What Can We Compute with Lateral Inhibition Circuits

In this paper we review the lateral Inhibition (LI) as a connectionist conceptual model invariant in front of level translation, moving up from physical to symbol and then to the knowledge level where LI is transformed into a "pattern of reasoning". Afterwards, we explore the possibilities of this pattern as a component of modelling fault-tolerant cooperative processes. LI can be considered as an adaptive mechanism that integrates excitation and inhibition in a dynamic balance between complementary information sources.