Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!sdd.hp.com!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!pacbell.com!ucsd!sdcc6!beowulf!schraudo From: schraudo@beowulf.ucsd.edu (Nici Schraudolph) Newsgroups: comp.ai.neural-nets Subject: Re: Introductory books... Keywords: books Message-ID: Date: 28 Jan 91 21:32:26 GMT References: <1991Jan22.173911.779@scrumpy@.bnr.ca> <6929@uqcspe.cs.uq.oz.au> Sender: news@sdcc6.ucsd.edu Lines: 37 Nntp-Posting-Host: beowulf.ucsd.edu Don't forget "Introduction to the Theory of Neural Computation" by Hertz, Krogh & Palmer (Addison-Wesley 1991). Though I haven't read my copy yet (it only just came out), my first impression is that this one is by far the best book on the topic. At NIPS, where it was first introduced, the general reaction was "looks like there's finally a good textbook on neural nets". And it's below $30 for the paperback. >Wasserman seemed to be the most popular choice. This statement made me cringe - am I the only one who considers Wasserman a, pardon me, next to useless book? I could go on and on about the reasons for my dislike, but IMHO this book is not even worth wasting bandwidth for that. Suffice it to say that Wasserman flatly enumerates some common architectures from an engineer's perspective ("how it works") without ever addressing the underlying fundamentals ("why it works") - important basic concepts such as clustering, principal components or gradient descent are not treated. It's also full of errors, and unhelpful diagrams drawn with what appears to be PCB board layout software from the '70s. Wasserman may be acceptable for the applied engineer who quickly wants to try a neural net at his problem, and feels he doesn't need to *understand* the algorithms he is using. This approach is of course easier on the reader, which may partly explain the apparent popularity of this book. For anyone who wants to do active research in the field I consider it quite inadequate. Disclaimer: sprinkle the above with IMHOs as needed. PS: I will send my copy of Wasserman to the first victim, er, person who is willing to pay for the postage. No-income people have priority. -- Nicol N. Schraudolph, C-014 | "And long cars in long lines University of California, San Diego | And great big signs, and they all say: La Jolla, CA 92093-0114 | Hallelujah. Every man for himself." nici%cs@ucsd.{edu,bitnet,uucp} | - Laurie Anderson, "Big Science".