Download:
by Mostefa Golea, Thomas R. Hancock, Mario Marchand
http://www.csi.uottawa.ca/~marchand/publications/nips92.pdf
Add To MetaCart
Abstract:
Neural networks with binary weights are very important from both the theoretical and practical points of view. In this paper, we investigate the learnability of single binary perceptrons and unions of µ-binary-perceptron networks, i.e. an “OR ” of binary perceptrons where each input unit is connected to one and only one perceptron. We give a polynomial time algorithm that PAC learns these networks under the uniform distribution. The algorithm is able to identify both the network connectivity and the weight values necessary to represent the target function. These results suggest that, under reasonable distributions, µ-perceptron networks may be easier to learn than fully connected networks.
Citations
|
1744
|
Eigenfaces for Recognition
– Turk, Pentland
- 1991
|
|
1364
|
A theory of the learnable
– Valiant
- 1984
|
|
159
|
Face recognition: the problem of compensating for changes in illumination direction
– Adini, Moses, et al.
- 1997
|
|
144
|
Learning Boolean formulae
– Kearns, Li, et al.
- 1995
|
|
40
|
Complexity Results on Learning by Neural Nets
– Lin, Vitter
- 1991
|
|
28
|
Parameterised images for recognition and reconstruction
– Craw, Cameron
- 1991
|
|
18
|
A comparison of two computer-based face identification systems with human perceptions of faces. Vision Research
– Hancock, Bruce, et al.
- 1998
|
|
16
|
Learning nonoverlapping perceptron networks from examples and membership queries
– Hancock, Golea, et al.
- 1994
|
|
10
|
Neural and psychophysical analysis of object and face recognition
– Biederman, Kalocsai
- 1998
|
|
9
|
Some approximations to the binomial distribution function
– Bahadur
- 1960
|
|
8
|
Recognition memory for photographs of faces
– Galper
- 1971
|
|
6
|
A Guided Tour to
– Hagerup
- 1989
|
|
6
|
On Learning Binary Weights for Majority Functions
– Venkatesh
- 1991
|
|
5
|
Optimal unsupervised learning in a single-layer linear feedforward neural network. Neural Networks 2:459–473
– TD
- 1989
|
|
5
|
Margaret Thatcher: a new illusion. Perception
– Thompson
- 1980
|
|
4
|
Storage Capacity of a Multilayer Neural Network with Binary weights”, Europhys
– Barkai
- 1991
|
|
4
|
A greedy method for learning µDNF functions under the uniform distribution
– Pagallo
- 1989
|
|
4
|
Face processing: human perception and principal components analysis", submitted to Memory and Cognition
– PJB, AM, et al.
- 1996
|
|
3
|
The visual representation of texture
– SC
- 1994
|
|
3
|
von der Malsburg C, Würtz RP, Konen W. Distortion invariant object recognition in the dynamic link architecture
– Lades, JC, et al.
- 1993
|
|
2
|
A statistical analysis of natural images matches psychophysically derived orientation tuning curves
– RJ, PJB
- 1991
|
|
2
|
Independent component analysis of face images: a representation for face recognition
– MS, TJ
- 1997
|
|
2
|
Healy P and Burton AM, The importance of 'mass
– Bruce, Hanna, et al.
- 1992
|
|
2
|
A computational examination of image segmentation and the initial stages of human vision, Perception
– RJ
- 1994
|
|
1
|
On Learning Simple Neural Concepts
– Marchand
- 1992
|
|
1
|
Familiarity, memorability and the e ect of typicality on the recognition of faces. Memory and Cognition
– JR, JD
- 1992
|
