G-maximization: An unsupervised learning procedure for discovering regularities

By Barak A. Pearlmutter, Geoffrey E. Hinton. 1987

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, background): “Many papers of the previous millennium, however, were about unsupervised learning (UL) without a teacher (e.g., Hebb, 1949; von der Malsburg, 1973; Kohonen, 1972, 1982, 1988; Willshaw and von der Malsburg, 1976; Grossberg, 1976a,b; Watanabe, 1985; Rumelhart and Zipser, 1986; Pearlmutter and Hinton, ”
Learning Factorial Codes by Predictability Minimization (1992, background): “'The G-Max algorithm (Pearlmutter and Hinton 1986) aims at a related objective: It tries to discover features that account for input redundancy.”
Reducing Overfitting in Deep Networks by Decorrelating Representations (2015, background): “It also appears in (Pearlmutter & Hinton, 1986) where it complements an objective which encourages units to capture higher order input statistics.”
Could information theory provide an ecological theory of sensory processing? (2011, background): “Some unsupervised learning algorithms that achieve this in some simple settings have appeared in Goodall (1960), Hinton and Sejnowski (1983), Pearlmutter and Hinton (1986), Barlow and Foldiak (1989), Redlich (1991) and Atick and Redlich (1992b).”
The Utility Driven Dynamic Error Propagation Network (background): “It is assumed that similar architectures are implementable for other connectionist learning algorithms such as G-Maximisation (Pearlmutter and Hinton, 1986), Boltzmann machines (Ackley, Hinton and Sejnowski, 1985) and Hopfield nets (Hopfield, 1982).”
Connectionist Learning Procedures (1989, background): “The derivative of this mutual information is relatively easy to compute and so it can be maximized by gradient ascent (Pearlmutter and Hinton, 1986).”