Title :
Dissolving variables in connectionist combinatory logic
Author :
Barnden, John ; Srinivas, K.
Abstract :
Variable binding arises because the production rules, etc. deployed in traditional artificial intelligence computations generally have variables as crucial components. The technique of recasting such computations in terms of special functions called combinators avoids the use of explicit variables and can therefore be a useful approach for connectionism. In the important case of the connectionist system that interprets rules rather than having them wired in, the system would merely process complex, but variable-free, combinator expressions in certain simple ways. Although a connectionist system of the type investigated still has to manipulate complex data structures, those structures and their manipulations are rendered more uniform, thereby dissolving an important part of the variable binding problem. A connectionist combinator reduction machine which mimics the graph reduction machine described by D.A. Turner (1979) has been implemented
Keywords :
cognitive systems; combinatorial mathematics; data structures; formal logic; neural nets; artificial intelligence computations; combinator reduction machine; complex data structures; connectionist combinatory logic; graph reduction machine; production rules; rule interpretation; variable binding;
Conference_Titel :
Neural Networks, 1990., 1990 IJCNN International Joint Conference on
Conference_Location :
San Diego, CA, USA
DOI :
10.1109/IJCNN.1990.137921