PDF Introduction to Deep Learning The MIT Press Eugene Charniak 9780262039512 Books

By Edwin Elliott on Thursday, May 23, 2019

PDF Introduction to Deep Learning The MIT Press Eugene Charniak 9780262039512 Books





Product details

  • Series The MIT Press
  • Hardcover 192 pages
  • Publisher The MIT Press (January 29, 2019)
  • Language English
  • ISBN-10 0262039516




Introduction to Deep Learning The MIT Press Eugene Charniak 9780262039512 Books Reviews


  • I highly recommend this text to anyone getting started with deep learning. It's a much quicker read than the standard Goodfellow et al text, which was really the only option for quite some time. This book has enough detail to get you started, and lots of practical advice on building real neural networks. It's a much more practical way to get started. If you read this and want more detail, you should move on to the Goodfellow book, but I think almost everyone should start here.
  • I was wondering how the author could pack so much information in less than 200 pages on deep learning. Then after a few days reading, I got impressed that the contents were quite accessible and practical for programmers like me. Also, it helped me find missing parts that I didn't know where to fit in when I was reading theory books. However, I encountered some typos here and there that disrupted the reading flow.
  • this is a good (but hardly great) price for such a book. if you are familiar with Python, but not a professional programmer, not at all with other languages (particularly C/C++ for memory management), you are enthusiastic about creating neural nets yourself, but find the why and how of it all too boring, then this COULD be a good place to start. however, this book also isn't quite clear enough as simply a list of instructions, to follow by rote.

    for some reason there did not seem to be a proof-reader? usually MIT Press is great about this. but besides the typos (forgivable), the grammar is convoluted. but far worse, the points of each paragraph are usually lost. it reads as if the original manuscript was twice as long and they just deleted every other sentence. one thought rarely connects to the next.

    it also appears the author does dutifully refer to other reading materials. these seem to have been assigned in some computer class, discussed, but the source material never really understood. i plodded ahead and read the whole thing, but sometimes someone goes on and on about a complex subject, and you realize this is so hard to understand because the person talking doesn't actually understand it even worse. but the author must really, really want folks (strangers?) to be impressed.
  • Very good, I'm happy with it, have examined a lot of other titles, this one provides the missing ring between theory and implementation, even if you're an expert, you'll like it.
  • Bought for my husband. He loves it