09.08.2022

MCML at KDD 2022: One Accepted Paper

28th ACM SIGKDD International Conference on Knowledge Discovery and Data (KDD 2022). Washington, DC, USA, 14.08.2022–18.08.2022

We are happy to announce that MCML researchers have contributed a total of 1 paper to KDD 2022. Congrats to our researchers!

Main Track (1 paper)

C. Leiber, L. G. M. Bauer, M. Neumayr, C. Plant and C. Böhm.
The DipEncoder: Enforcing Multimodality in Autoencoders.
KDD 2022 - 28th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Washington, DC, USA, Aug 14-18, 2022. DOI

Abstract

Hartigan’s Dip-test of unimodality gained increasing interest in unsupervised learning over the past few years. It is free from complex parameterization and does not require a distribution assumed a priori. A useful property is that the resulting Dip-values can be derived to find a projection axis that identifies multimodal structures in the data set. In this paper, we show how to apply the gradient not only with respect to the projection axis but also with respect to the data to improve the cluster structure. By tightly coupling the Dip-test with an autoencoder, we obtain an embedding that clearly separates all clusters in the data set. This method, called DipEncoder, is the basis of a novel deep clustering algorithm. Extensive experiments show that the DipEncoder is highly competitive to state-of-the-art methods.

MCML Authors