本文主要介绍深度学习以及复杂layman形式运算的深度网络。通过真实数据的实验，定量比较，证实（或证伪）深度神经网络理论在外汇交易中的优势。当前主要用途是分类，基于深度神经网络模型创建一个指标和一个EA，根据客户端/服务器的方式进行运作，并对它们进行测试。(This paper mainly introduces deep learning and deep network of complex layman form operations. Through real data experiments, quantitative comparison proves that (or falsified) deep neural network theory has the advantage in foreign exchange trading. The current main use is classification, based on a deep neural network model to create an indicator and a EA that operates on the client / server way and tests them.)