|
CNS 185 is an interdisciplinary course combining material from several areas: mathematics, computer science, engineering, and the natural sciences of biology and physics. Listed below are topics from these areas that you should already be familiar with. However, we don't want to discourage anybody from taking the course: if you see something on the list that you only half-remember from a class long ago, don't worry, chances are that in doing 185 your memory will be refreshed as necessary. (Ideally, of course, you will take the trouble to refresh your memory on these topics before the next lecture.) Also listed are a number of topics on which a number of handouts will be provided to help with learning them. To get a background reference list that gives pointers to books containing detailed descriptions of more complicated material, click here. Note that this reference list is not required reading or knowledge for this course; on the contrary, it is a guide to some of the information you may need beyond CNS 185. The lecture schedule describes the material that will be covered, and you may note that there are no prerequisites or references in artificial or biological neural networks since these are the main topics of the course. If, after reading this, you are still unsure about whether this class is appropriate for you, do not hesitate to contact one of the TAs.
MathThe most important concepts you will need are basic linear algebra (matrices, vectors, eigenvalues and eigenvectors) and simple linear differential equations. Familiarity with basic math, as in basic differentiation and integration, and the basic concepts of probability, will be assumed. Previous contact with information theory would prove useful. There was a supplementary math class concurrent with CNS 185; it started out with linear algebra and covered many useful topics. However there was not sufficient interest from students or CNS faculty to continue it. A new class covering a similar range of topics might be taught in the fall of 1999, look for more details at the beginning of the semester. For now, you can visit the web page of the old math class for some background information. ComputingYou should have some experience with a high-level programming language (Pascal, Basic, C, ideally MATLAB). The homeworks include a fair amount of programming -- if you have never programmed before, this is probably not a class for you (by the way, CCO runs some very good programming classes). MATLAB code will be provided for several problem sets, so it is recommended that students use MATLAB to complete their assignments. Nevertheless, students are free to complete the assingmants in whatever programming language they choose. If you have used any other high-level programming language before, MATLAB should not take too long to pick up. If there are enough requests, the TAs will provide a brief tutorial on MATLAB during the first week of the course. Otherwise, if you are having trouble getting used to MATLAB programming, download the MATLAB Primer and come to TA hours with specific questions.
Physics/EngineeringYou should know Newton's laws and the concept of energy. You should also know some simple passive circuit analysis, as in Ohm's law and Kirchoff's current and voltage laws. If you don't know these last, ask an engineering friend -- it will take you 5 minutes to learn them. |