Title :
Learning exponential state-growth languages by hill climbing
Author_Institution :
Dept. of Psychol., Connecticut Univ., Storrs, CT, USA
fDate :
3/1/2003 12:00:00 AM
Abstract :
Training recurrent neural networks on infinite state languages has been successful with languages in which the minimal number of machine states grows linearly with the length of the sentence, but has faired poorly with exponential state-growth languages. The new architecture learns several exponential state-growth languages in near perfect by hill climbing.
Keywords :
fractals; learning (artificial intelligence); probability; recurrent neural nets; exponential state-growth languages; fractal learning; hill climbing; infinite state languages; probability distribution; recurrent neural networks; recursive computations; Computer architecture; Computer networks; Counting circuits; Fractals; Learning automata; Neural networks; Probability distribution; Recurrent neural networks; Turing machines; Vocabulary;
Journal_Title :
Neural Networks, IEEE Transactions on
DOI :
10.1109/TNN.2003.809421
