说明：  一个双向LSTM程序 Long Short Term 网络—— 一般就叫做 LSTM ——是一种 RNN 特殊的类型，可以学习长期依赖信息。LSTM 由Hochreiter & Schmidhuber (1997)提出，并在近期被Alex Graves进行了改良和推广。在很多问题，LSTM 都取得相当巨大的成功，并得到了广泛的使用。 LSTM 通过刻意的设计来避免长期依赖问题。记住长期的信息在实践中是 LSTM 的默认行为，而非需要付出很大代价才能获得的能力！ 所有 RNN 都具有一种重复神经网络模块的链式的形式。在标准的 RNN 中，这个重复的模块只有一个非常简单的结构，例如一个 tanh 层。
(A bidirectional LSTM program Long short term network, commonly known as LSTM, is a special type of RNN that can learn long-term dependent information. LSTM was proposed by Hochreiter & schmidhuber (1997) and recently improved and promoted by Alex graves. In many problems, LSTM has achieved great success and has been widely used. LSTM is designed to avoid long-term dependency. Remember that long-term information is the default behavior of LSTM in practice, not the ability to acquire it at a great cost! All RNNs have a chained form of repetitive neural network modules. In the standard RNN, this repetitive module has only a very simple structure, such as a tanh layer.)