We introduce an approach to modular dimensionality reduction, allowing efficient learning of multiple complementary representations of the same object. Modules are trained by optimising an unsupervised cost function which balances two competing goals: Maintaining the inner product structure within the original space, and encouraging structural diversity between complementary representations. We derive an efficient learning algorithm which outperforms gradient based approaches without the need to choose a learning rate. We also demonstrate an intriguing connection with Dropout. Empirical results demonstrate the efficacy of the method for image retrieval and classification.
CITATION STYLE
Reeve, H. W. J., Mu, T., & Brown, G. (2019). Modular dimensionality reduction. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 11051 LNAI, pp. 605–619). Springer Verlag. https://doi.org/10.1007/978-3-030-10925-7_37
Mendeley helps you to discover research relevant for your work.