深度学习面试题

1. AI, Machine Learning和Deep Learning的区别

2. Deep learning的限制

• Deep learning usually requires large amounts of training data.
  Deep learning通常需要更大量的训练数据

• Success of deep learning are purely empirical, deep learning algorithms have veen criticized as uninterpretable “black-boxes”.
  深度学习的成功纯粹是经验主义的，深度学习算法被批评为无法解释的“黑匣子”。

• Deep learning thus far has not been well integrated with prior knowledge.
  到目前为止，深度学习还没有很好地与已有知识相结合。

• Deep neural networks are easily fooled.
  深度神经网络很容易被愚弄。

来源：