ICANN 2012: Artificial Neural Networks and Machine Learning – ICANN 2012 pp 482-490 | Cite as
A Dynamic Binding Mechanism for Retrieving and Unifying Complex Predicate-Logic Knowledge
Abstract
We show how to encode, retrieve and process complex structures equivalent to First-Order Logic (FOL) formulae, with Artificial Neural Networks (ANNs) designed for energy-minimization. The solution constitutes a binding mechanism that uses a neural Working Memory (WM) and a long-term synaptic memory (LTM) that can store both procedural and declarative FOL-like Knowledge-Base (KB). Complex structures stored in LTM are retrieved into the WM only upon need, where they are further processed. The power of our binding mechanism is demonstrated on unification problems: as neurons are dynamically allocated from a pool, most generally unified structures emerge at equilibrium. The network’s size is O(n·k), where n is the size of the retrieved FOL structures and k is the size of the KB. The mechanism is fault-tolerant, as no fatal failures occur when random units fail. The paradigm can be used in a variety of applications, such as language processing, reasoning and planning.
Keywords
Binding-Problem Logic Unification Neural-Symbolic InferencePreview
Unable to display preview. Download preview PDF.
References
- 1.Ballard, D.H.: Parallel logical inference and energy minimization. In: Proceedings of the AAAI National Conference on Artificial Intelligence, pp. 203–208 (1986)Google Scholar
- 2.Barrett, L., Feldman, J.A., Dermed, M.L.: A (somewhat) new solution to the binding problem. Neural Computation 20, 2361–2367 (2008)MATHCrossRefGoogle Scholar
- 3.Browne, A., Sun, R.: Connectionist variable binding. Springer, Heidelberg (2000)Google Scholar
- 4.d’Avila Garcez, A.S., Lamb, L.C., Gabbay, D.M.: Neuro-Symbolic Cognitive Reasoning, Cognitive Technologies. Springer (2008)Google Scholar
- 5.Feldman, J.: The Binding Problem(s) (2010), http://www.computational-logic.org/content/events/iccl-ss-2010/slides/feldman/papers/Binding8.pdf
- 6.Fodor, J.A., Phylyshyn, Z.W.: Connectionism and cognitive architecture: A critical analysis. In: Pinker, Mehler (eds.) Connectionism and Symbols, pp.3–71. MIT Press (1988)Google Scholar
- 7.Hölldobler, S.: A Structured Connectionist Unification Algorithm. In: Proceedings of the Eighth National Conference on Artificial Intelligence, vol. 1 (1990)Google Scholar
- 8.Hölldobler, S., Kurfess, F.: CHCL-A Connectionist Inference System. In: Fronhöfer, B., Wrightson, G. (eds.) Dagstuhl Seminar 1990. LNCS (LNAI), vol. 590, pp. 318–342. Springer, Heidelberg (1992)CrossRefGoogle Scholar
- 9.Jackendoff, R.: Foundations of Language: Brain, Meaning, Grammar, Evolution. Oxford University Press (2002)Google Scholar
- 10.Komendantskaya, E.: Unification neural networks: unification by error-correction learning. Logic Jnl IGPL (2010)Google Scholar
- 11.Paterson, M., Wegman, M.: Linear unification. J. Comput. Syst. Sci. 16(2), 158–167 (1978)MathSciNetMATHCrossRefGoogle Scholar
- 12.Pinkas, G.: Reasoning, non-monotonicity and learning in connectionist networks that capture propositional knowledge. Artificial Intelligence 77, 203–247 (1995)MathSciNetMATHCrossRefGoogle Scholar
- 13.Pinkas, G.: Symmetric neural networks and logic satisfiability. Neural Comp. 3(2) (1991)Google Scholar
- 14.Pinkas, G.: Constructing proofs symmetric networks. In: NIPS-4, pp. 217–224 (1992)Google Scholar
- 15.Plate, T.: Holographic Reduced Representations. IEEE Trans. Neural Network 6(3) (2003)Google Scholar
- 16.Shastri, L., Ajjanagadde, V.: From associations to systematic reasoning: A connectionist representation of rules, variables and bindings. BBS 16(3), 417–494 (1993)Google Scholar
- 17.Lima, P.M.V.: Resolution-Based Inference on Artificial Neural Networks. Ph.D. Thesis, Department of Computing. Imperial College London, UK (2000)Google Scholar
- 18.Stolke, A.: Unification as Constraint Satisfaction in Structured Connectionist Networks. Neural Computation 1(4), 558–566 (1989)Google Scholar
- 19.Van der Velde, F., Kamps, M.: Neural blaclboard architectures of combinatorial structures and cognition. Behav. Brain Sci. 29, 1–72 (2006)Google Scholar
- 20.Bader, S., Hitzler, S., Holdobler, S.: Connectionist model generation: A first-order approach. Neurocomputing 71, 2420–2432 (2008)CrossRefGoogle Scholar