1. Deep learning and artificial intelligence. In essence, artificial intelligence is a broader concept than deep learning. At present, artificial intelligence can be divided into weak artificial intelligence and strong artificial intelligence In fact, the so-called "artificial intelligence" that can be realized by current technology is weak AI, while that of Ultron is strong AI (even boss level). Deep learning, a technology or idea in AI, was once listed as one of the top ten breakthrough technologies in 20 13 by MIT Science and Technology Review. Or in other words, deep learning (I prefer to call it thinking, that is, end-to-end) may be the breakthrough to achieve strong AI in the future.

2. deep learning and ML. There is a subtle relationship between DL and ML. When DL has no fire, it exists in ML as a neural network abbreviation learning algorithm. With the rise of computing resources and big data, neural networks have become today's DL. There are generally two views on DL in academic circles. One is to regard DL as a feature extractor, which is only used to extract powerful features; ; Another hope is to develop it into a new branch of learning, which is the end-to-end "deep learning thought" I mentioned above.