%0 Conference Proceedings
%T Non-deterministic Behavior of Ranking-based Metrics when Evaluating Embeddings
%A Anguelos Nicolaou
%A Sounak Dey
%A V.Christlein
%A A.Maier
%A Dimosthenis Karatzas
%B International Workshop on Reproducible Research in Pattern Recognition
%D 2018
%V 11455
%F Anguelos Nicolaou2018
%O DAG; 600.121; 600.129
%O exported from refbase (http://refbase.cvc.uab.es/show.php?record=3178), last updated on Wed, 10 Mar 2021 11:46:48 +0100
%X Embedding data into vector spaces is a very popular strategy of pattern recognition methods. When distances between embeddings are quantized, performance metrics become ambiguous. In this paper, we present an analysis of the ambiguity quantized distances introduce and provide bounds on the effect. We demonstrate that it can have a measurable effect in empirical data in state-of-the-art systems. We also approach the phenomenon from a computer security perspective and demonstrate how someone being evaluated by a third party can exploit this ambiguity and greatly outperform a random predictor without even access to the input data. We also suggest a simple solution making the performance metrics, which rely on ranking, totally deterministic and impervious to such exploits.
%U https://link.springer.com/chapter/10.1007/978-3-030-23987-9_5
%U http://refbase.cvc.uab.es/files/NDC2018.pdf
%P 71-82