 How good does the AI system perform, and how many intensive human work should be involved? And taking an example of the spiritual conditions, there are hundreds of thousands of speeds that were carefully human transgressions are needed. Another issue of the AI system is very powerful competition ability is needed. Like Africa, 1,000 more CPUs, and the current AI system cannot work without the network connection. With the competition power and the big data, we need a big container. D-Model is such kind of big data that can make steady progress along with the increasing of the data. And we can see the big difference for the AI system between the AI systems and the human brain in the way of the learning and learning efficiency. And the more big difference is that actually the system is independently disconnected. They cannot show the knowledge and the data. The human brain growing like a tree with a consistent action with the environment. And we can see the neural connection changes over the development from the newborn baby to the adult. And what kind of aspect we can learn from the brain. Let's see the left-down pictures with the many animals and overlap. And through the convolutional neural network, we just add a backward connection in the network. And a very good separation of the animals and the recognition can be a product. This is another example from the top-down view of the brain-inspired models. That models are different brain regions working coordinate from sending the information and makes the output decisions without human interventions. And this is more like brain-like models that using the spiking neural network neurons and purely learning using the biological rules achieves a comparable result with the state or art of the deep neural network but with very little less energy. So currently AI system based on the big data can be consistently learned from the brains. Even plus the wearable equipment, we can let the robotics to learn the human actions from the human's movement parameters. The state of the art of the AI is that the machine intelligence is rapidly approaching the human abilities in perception intelligence but still far away in cognition level, especially for natural language understanding and like the semantic modeling reasoning. Most significant output of the brain-inspired model is brain chip with different architectures and together with many training tools and learning data that they can integrate together to solve some vertical applications. So more about brains in this slide, the high process that brain can optimize the cost of functions and the cost of focusing at adverse from the areas and over the development. So through this research way, the more adaptive models can be achieved. Fortunately, if we can lead to the more general intelligence that executes all the tasks and can show the knowledge, thanks.