A Probability Path

By Sidney I. Resnick. 1999

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

P3O: Policy-on Policy-off Policy Optimization (2019, methodology): “is the Radon-Nikodym derivative of the two densities (Resnick, 2013) and it re-balances the data to put more weight on unlikely samples in q(x) that are likely under the test data p(x).”
Is network traffic approximated by stable Levy motion or fractional Brownian motion (2002, methodology): “We check Lyapunov’s condition ([14], page 286, [40], page 319) for asymptotic normality of the sums”
Support points (2016, background): “2 of [57]: one can show that any two convergent subsequences of {Fn} must converge to the same limit, thereby proving the convergence of Fn to F .”
Limit Laws for Random Vectors with an Extreme Component (2005, unknown): “, [35], page 253) [∫ [u≤x] f(X−β(t))/α(t)|Y/t=v(u|v) du ] → ∫ [u≤x] v2g(u, v) du.”
Mapping complex tissue architecture with diffusion spectrum magnetic resonance imaging (2005, unknown): “[21] the moments can be expressed in terms of DTs of corresponding order, thus bringing up the generalized DT formalism developed by Liu et al.”
Extended Conditional Independence and Applications in Causal Inference (2015, unknown): “Z whenever AY,Z ∈ σ(Y,Z). (A.1) To show (A.1), consider DA X = AY,Z ∈ σ(Y,Z) : Eσ 1AX AY,Z = Eσ{wφ(Z) A Y,Z } and Π = {AY,Z ∈ σ(Y,Z) : AY,Z = AX ∩AY for some AY ∈ σ(Y),AZ ∈ σ(Z)}. Then σ(Π) = σ(Y,Z) [Resnick, 2014, p. 73]. We will show that DA X is a d-system that contains Π; it will then follow fro”
DDPG++: Striving for Simplicity in Continuous-control Off-Policy Reinforcement Learning (2020, background): “which is the Radon-Nikodym derivative of the two densities [14] and it re-balances the data to put more weight on unlikely samples in q(x) that are likely under the test data p(x).”
Detecting a conditional extreme value model (2009, unknown): “For N finite we know that N∫ 1 N∫ 1 µ∗n([x,∞]× (y,∞]) dx x dy y P→ N∫ 1 N∫ 1 µ∗([−∞, H←(1/x)]× (y,∞])dx x dy y ,(2.17) since (2.15) implies that the integrand converges in probability and we can use Pratt’s Lemma (Resnick, 1999, page 164) for the convergence of the integral.”
Predicting Binary Outcomes (2013, unknown): “6(c), p. 156, Resnick 1999).”
Meta-Q-Learning (2019, background): “The importance ratio β ( x ) = d p d q ( x ) , also known as the propensity score, is the Radon-Nikodym derivative (Resnick, 2013) of the two data densities and measures the odds of a sample x coming from the distribution p versus the distribution q .”