenario, and you can't use it for reference. I need to make a profit in the stock market, only the correlation is not to use the stock trading algorithm to make arbitrage. In artificial intelligence data analysis calculation there are many algorithms, I want to say is that some algorithm is useful in a specific category in many algorithms, I talk about the algorithm, my background is the calculation theory of logic background, I emphasized on any one industry, to analyze the excavation inside the pain point from the perspective of the origin and theoretical logic. If you use machine learning or neural network, you can calculate the resolution bias, which is bias, if not means that your algorithm is not informed of certain black box algorithm, the algorithm is useful you but you have no way to prove that you are right as long as the algorithm, Bayesian statistics can be calculate the resolution bias. What is the criteria of science, there is also a Bayesian hierarchical Bayesian, now popular is the deep learning neural network inside multilayer Bayesian network to dividends, belongs to the multilayer, and with Bayesian networks can be used to explore the data behind, so Bayesian network can make some deep learning can not do. For example, a wide range of infectious diseases such as the spread of SARS nodes to explore, for example like SARS, bird flu, if starting from Beijing, Wuhan, Zhengzhou, after Ji'nan, but some of the city's communication node from the infectious disease statistics chart and the data inside is not, as long as this situation with hierarchical Bayesian the network, can discover the spread of infectious diseases hiding nodes, we can discover hidden relations between nodes, but also can discover hidden nodes behind a layer of nodes, based on the statistical data of infectious diseases, as long as a way to discover the relationship between nodes and implicit, other deep learning machine learning approach regardless of all.
Just as we know the probability map, the industry now in the natural language understanding research inside machine learning to use the best, is that it can use a large amount of data to Machine Translation, but only the application of the traditional Machine Translation, this way the statistical significance on the traditional to find corresponding relation on a wide range of data, it is not enough. The latest research in the study is to use the probability graph to do the natural language understanding and machine learning, which can achieve better translation results.
There has been a famous debate recently that deep learning is the overthrow of everything. Is said to have deep learning what are capable, there is another problem, alchemy good or chemistry, if not quite understand the definite conclusion that it can not prove it, the result is qualitative, so it is a kind of alchemy alchemy, behind every element is how to response. They remind the repercussions of the chemical law is clear, this is science. What is the name of science, science only conclude that the specification is certainty, whether with certainty, if you find a rule, before the miscellaneous no chapter of this situation, with some certainty, as long as the law is in the affirmative, what you find is really a scientific law, if not sure it is not a science. Can data science be established? Everyone in the world are now speculation, speculation, big data artificial intelligence data science, if the criteria data science is only in this way was, not sure whether or not the correct way to determine if it is not a science, he is only a statistical result of statistics in science speaking, all of the scholars believe that the statistics is not scientific, because it is random.
We are now talking about small data and zero data. Now many companies say the key to artificial intelligence is whether it can have big data.