Inv-SENnet: Invariant Self Expression Network for clustering under biased data

TL;DR

Inv-SENnet introduces an adversarial framework that effectively removes biases related to unwanted attributes during subspace clustering, improving the interpretability and accuracy of clustering results on biased datasets.

Contribution

The paper presents a novel adversarial regularization method for bias removal in subspace clustering, addressing a key limitation of existing algorithms.

Findings

Effective bias removal demonstrated on synthetic datasets

Improved clustering accuracy on real-world datasets

Adversarial training reduces mutual information between data and biases

Abstract

Subspace clustering algorithms are used for understanding the cluster structure that explains the dataset well. These methods are extensively used for data-exploration tasks in various areas of Natural Sciences. However, most of these methods fail to handle unwanted biases in datasets. For datasets where a data sample represents multiple attributes, naively applying any clustering approach can result in undesired output. To this end, we propose a novel framework for jointly removing unwanted attributes (biases) while learning to cluster data points in individual subspaces. Assuming we have information about the bias, we regularize the clustering method by adversarially learning to minimize the mutual information between the data and the unwanted attributes. Our experimental result on synthetic and real-world datasets demonstrate the effectiveness of our approach.