Addressing Feature Imbalance in Sound Source Separation

TL;DR

This paper introduces FEABASE, a novel method to address feature imbalance in sound source separation, improving neural network performance by balancing reliance on spatial and timbre features.

Contribution

The paper presents FEABASE, a new technique to mitigate feature preference in high-dimensional regression tasks like source separation, enhancing data utilization.

Findings

Effective in balancing spatial and timbre features

Improves source separation performance

Addresses feature preference in high-dimensional tasks

Abstract

Neural networks often suffer from a feature preference problem, where they tend to overly rely on specific features to solve a task while disregarding other features, even if those neglected features are essential for the task. Feature preference problems have primarily been investigated in classification task. However, we observe that feature preference occurs in high-dimensional regression task, specifically, source separation. To mitigate feature preference in source separation, we propose FEAture BAlancing by Suppressing Easy feature (FEABASE). This approach enables efficient data utilization by learning hidden information about the neglected feature. We evaluate our method in a multi-channel source separation task, where feature preference between spatial feature and timbre feature appears.