The twenty last years have been marked by an increase inavailable data and computing power. In parallel to this trend, thefocus of neural network research and the practice of trainingneural networks has undergone a number of important changes, forexample, use of deep learning machines. The second edition of thebook augments the first edition with more tricks, which haveresulted from 14 years of theory and experimentation by some of theworld's most prominent neural network researchers. These tricks canmake a substantial difference (in terms of speed, ease ofimplementation, and accuracy) when it comes to putting algorithmsto work on real problems. |