Technical Report Number
Connectionist networks with symmetric weights (like Hopfield networks and Boltman Machines) use gradient descent to find a minimum for quadratic energy functions. We show an equivalence between the problem of satisfiability in propositional calculus and the problem of minimizing those energy functions. The equivalence is in the sense that for an satisfiable Well Formed Formula (WFF) we can find a quadratic function that describes it, such that the set of solutions that minimize the function is equal to the set of truth assignments that satisfy the WFF. We also show that in the same sense every quadratic energy function describes some satisfiable WFF. Algorithms are given to transform any propositional WFF into an energy function that describes it and vice versa. High-order models that use Sigma-Pi units are shown to be equivalent to the standard quadratic models with additional hidden units. An algorithm to convert high-order networks to low-order ones is used to implement a satisfiability problem-solver on a connectionist network. The results give better understanding of the role of hidden units and of the limitations and capabilities of symmetric connectionist models. The techniques developed for the satisfiability problem may be applied to a wide range of other problems, such as: associative memories, finding maximal consistent subsets, automatic deduction and even non-monotonic reasoning.
Pinkas, Gadi, "Symmetric Neural Nets and Propositional Logic Satisfiability" Report Number: WUCS-91-02 (1991). All Computer Science and Engineering Research.