Categorical Unsupervised Variational Acoustic Clustering

Luan Vinícius Fiorio, Ivana Nikoloska, Ronald M. Aarts · 2025

Read Categorical Unsupervised Variational Acoustic Clustering by Luan Vinícius Fiorio, Ivana Nikoloska, Ronald M. Aarts (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-07652.txt.

Description

Luan Vinícius Fiorio, Ivana Nikoloska, and Ronald Aarts cluster acoustic scenes categorically in two thousand twenty five. Urban sound data overlap heavily in time and frequency. Soft clustering blurs when points pile on top of each other. A categorical model enforces sharper cluster assignments instead here.

The Gumbel-Softmax distribution makes the categorical choice trainable by gradient methods. Backpropagation flows through the soft approximation during learning. Softmax temperature becomes the main tuning knob for sharpness. Results impress across all the datasets they tried. Strong overlap no longer defeats the clustering quality.

Catalog record

CREATORLuan Vinícius Fiorio, Ivana Nikoloska, Ronald M. Aarts
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE23 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-07652

Browse the full library · About the Tower of Babel