We justify the need for a connectionist implementation of compositional rul
e of inference (COI) and propose a network architecture for the same. We ca
ll it COIN-the compositional rule of inferencing. Given a relational repres
entation of a set of rules, the proposed architecture can realize the COI.
The outcome of COI depends on the choice of the implication function and al
so on choice of inferencing scheme. The problem of choosing an appropriate
implication function is avoided through neural learning. The system automat
ically finds an "optimal" relation to represent a set of fuzzy rules. We su
ggest a suitable modeling of connection weights so as to ensure learned wei
ghts lie in [0, 1]. We demonstrate through numerical examples that the prop
osed neural realization can find a much better representation of the rules
than that by usual implication and hence results in much better conclusions
than the usual COI. Numerical examples exhibit that COIN outperforms not o
nly usual COI but also Some of the previous neural implementations of fuzzy
logic. (C) 1999 John Wiley & Sons, Inc.