AI-Assisted Statistics for Data Scientists, 3rd Edition
by Peter Bruce, Andrew Bruce, Peter Gedeck
Chapter 9. Deep Learning
Neural nets had their greatest impact following the 2006 publication of a pair of papers by AI researcher Geoffrey Hinton1 and the rapidly growing popularity of what came to be called deep learning. Deep learning involves large and complex networks with many layers, incorporating processes for dimension reduction and feature discovery. The advent of big data, and of powerful graphics processing unit (GPU) chips, has accelerated the growth of deep learning. Deep learning has revolutionized fields such as computer vision, natural language processing, speech recognition, and generative AI, enabling breakthroughs in applications like image classification, machine translation, and voice assistants.
Advantages of Depth
The data we have dealt with so far in this book is structured data, typically from organized databases. The predictor variables, or features, that we think might be meaningful in predicting an outcome of interest already exist in our data. In predicting loan default using the LendingClub data Table 6-1, for example, we used features that have clear predictive value such as debt-to-income level or the ratio of loan payment to income. With tasks like text mining and image recognition, structured high-level predictor information like this is not available. Even with some structured data, there may be complex, nonlinear relationships among the features that are not easily captured by traditional statistical models. An example of the latter is complex ...
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