Reproduce “Few-Shot Adaptation for Anomalous Sound Detection at Low False-Positive Rate” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Simone Lugani, Christof Pichler, Hannes Wegleiter, Christoph Feilmayr, Rodolfo Zunino, Edoardo Ragusa
The authors present a two-stage acoustic fault detector for industrial machinery. It learns normal sound, then fine-tunes its encoder on normal and few-shot fault examples while keeping the normality reference fixed. On recreated faults at one blast-furnace facility, adaptation raised sample-level true-positive rate at a 0.1% false-positive rate from 9.4% ± 0.8 to 60.6% ± 0.1, with results limited to held-out machines.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.