The performance of universal encoding

By R. Krichevsky, V. K. Trofimov. 1981

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

Distributed Mean Estimation with Limited Communication (2016, methodology): “Instead, we propose to further encode the transmitted values using universal compression schemes [17, 12].”
Learning Discrete Distributions from Untrusted Batches (2017, background): “The problem of learning a discrete distribution given access to independent samples has been intensely studied over the past century (see, e.g. [Gil71, KT81] and the references therein).”
The Forget-me-not Process (2016, methodology): “The following base models were chosen for each test domain: for the Mysterious Bag of Coins (MBOC), we used the KT-estimator [13], a beta-binomial model; for A Fistful of Digits (FOD), we used MADE [9], a recently introduced, general purpose neural density estimator, with 500 hidden units, trained o”
Skip Context Tree Switching (2014, methodology): “In Section 4 we describe a large experiment whose performance is much improved by the use of a SAD-like estimator.”
On Prediction Using Variable Order Markov Models (2004, methodology): “(1997), the above simple form of the KT estimator is equivalent to the original form, given in terms of a Dirichlet distribution by Krichevsky and Trofimov (1981). The (original) KT-”
Entropy Inference and the James-Stein Estimator, with Application to Nonlinear Gene Association Networks (2008, background): “0 no prior maximum likelihood 1/2 Jeffreys prior (Jeffreys, 1946) Krichevsky and Trofimov (1981) 1 Bayes-Laplace uniform prior Holste et al. (1998) 1/p Perks prior (Perks, 1947) Schurmann and Grassberger (1996) p n/p minimax prior (Trybula, 1958)”
Online Learning of Non-stationary Sequences (2003, methodology): “1 Optimized regret bound for Learn-α The optimized regret bound for Learn-α(δ) is thus (approximately) 12 log T +c, which is comparable to analysis of universal coding for word-length T [11].”
Competitive Distribution Estimation: Why is Good-Turing Good (2015, background): “Following [8], researchers have studied rn(∆k) and related quantities [9].”
Efficient Tracking of Large Classes of Experts (2011, unknown): “(10) This, together with (9), implies (8).”
Universal Estimation of Information Measures for Analog Sources (2009, background): “In addition, for memoryless sources, different approaches [139, 177, 184, 265] have been designed to overcome difficulties in the under-sampled regime.”
Prediction, learning, and games (2006, unknown): “Trofimov [186].”
On Learning Distributions from their Samples (2015, background): “As shown by Krichevsky (1998), no add-constant estimator can achieve a natural lower bound of rKL2,n ≥ 12n + o ( 1 n ) .”