Chapter 3 was about images. I had been a lawyer and members of my family will tell you that I’m not a visual person. I like numbers but I swim in words. And for words, a different form of deep learning is dominant: a Recurrent Neural Network (RNN). And RNNs also learn from examples. But, in order to “understand” a specific topic, the examples must come from (and be related to) that topic.
The bottom line of RNNs is that the training data comes from examples that are labeled (the word Richard Socher used), or categorized or classified.1
Initially, when I found it in 2015, I was impressed with a YouTube presentation in November of 2012, in which Rick Rashid, then the head of Microsoft Research was speaking to a Chinese audience. ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access