Competition and Multiple Cause Models

By Peter S. Dayan, Richard S. Zemel. 1995

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Deep learning in neural networks: An overview (2014, neutral): “Many UL methods are designed to maximize information-theoretic objectives (e.g., Linsker, 1988; Barlow et al., 1989; MacKay and Miller, 1990; Plumbley, 1991; Schmidhuber, 1992b,c; Schraudolph and Sejnowski, 1993; Redlich, 1993; Zemel, 1993; Zemel and Hinton, 1994; Field, 1994; Hinton et al., 1995; D”
Generalization in Interactive Networks: The Benefits of Inhibitory Competition and Hebbian Learning (2001, methodology): “Speciﬁcally, the input vocabulary consists of all 45 combinations of 5 horizontal and 5 vertical bars in a 5x5 grid, and the output mapping is a localist identiﬁcation of the two input bars (this is similar to the bars tasks used by F¨oldi´ak, 1990; Saund, 1995; Zemel, 1993; Dayan & Zemel, 1995).”
Feature Extraction Through LOCOCODE (1999, methodology): “This task is adapted from Dayan and Zemel (1995) (see also Földiák, 1990; Zemel, 1993; Saund, 1995) but is more difficult (compare Baumgartner, 1996).”
Learning with Mixtures of Trees (2000, methodology): “Typical values for l in the literature are l = 4, 5 ; we choose l = 5 following Dayan and Zemel (1995). The other parameter values were pb = 0.”
Varieties of Helmholtz Machine (1996, methodology): “unit activation functions The original HM involved networks of binary stochastic sigmoid, noisy-or (Pearl, 1988; Saund, 1995) or competitive ( Dayan & Zemel, 1995 ) units.”
Variational Learning in Nonlinear Gaussian Belief Networks (1999, background): “The goal of the much simpler “bars problem” (Dayan and Zemel 1995) is to learn without supervision to detect bars of two orthogonal orientations and to model the constraint that each image consists of bars of the same orientation.”