An Adaptive Optimization Method Based on Learning Rate Schedule for Neural Networks

Artificial intelligence (AI) is achieved by optimizing the cost function constructed from learning data. Changing parameters in an AI process (or for convenience). If well performed, then value of global minimum. In order to obtain well-learned learning, parameter should be no change at One useful optimization method momentum method; however, has difficulty stopping when satisfies minimum (non-stop problem). The proposed based on method. solve non-stop problem method, we use our Therefore, as processes, mechanism reduces amount effect function. We verified through proof convergence and numerical experiments with existing methods ensure that works well.