Fetching the paper…
Reading the bibliography…
We introduce a notion of "effective dimension" of a statistical model based on the number of cubes of size $1/\sqrt{n}$ needed to cover the model space when endowed with the Fisher Information Matrix as metric, $n$ being the number of observations.
Rissanen J. Modeling By Shortest Data Description
1978
Earlier work this paper cites.
Mandelbrot B. The fractal geometry of nature
1982
Earlier work this paper cites.
Weigend A S, Rumelhart D E. The effective dimension of the space of hidden units
1991
Earlier work this paper cites.
Opper M. Learning and generalization in a two-layer neural network: The role of the Vapnik-Chervonvenkis dimension
1994
Earlier work this paper cites.
Rissanen J. Fisher information and stochastic complexity
1996
Earlier work this paper cites.
Takeuchi J, Barron A. Asymptotically minimax regret for exponential families
1997
Cited alongside, same era.
Takeuchi J, Barron A. Asymptotically minimax regret by Bayes mixtures
1998
Cited alongside, same era.
Vapnik V. Statistical learning theory
1998
Cited alongside, same era.
Takeuchi J. On minimax regret with respect to families of stationary stochastic processes
2000
Cited alongside, same era.
Bialek W, Nemenman I, Tishby N. Predictability, complexity, and learning
2001
Cited alongside, same era.
Geiger D, Heckerman D, Meek C. Asymptotic Model Selection for Directed Networks with Hidden Variables
Cited in the paper.
Zhang N L, Kocka T. Effective dimensions of Hierarchical Latent Class Models
2004
Later among the works it cites.
Grünwald P D. The Minimum Description Length Principle
2007
Later among the works it cites.
2017
Later among the works it cites.
Ravichandran K, Jain A, Rakhlin A. Using effective dimension to analyze feature transformations in deep neural networks
2019
Later among the works it cites.
alphaXiv searches the wider corpus for related work and actual follow-ups.
alphaXiv is searching for related work…