On the Learnability of Recursive Data

On the Learnability of Recursive Data (1999)

Cached

Download Links

[www.informatik.uni-osnabrueck.de]

by Barbara Hammer

Venue:	Mathematics of Control, Signals, and Systems
Citations:	10 - 8 self

BibTeX

@ARTICLE{Hammer99onthe,
    author = {Barbara Hammer},
    title = {On the Learnability of Recursive Data},
    journal = {Mathematics of Control, Signals, and Systems},
    year = {1999},
    volume = {12},
    pages = {12--62}
}

Bookmark

OpenURL

Abstract

We establish some general results concerning PAC learning: We find a characterization of the property that any consistent algorithm is PAC. It is shown that the shrinking width property is equivalent to PUAC learnability. By counterexample PAC and PUAC learning are shown to be different concepts. We find conditions ensuring that any nearly consistent algorithm is PAC or PUAC, respectively.

Citations

383	Identification and control of dynamic systems using neural networks - Narendra, Parthasarathy - 1990
214	Learning long-term dependencies with gradient descent is difficult - Bengio, Simard, et al. - 1994
186	Training a 3-node neural networks is np-complete - Blum, Rivest - 1992
175	Scale-sensitive dimensions, uniform convergence, and learnability - Alon, Ben-David, et al. - 1997
139	On the computational power of neural nets - Siegelmann, ED - 1995
103	Neural net architectures for temporal sequence processing. In: A.S. Weigend and N.A. Gershenfeld (Eds.), Time Series Prediction: Forecasting the future and understanding the past - Mozer - 1994
88	A Theory of Learning and Generalization - Vidyasagar - 1997
31	The dynamic universality of sigmoidal neural networks - Kilian, Siegelmann - 1996
29	Neural nets with superlinear VC-dimension - Maass - 1994
26	Backpropagation Can Give Rise to Spurious Local Minima Even for Networks Without - Sontag, Sussman - 1989
23	On the complexity of training neural networks with continuous activation - DasGupta, Siegelmann, et al. - 1995
23	Vapnik-Chervonenkis dimension of recurrent neural networks - Koiran, Sontag - 1998
22	ample complexity for learning recurrent perceptron mappings - Dasgupta, Sontag - 1996
21	Stable encoding of large finite-state automata in recurrent neural networks with sigmoid discriminants - Omlin, Giles - 1996
17	Some experiments on the applicability of folding architectures to guide theorem proving - Schulz, Küchler, et al. - 1997
12	Neural networks can approximate mappings on structured objects - Hammer, Sperschneider - 1997
10	Recurrent neural networks: Some systems-theoretic aspects,” in Dealing with Complexity: a Neural Network Approach - Sontag - 1997
9	Protein secondary structure prediction with partially recurrent neural networks - Reczko - 1993
5	On the generalization of Elman networks - Hammer - 1997
5	Training a sigmoidal network is difficult - Hammer - 1998
1	Neural networks for control, Essays on Control - Sontag - 1993