Parallel structure and parallel processing power. Artificial neural networks are structurally parallel, and the processing order is parallel and synchronized. In other words, the computational function of the neural network is distributed across multiple processing units.<br>(2) Knowledge distributed storage. In neural networks, knowledge is not stored in a particular storage unit, all memorable information is stored in the interconnect weights between neurons, and a single weight does not have a significant impact on the performance of the entire neural network.<br>(3) fault tolerance is strong, the human brain has a strong fault tolerance, brain cell metabolism does not affect our memory and thinking ability. Similarly, artificial neural networks also have a strong fault tolerance, local or partial neurons failure, will not affect global activities.
正在翻译中..