| Hans W. Gusgen and Ste#en Holldobler. Connectionist inference systems. In Bertram Fronhofer and Graham Wrightson, editors, Parallelization in Inference Systems, volume 590 of Lecture Notes in Artificial Intelligence, pages 82--120. Springer, Berlin, 1992. |
....by introducing a second subsymbolic level where symbols are represented by features. Thus e.g. similarity matching becomes possible. Similarity matching is a way to cope with incomplete knowledge. It enables rules that do not match exactly to be used if directly applicable rules are missing. In [GH92] a number of different ideas and models are presented how to do rule based reasoning with neural networks, whereas in [HK92] one sophisticated model is presented which can do inferences for a subset of FOL. This work gives interesting insight in the computational power of connectionist networks, ....
H. W. Gusgen and S. Holldobler. Connectionist Inference Systems. In B. Fronhofer and G. Wrightson, editors, Parallelization in Inference Systems, pages 82--120, Dagstuhl, Germany, 1992. LNAI 590, Springer-Verlag.
....as reasoning has been suffering from a certain inadequacy in the representations used. The ease and elegance with which symbolic manipulation techniques are applied to dynamically changing data structures in conventional processing paradigms has not yet been matched with connectionist means (see [Gusgen and Holldobler, 1991] in this book) One particular problem with respect to logic is the treatment of variable bindings; although some proposals to solve this problem have been made (see e.g. Ballard, 1986; Touretzky and Hinton, 1988; Shastri and Ajjanagadde, 1990b] their integration into a logical framework is not ....
H.-W. Gusgen and S. Holldobler. Connectionist inference systems. In B. Fronhofer and G. Wrightson, editors, Parallelization in Inference Systems. Springer, 1991. (to appear).
No context found.
Hans W. Gusgen and Ste#en Holldobler. Connectionist inference systems. In Bertram Fronhofer and Graham Wrightson, editors, Parallelization in Inference Systems, volume 590 of Lecture Notes in Artificial Intelligence, pages 82--120. Springer, Berlin, 1992.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC