Ask HN: Best books on AI? • Hacker News Board

Re: Ask HN: Best books on AI?

#41

Post by gugagore » Sat, Mar 18, 2017, 1:13 PM UTC

flor1s wrote: post #17

iamlucaswolf wrote: post #16

To iterate on what others said, but what was not emphasized enough from my point of few: AI is academic (as a synonym for 'theoretical' and 'math-intensive'). Once you look beyond purely symbolic AI, which proved to be infeasible as @curuinor pointed out somewhere here, you will need to build up at least basic knowledge in probability theory and linear algebra. The path I'm following at the moment is a quite rigorous…

I think to be successful at machine learning you also need a good understanding of calculus, besides probability and linear algebra.

but perhaps not too much: https://news.ycombinator.com/item?id=9189553

Re: Ask HN: Best books on AI?

#42

Post by pieterp » Sat, Mar 18, 2017, 1:40 PM UTC

The Second Machine Age

Re: Ask HN: Best books on AI?

#43

Post by Scarblac » Sat, Mar 18, 2017, 1:52 PM UTC

kleiba wrote: post #22

curuinor wrote: post #15

Earlier quoted context omitted.

Unfortunately, there is not a drop of numerics in that book. It's a good book for learning about symbolic AI. There is, to a solid first order approximation, zero symbolic AI in a system like, say, Google speech recognition.

Not surprisingly, I would say, since I wouldn't count speech recognition as an AI task.

Nothing is counted as an AI task after someone finally manages to do it.

Re: Ask HN: Best books on AI?

#44

Post by learningtool » Sat, Mar 18, 2017, 2:04 PM UTC

A good suggestion seems to be not to read a lot of books, but to put intro practice what you read and apply it to problems you need to solve, that way you learn a lot more effectively

Re: Ask HN: Best books on AI?

#45

Post by akssri » Sat, Mar 18, 2017, 2:05 PM UTC

Graphical models might also be something folks might want to consider. Ideas from PGMs are often behind many advances in ML.

The canonical text is by Daphne Koller; a course I took used Martin Wainwright's monograph though - the book is briefer and dives into the math quicker.

https://people.eecs.berkeley.edu/~wainwrig/Papers/WaiJor08_F...

Re: Ask HN: Best books on AI?

#46

Post by mnault000 » Sat, Mar 18, 2017, 2:21 PM UTC

Try this site: http://www.allitebooks.com/?s=artificial And the books are free and downloadable :)

Re: Ask HN: Best books on AI?

#47

Post by taktoa » Sat, Mar 18, 2017, 2:24 PM UTC

JamilD wrote: post #39

krosaen wrote: post #30

Some on this thread have recommended Norvig's PAIP, but that's kind of an old school AI book in that it focuses on heuristic search and logic (implementing prolog in lisp at one point, very impressive stuff actually); but is lacking any coverage of statistical machine learning, which is the approach that underlies most of the cool stuff these days. It's still a great book, but I'd instead recommend a path that focuse…

You don't need to focus on just one side or the other. AlphaGo wasn't made by just hacking together a couple neural nets, it built on heuristic search, MCTS, and the "old-school" AI. I agree that basic statistics + Bishop's book is a great way to start getting into machine learning -- but AI is a much broader field than that.

"old-school" AI is just a pretentious name for algorithms/graph theory/combinatorics, plus a bit of (now very outdated) PLT mixed in.

"new-school" AI (machine learning) is just a more pretentious name for statistics/control theory/randomized algorithms.

Re: Ask HN: Best books on AI?

#48

Post by danielam » Sat, Mar 18, 2017, 2:25 PM UTC

brogrammer2 wrote: post #2

Though it is an academic book, Artificial Intelligence: A Modern Approach by Stuart Russell and Peter Norvig. The first chapter in the book provides a detailed analysis of how other disciplines contribute to the idea of AI - from Philosophy to Psychology, Biology to Computer Science. Makes for an interesting read, even for a non-tech reader.

I wouldn't exaggerate the "detailed analysis" here. For example, I found the philosophical parts quite weak and superficial.

Re: Ask HN: Best books on AI?

#49

Post by visarga » Sat, Mar 18, 2017, 2:31 PM UTC

Imagine you want to learn the Roman alphabet. Which book should you use? Any book. There are so many good books and courses that it's almost useless to select. Don't worry about selecting the initial book, just use any course or book in the beginning, and later when you will know exactly what fits your needs, you will be able to fine tune.

Re: Ask HN: Best books on AI?

#50

Post by danielmorozoff » Sat, Mar 18, 2017, 2:36 PM UTC

Once you get your feet wet, the first year PHD course book for good theory is Pattern Recognition and Machine Learning by Bishop.

But it could be a bit too theoretical - it provides a foundational mathematical framework and got me thinking about problems in a better way.