Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Ask HN: Best books on AI?

news.ycombinator.com

Post reply on HN
Print view View original

61–70 of 95 posts

Re: Ask HN: Best books on AI?

#61

Post by psb217 »

king_magic wrote: post #37
curuinor wrote: post #13

Earlier quoted context omitted.

Less symbolic failures. There was a huge and abiding torrent of neural net stuff that dealt with evolving topologies in late 90's. I see very little of it in any way shape or form in industry or academia today, because it's a lot of computation for basically no gain. They thought that layerwise pretraining of neural nets was the way to go in 2006, before they realized that initializations, normalization, and better a…

Is pretraining really all that much of a failure? I haven't really found an authoritative answer on whether or not pretraining is worth it these days. Hinton's 2012(?) Coursera course still focuses pretty deeply on generative/layer-by-layer pretraining with RBMs but I'm just not really sure if that's fallen by the wayside today. Or maybe it's still useful only in specific circumstances?

AlphaGo relied heavily on (supervised) pretraining, and that seemed fairly successful.

Re: Ask HN: Best books on AI?

#62

Post by abecedarius »

nextos wrote: post #35
arihant wrote: post #26

You cannot really go practical in AI without academic rigor. You can do recipes of what's already been done using a TensorFlow book, but that's how far one can go. If one is serious in getting in AI today, a great way to do is read the following books, in order: 1. AI: A Modern Approach by Stuart Russell and Peter Norvig. 2. Deep Learning by Ian Goodfellow and Yoshua Bengio. It is amazing how approachable both books…

I love Paradigms of Artificial Intelligence by Peter Norvig much more than AIMA, which I found excessively encyclopedic and shallow. While some will argue it is dated, I think it presents many timeless ideas that will get in vogue soon with little tweaks to their inference schemes. Same for The Art of Prolog.

PAIP is one of my favorite books ever, but taken as a book about the craft of programming, not about AI. AI has grown, and the broadness of AIMA matches the subject. (It does need another update, and I heard they're working on one.)

Re: Ask HN: Best books on AI?

#63

Post by nabla9 »

nextos wrote: post #35
arihant wrote: post #26

You cannot really go practical in AI without academic rigor. You can do recipes of what's already been done using a TensorFlow book, but that's how far one can go. If one is serious in getting in AI today, a great way to do is read the following books, in order: 1. AI: A Modern Approach by Stuart Russell and Peter Norvig. 2. Deep Learning by Ian Goodfellow and Yoshua Bengio. It is amazing how approachable both books…

I love Paradigms of Artificial Intelligence by Peter Norvig much more than AIMA, which I found excessively encyclopedic and shallow. While some will argue it is dated, I think it presents many timeless ideas that will get in vogue soon with little tweaks to their inference schemes. Same for The Art of Prolog.

It's Paradigms of Artificial Intelligence Programming (PAIP) http://norvig.com/paip.html

AIMA provides better introduction for wider area of subjects but PAIP is one of most elegant and timeless books for both programming and old school AI.

Re: Ask HN: Best books on AI?

#64

Post by king_magic »

curuinor wrote: post #51
king_magic wrote: post #37

Earlier quoted context omitted.

Is pretraining really all that much of a failure? I haven't really found an authoritative answer on whether or not pretraining is worth it these days. Hinton's 2012(?) Coursera course still focuses pretty deeply on generative/layer-by-layer pretraining with RBMs but I'm just not really sure if that's fallen by the wayside today. Or maybe it's still useful only in specific circumstances?

Saxe Ganguli McClelland, 2013, about linear nets and orthogonal initialization. But then, read Li Jiao Han Weissman 2017 (maybe preprint), "Demystifying ResNet", which makes a nice claim about the niceness being conditioning of Hessian at init. Tldr: it's good conditioner but you can do better ab initio

Interesting, I will check these out!

Re: Ask HN: Best books on AI?

#66

Post by EternalData »

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…

+1 to Python Machine Learning, I feel like some of artificial intelligence is understanding a lot of finer (but critical) mathematical nuances, but some of it is just getting your hands dirty, and the second is definitely more accessible for somebody who is starting to learn.

Re: Ask HN: Best books on AI?

#67

Post by sehugg »

You know, I have never found a "casual" book covering tree-search techniques, from minimax to Monte Carlo Tree Search. Still relevant for game/agent AI (AlphaGo used MCTS for example).

You'd think there would have been 100 "How To Make a Computer Chess Engine in BASIC" books back in the 80s, and continuing to the present day, but I can't find them. Lots of papers and online tutorials, and some stuff in textbooks, but no accessible hands-on books.

Re: Ask HN: Best books on AI?

#68

Post by goatlover »

nojvek wrote: post #57
taktoa wrote: post #47

Earlier quoted context omitted.

"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.

What is to say the human brain doesn't engage in graph combinotrics with some learned heuristics. Some of the old ideas with enough computation power are actually pretty amazing.

Because the human brain isn't a symbolic machine. It's a living organism made up of cells trying to survive and reproduce in the real world. All our symbol manipulation capabilities are built up on older abilities based on feeling, experiencing, socializing, and moving about.

Re: Ask HN: Best books on AI?

#69

Post by j_s »

First, read fucking Hastie, Tibshirani, and whoever. Chapters 1-4 and 7. If you don't understand it, keep reading it until you do.

You can read the rest of the book if you want. You probably should, but I'll assume you know all of it.

Take Andrew Ng's Coursera. Do all the exercises in Matlab and python and R. Make sure you get the same answers with all of them.

Now forget all of that and read the deep learning book. Put tensorflow or torch on a Linux box and run examples until you get it. Do stuff with CNNs and RNNs and just feed forward NNs.

Once you do all of that, go on arXiv and read the most recent useful papers. The literature changes every few months, so keep up. There. Now you can probably be hired most places. If you need resume filler, so some Kaggle competitions. If you have debugging questions, use StackOverflow. If you have math questions, read more. If you have life questions, I have no idea.

source: fizixer https://news.ycombinator.com/item?id=13890952

FWIW, a "super harsh" guide to (learning) ML [1] was posted on reddit a few days ago.

[1] https://redd.it/5z8110

Edit: The entire Reddit discussion feels slightly similar to this one, if more snarky. The first reply there also links all the resources listed above. I don't really know enough to add anything.

Re: Ask HN: Best books on AI?

#70

Post by radicality »

j_s wrote: post #69

First, read fucking Hastie, Tibshirani, and whoever. Chapters 1-4 and 7. If you don't understand it, keep reading it until you do. You can read the rest of the book if you want. You probably should, but I'll assume you know all of it. Take Andrew Ng's Coursera. Do all the exercises in Matlab and python and R. Make sure you get the same answers with all of them. Now forget all of that and read the deep learning book.…

Statistical learning is only one part of AI. If the person wants to learn about AI, wouldn't starting with something like "AI: A modern approach" he a better start?
Post reply on HN