Using UMAP to Inspect Audio Data for Unsupervised Anomaly Detection under Domain-Shift Conditions

PUBLISHED ON AUG 20, 2021 — CATEGORIES: publications

I’m very happy to announce that my first publication with Prof. Mark Plumbley, reflecting work done in the scope of the AI for Sound fellowship at CVSSP (Surrey University), has been accepted for presentation at the DCASE2021 Workshop.

We have briefly covered the main ideas in an entry at the AI4S page, check it out!

For a more detailed overview, check the resources below:

Poster (see PDF below for higher resolution):


Presentation:


Useful links:

Work supported by EPSRC grants EP/T019751/1 (AI for Sound) and EP/T022205/1 (JADE2 Tier 2 HPC facility)

TAGS: dcase, domain shift, interpretability, umap, unsupervised anomaly detection