Immune systems are adaptive systems in which learning takes place by evolutionary mechanisms similar to biological evolution. Their major function is to provide a defense mechanism for the body that can identify dangerous foreign material and eliminate it. Understanding the immune system is important, both because of its role in complex diseases such as AIDS and because of potential applications to computational problems. The mechanisms of the immune system are remarkably complex and poorly understood, even by immunologists.

Our immune system models are based on a universe in which antigens (foreign material) and lymphocytes (the cells that do the recognition of foreign material) are represented by strings of discrete symbols. The strings represent receptors on B cells and T cells and epitopes on antigens—the regions of the cells in which binding takes place. The complex chemistry of molecular binding is modeled in our system by string matching.

We have used this model to study several different aspects of the immune system, including its ability to detect common patterns (schemas) in noisy environments (Forrest et al., 1993a), its ability to discover and maintain coverage of diverse pattern classes (Smith et al., 1993), and its ability to learn effectively, even when not all antibodies are expressed and not all antigens are presented (Hightower et al., 1995, Hightower et al., 1996). Current research directions include modeling cross-reactive memory (work by Derek Smith) and location-specific behavior (immune cells that behave differently depending on their physical location in the body). This work is in collaboration with Alan Perelson.