Internet resources for learning artificial intelligence (AI), especially regarding the subtopic of machine learning (ML), are expanding rapidly. Two resources, presented as online books, are the subject of this review. Both resources overlap in coverage of ML and deep learning (DL), and both are called “books” (PDFs), but other than that the comparison is apples-and-oranges different.
Neural networks and deep learning is free, whereas Deep learning for computer vision with Python is not. Neural networks and deep learning was posted at a specific date, whereas Deep learning for computer vision with Python and its supporting information are updated periodically to stay current (Dr. Rosebrock is very attentive to emailed questions). Neural networks and deep learning explores the mathematics of neural networks in depth, whereas Deep learning for computer vision with Python skims over the math and concentrates on coding. Neural networks and deep learning comes across as a labor of love, an ode to a beautiful concept, whereas Deep learning for computer vision with Python is entrepreneurial outreach, information sold at a reasonable price so the buyer can, in turn, use it to make their own money. Neural networks and deep learning should be read outdoors in the summer shade with a cold drink, reclining peacefully under an azure sky, whereas Deep learning for computer vision with Python is raw “git ’er done!”--a pure and beautiful concept in its own right--meant to be open in one of two monitors with your hands poised over the keyboard.
So which one is better? Wrong thinking. There will probably be one that you prefer, but they’re not mutually exclusive. Read both. Absorb as much mathematical foundation as you can from Neural networks and deep learning, then start in on Deep learning for computer vision with Python and start coding. Or, read them in parallel as complementary sources that reinforce each other (the method I used). There’s no reason to rule out either one.
Note that Deep learning for computer vision with Python is sold as one of three packages. There is the lowest-cost “Starter Bundle,” meant to teach ML. The “Practitioner Bundle” adds resources to the “Starter Bundle” to take the student deeper into practical coding of computer vision applications and best practices. The final “ImageNet Bundle” includes data and instructions to build large-scale neural nets. Visit https://www.pyimagesearch.com/ to get more information. Neural networks and deep learning can be found at http://neuralnetworksanddeeplearning.com/.
Before closing this review, however, I have to mention a third resource. David Venturi of freeCodeCamp (https://www.freecodecamp.org/) has collected “Every single machine learning course on the Internet, ranked by your reviews” . His goal was to create his own online data science master’s program, and as a side effect created a great resource for anyone wanting to investigate ML. It’s well worth the time it takes to read his post.
More reviews about this item: Goodreads