Pattern Recognition in a Bucket
This paper demonstrates that the waves produced on the surface of water can be used as the medium for a “Liquid State Machine” that pre-processes inputs so allowing a simple perceptron to solve the XOR problem and undertake speech recognition. Interference between waves allows non-linear parallel computation upon simultaneous sensory inputs. Temporal patterns of stimulation are converted to spatial patterns of water waves upon which a linear discrimination can be made. Whereas Wolfgang Maass’ Liquid State Machine requires fine tuning of the spiking neural network parameters, water has inherent self-organising properties such as strong local interactions, time-dependent spread of activation to distant areas, inherent stability to a wide variety of inputs, and high complexity. Water achieves this “for free”, and does so without the time-consuming computation required by realistic neural models. An analogy is made between water molecules and neurons in a recurrent neural network.
KeywordsSpeech Recognition Recurrent Neural Network Separation Property Sound Sample Neural Network Parameter
Unable to display preview. Download preview PDF.
- 3.Auer, P., Burgsteiner, H., et al.: The p-Delta Learning Rule for Parallel Perceptrons (2002) (submitted for publication), http://www.cis.tugraz.at/igi/pauer/publications.html
- 7.MacLennan, B.: Field Computation in Natural and Artificial Intelligence. Knoxville, University of Tennessee. Report UT-CS-99-422. (1999), http://www.cs.utk.edu/mclennan/fieldcomp.html
- 10.Walmsley, I.: Computing with interference: All-optical single-query 50- element database search. In: Conference on Lasers and Electro-Optics/Quantum Electronics and Laser Science, Baltimore, Maryland (2001)Google Scholar
- 12.Muller, G., Newman, S.: Origination of Organismal Form: Beyond the Gene in Developmental and Evolutionary Biology. MIT Press, Cambridge (2003)Google Scholar