10.11.2016 Views

Learning Data Mining with Python

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Classifying Objects in Images Using Deep <strong>Learning</strong><br />

The following screenshot shows the Mars rover designed and used by NASA; it<br />

made significant use of computer vision:<br />

Deep neural networks<br />

The neural networks we used in Chapter 8, Beating CAPTCHAs <strong>with</strong> Neural Networks,<br />

have some fantastic theoretical properties. For example, only a single hidden layer is<br />

needed to learn any mapping (although the size of the middle layer may need to be<br />

very, very big). Neural networks were a very active area of research in the 1970s and<br />

1980s, and then these networks were no longer used, particularly compared to other<br />

classification algorithms such as support vector machines. One of the main issues<br />

was that the computational power needed to run many neural networks was more<br />

than other algorithms and more than what many people had access to.<br />

Another issue was training the networks. While the back propagation algorithm has<br />

been known about for some time, it has issues <strong>with</strong> larger networks, requiring a very<br />

large amount of training before the weights settle.<br />

[ 246 ]

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!