Luis C. Lamb, Rafael V. Borges, Artur S. d'Avila Garcez
The importance of the efforts towards integrating the symbolic and connectionist paradigms of artificial intelligence has been widely recognised. Integration may lead to more effective and richer cognitive computational models, and to a better understanding of the processes of artificial intelligence across the field. This paper presents a new model for the representation, computation, and learning of temporal logic in connectionist systems. The model allows for the encoding of past and future temporal logic operators in neural networks, through a neural-symbolic translation algorithms introduced in the paper. The networks are relatively simple and can be used for reasoning about time and for learning by examples with the use of standard neural learning algorithms. We validate the model in a well-known application dealing with temporal synchronisation in distributed knowledge systems. This opens several interesting research paths in cognitive modelling, with potential applications in agent technology, learning and reasoning.
Subjects: 4. Cognitive Modeling; 14. Neural Networks
Submitted: Apr 24, 2007