Lens is a neural network simulator designed for speed and ease of customization. On large networks, Lens is several times faster than most commonly used simulators. Although intended primarily for backpropagation networks, it also currently supports deterministic Boltzmann machines and Kohonen networks and could easily be extended to other Hebbian or Bayesian models. Lens is written entirely in the C and Tcl languages and operates on both Unix and Windows platforms. This report gives a brief overview of Lens and describes some of the interesting aspects of its design.

10 pages

Return to: SCS Technical Report Collection
School of Computer Science homepage

This page maintained by reports@cs.cmu.edu