An Information Divergence Estimation over Data Streams

Emmanuelle Anceaume (1, 2) , Yann Busnel (3)

1 CIDER
2 CIDRE - Confidentialité, Intégrité, Disponibilité et Répartition
3 GDD - Gestion de Données Distribuées [Nantes]

In this paper, we consider the setting of large scale distributed systems, in which each node needs to quickly process a huge amount of data received in the form of a stream that may have been tampered with by an adversary. In this situation, a fundamental problem is how to detect and quantify the amount of work performed by the adversary. To address this issue, we have proposed in a prior work, AnKLe, a one pass algorithm for estimating the KL divergence of an observed stream compared to the expected one. Experimental evaluations have shown that the estimation provided by AnKLe is accurate for different adversarial settings for which the quality of other methods dramatically decreases. In the present paper, considering n as the number of distinct data items in a stream, we show that AnKLe is an (ε,δ)-approximation algorithm with a space complexity Õ(1/ε + 1/ε^2) bits in "most" cases, and Õ(1/ε + (n−ε−1)/ε^2) otherwise. To the best of our knowledge, an approximation algorithm for estimating the KL divergence has never been analyzed before.

Mots clés: Data stream; Divergence; Randomized approximation algorithm

Domaines: Calcul parallèle, distribué et partagé [cs.DC]

Emmanuelle Anceaume, Yann Busnel. An Information Divergence Estimation over Data Streams. 11th IEEE International Symposium on Network Computing and Applications (IEEE NCA12), Aug 2012, Cambridge, MA, United States. pp.Number 72. ⟨hal-00725097⟩