Pith. sign in

REVIEW

Investigating Label Noise Sensitivity of Convolutional Neural Networks for Fine Grained Audio Signal Labelling

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1805.10880 v1 pith:5ASWJSNW submitted 2018-05-28 cs.SD cs.LGeess.AS

Investigating Label Noise Sensitivity of Convolutional Neural Networks for Fine Grained Audio Signal Labelling

classification cs.SD cs.LGeess.AS
keywords audiosignalconvolutionalfinegrainedlabelnetworksneural
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

We measure the effect of small amounts of systematic and random label noise caused by slightly misaligned ground truth labels in a fine grained audio signal labeling task. The task we choose to demonstrate these effects on is also known as framewise polyphonic transcription or note quantized multi-f0 estimation, and transforms a monaural audio signal into a sequence of note indicator labels. It will be shown that even slight misalignments have clearly apparent effects, demonstrating a great sensitivity of convolutional neural networks to label noise. The implications are clear: when using convolutional neural networks for fine grained audio signal labeling tasks, great care has to be taken to ensure that the annotations have precise timing, and are free from systematic or random error as much as possible - even small misalignments will have a noticeable impact.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.