Like a pianist who can play skillfully without looking at the keys, mechanical engineers at the University of California, Los Angeles, designed a new material that can learn behavior and develop its own "muscle memory" over time, allowing real-time adaptation to changing external forces. The material is composed of a structural system with adjustable beams, which can change its shape and behavior according to dynamic conditions. The research, published in Science Robot on the 19th, will have important applications in architecture, aircraft and imaging technology. Jonathan Hopkins, a professor of mechanical and aerospace engineering at the University of California, Los Angeles School of Engineering, who led the research, said that this artificial intelligence material can learn the behavior and characteristics that should be displayed when exposed to environmental conditions. For example, when the material is placed in the wing of an aircraft, it can learn the stroke mode during flight and change the shape of its wing to improve the efficiency and maneuverability of the aircraft; The building structure injected with this material can also self adjust the stiffness of some areas to improve its overall stability during earthquakes or other natural or man-made disasters. Scientists have used and adjusted the existing concept of artificial neural network. Artificial neural network is exactly the algorithm driving machine learning. Researchers have developed the mechanical equivalent of artificial neural network components in the interconnection system. This mechanical neural network consists of individually tunable beams oriented in a triangular lattice pattern. Each beam has voice coils, strain gauges, and flexors that enable the beam to change its length, adapt to changing environments in real time, and interact with other beams in the system. Then, the optimization algorithm controls the whole system by obtaining data from each strain gauge and determining the combination of stiffness values. In order to check the effectiveness of the strain gauge monitoring system, the research team also used the camera trained on the system output node. The early prototype of the system lags between the input of applied force and the output of mechanical neural network response, which affects the overall performance of the system. The team tested multiple iterations of strain gauges and bending in the beam, as well as different lattice patterns and thicknesses. The final design overcame the lag and accurately distributed the applied force in all directions. At present, the system is about the size of a microwave oven, but researchers plan to simplify the design of mechanical neural networks so that thousands of networks can be manufactured in 3D lattices at microscale for practical material applications. Why does the editor in chief tick a material to learn by himself? This depends on the artificial neural network, which endows this new material with intelligent and adaptive characteristics. In fact, the same basic principles have been used in the hot machine learning in recent years. In the future, in addition to using this new material in vehicles and building materials, it can also be used in the battlefield, such as integrated into armor, so as to deflect the shock wave; Or used in the medical field, acoustic imaging technology will also be greatly developed. (Liu Xinshe)