Research map / Label-efficient and robust learning / OOD and anomaly detection
HARMONY: Hierarchical Anchor Retrieval on Manifold for Oblivious-source Acoustic Anomaly Detection
Yuhang Zhang, Yang Li, Mingyuan Li, Qiang Wu
Read the paperCheck your idea against 137,217 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes HARMONY, a metadata-free universal acoustic anomaly detection method that uses hierarchical anchor retrieval on a manifold to monitor mixed machine populations without machine IDs.
Problem
Acoustic anomaly detection usually needs per-machine or per-domain training with explicit metadata, but machine identifiers are often incomplete, unreliable, or unavailable in deployment; the goal is a single model that handles mixed machine populations without IDs.
Approach
Argues the main failure mode of metadata-free universal ASD is not a weak detector but the lack of machine-specific reference structure; uses hierarchical anchor retrieval on a manifold to recover the relevant reference context for each oblivious input (based on the title and partial abstract only).
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| remove the learned projection (use raw BYOL-A features) (ablation) | HARMONY · DCASE 2020 | AUC (%): 62.40 -> 90.00 | Improved |
| add spectral anchor initialization (ablation) | HARMONY · DCASE 2020 | AUC (%): 81.40 -> 90.00 | Improved |
| add soft assignment (ablation) | HARMONY · DCASE 2020 | AUC (%): 87.60 -> 90.00 | Improved |
| replace flat search with hierarchical IVF retrieval (ablation) | HARMONY · DCASE 2020 | AUC (%): 90.00 -> 89.80; latency 42.3 ms -> 5.4 ms | Mixed |
| train per-region detectors on HARMONY-routed partitions instead of unified training | PatchCore · DCASE 2020 | AUC (%): 62.10 -> 91.80 | Improved |
| train per-region detectors on repartitioned data | Audio-MAE · DCASE 2020 | AUC (%): 86.10 -> 92.40 | Improved |
| per-region training increases cost for trained detectors | Audio-MAE · DCASE 2020 | training time (h): 14.2 -> 108.0 | Hurt |
| apply HNSW to the HARMONY representation instead of HARMONY routing (ablation) | HARMONY · DCASE 2020 | AUC (%): 89.80 -> 89.71 | No clear effect |
| use vMF soft prototypes instead of HARMONY (ablation) | HARMONY · DCASE 2020 (routed) | AUC (%): 89.80 -> 85.71 | Hurt |
| add projection to raw BYOL-A features (geometry metrics) (ablation) | HARMONY · DCASE 2020 | quantization error: 0.842 -> 0.181 | Improved |
| contaminate training data with anomalies (ablation) | HARMONY · DCASE 2020 | AUC (%): 89.65 -> 83.31 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Dinomaly — Dinomaly: The Less Is More Philosophy in Multi-Class Unsupervised Anomaly Detection (CVPR 2025) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- RAID: Retrieval-Augmented Anomaly Detection (CVPR 2026 · similarity 0.84)
- UniMMAD: Unified Multi-Modal and Multi-Class Anomaly Detection via MoE-Driven Feature Decompression (CVPR 2026 · similarity 0.83)
- Towards Lightweight, Model-Agnostic and Diversity-Aware Active Anomaly Detection (ICLR 2023 · similarity 0.82)
- Energy-Based Models for Anomaly Detection: A Manifold Diffusion Recovery Approach (NeurIPS 2023 · similarity 0.82)
- EchoTraffic: Enhancing Traffic Anomaly Understanding with Audio-Visual Insights (CVPR 2025 · similarity 0.82)
- Adaptive Compression and Targeted Perturbation: A Unified Framework for Generalized Audio Deepfake Detection (NeurIPS 2026 · similarity 0.81)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
