Ir directamente a la navegación principal Ir directamente a la búsqueda Ir directamente al contenido principal

Non-deterministic Behavior of Ranking-Based Metrics When Evaluating Embeddings

Anguelos Nicolaou*, Sounak Dey, Vincent Christlein, Andreas Maier, Dimosthenis Karatzas

*Autor correspondiente de este trabajo

Producción científica: Capítulo de libroCapítuloInvestigaciónrevisión exhaustiva

Resumen

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.

Idioma originalInglés
Título de la publicación alojadaReproducible Research in Pattern Recognition - Second International Workshop, RRPR 2018, Revised Selected Papers
EditoresBertrand Kerautret, Miguel Colom, Daniel Lopresti, Pascal Monasse, Hugues Talbot
Páginas71-82
Número de páginas12
DOI
EstadoPublicada - 2019

Serie de la publicación

NombreLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volumen11455 LNCS
ISSN (versión impresa)0302-9743
ISSN (versión digital)1611-3349

Huella

Profundice en los temas de investigación de 'Non-deterministic Behavior of Ranking-Based Metrics When Evaluating Embeddings'. En conjunto forman una huella única.

Citar esto