How can intelligence analysis be improved

Analysis of the four elements of artificial intelligence

The intelligence of artificial intelligence is contained in big data.

Computing power provides basic computing power support for artificial intelligence.

The algorithm is the fundamental way of realizing artificial intelligence and an effective method for obtaining data intelligence.

Big data, computing power and algorithms are used as inputs. The actual value can only be reproduced if the output takes place in actual scenarios.
To give a very clear analogy: if we look at cooking as our scene, big data corresponds to the ingredients needed to cook, and the computing power corresponds to the gas / electricity / firewood needed to cook. Algorithm It corresponds to cooking methods and spices.

——- Next I will explain the four elements of artificial intelligence one after the other ——-

1) big data
During this time, big data is constantly being generated. Data collected on mobile devices, cheap cameras, ubiquitous sensors, etc. This data takes various forms and most of it is unstructured data. If it is to be used by artificial intelligence algorithms, a lot of preprocessing is required.

2) computing power
The development of artificial intelligence places higher demands on computing power. The following is a comparison of the computing power of different chips. Among these, the GPU is the most widely used in the field of artificial intelligence, ahead of other chips. Both the GPU and the CPU do well for floating point calculations. In general, the GPU's floating point capability is about ten times that of the CPU.

In addition, the Deep Learning Acceleration Framework is optimized on the GPU in order to improve the computing power of the GPU again, which contributes to the acceleration of the calculation of the neural network. For example, cuDNN has a customizable data layout, supports the flexible dimensional sorting of four-dimensional tensors, steps and sub-areas and is used as the input and output of all routines. In the convolution operation of the convolutional neural network, the matrix operation is realized while reducing the memory and greatly improving the performance of the neural network.
3) algorithm
The common algorithms are mainly divided into conventional machine learning algorithms and neural network algorithms. The rapid development of neural network algorithms has reached a peak in recent years due to the development of deep learning.

4) scene
The classic application scenarios of artificial intelligence include:
1. Analysis of the user portrait
2. Risk control based on creditworthiness
3. Fraud detection
4. Robo-Advisor
5. Smart review
6. Intelligent customer service robot
7. Machine translation
8. Face recognition