Using Umap to Inspect Audio Data for Unsupervised Anomaly Detection Under Domain-Shift Conditions

Fernandez, A; Plumbley, Mark D.

doi:10.31219/osf.io/ren6d

Cited by 1 publication

(2 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Furthermore, autoencoders can localize anomalies in the input space by visualizing an element-wise reconstruction error as done in [19], [21]. However, training ASD models by using an auxiliary task usually enhances their performance [46]. Even for IM-based models, performance can be significantly improved when utilizing meta information such as machine types.…”

Section: Introductionmentioning

confidence: 99%

“…This has been visualized using local interpretable model-agnostic explanations (LIME) [49] applied to sounds (SLIME) [50]. Furthermore, uniform manifold approximation and projection (UMAP) [51] has been used to visualize representations of the data such as stacked consecutive frames of log magnitude spectrograms, log-mel magnitude spectrograms, or openL3 embeddings [46].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Why Do Angular Margin Losses Work Well for Semi-Supervised Anomalous Sound Detection?

Wilkinghoff,

Kurth

2024

IEEE/ACM Trans. Audio Speech Lang. Process.

View full text Add to dashboard Cite

State-of-the-art anomalous sound detection systems often utilize angular margin losses to learn suitable representations of acoustic data using an auxiliary task, which usually is a supervised or self-supervised classification task. The underlying idea is that, in order to solve this auxiliary task, specific information about normal data needs to be captured in the learned representations and that this information is also sufficient to differentiate between normal and anomalous samples. Especially in noisy conditions, discriminative models based on angular margin losses tend to significantly outperform systems based on generative or one-class models. The goal of this work is to investigate why using angular margin losses with auxiliary tasks works well for detecting anomalous sounds. To this end, it is shown, both theoretically and experimentally, that minimizing angular margin losses also minimizes compactness loss while inherently preventing learning trivial solutions. Furthermore, multiple experiments are conducted to show that using a related classification task as an auxiliary task teaches the model to learn representations suitable for detecting anomalous sounds in noisy conditions. Among these experiments are performance evaluations, visualizing the embedding space with t-SNE and visualizing the input representations with respect to the anomaly score using randomized input sampling for explanation.

show abstract

Section: Introductionmentioning

confidence: 99%