Research map / Label-efficient and robust learning / OOD and anomaly detection

Quantifying Statistical Significance of Deep Nearest Neighbor Anomaly Detection via Selective Inference

NeurIPS 2025 · Label-efficient and robust learning › OOD and anomaly detection · approach: Unified multi-class anomaly detection · 0 citations (OpenAlex)

Read the paperCheck your idea against 108,967 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 selective inference to give p-values/statistical significance for deep k-nearest-neighbor anomaly detection decisions, using only normal training data.

Problem

Semi-supervised anomaly detection with deep kNN scoring lacks a mechanism to quantify statistical uncertainty or significance of its detections, which is needed when only normal data is available.

Approach

Applies selective inference to deep kNN anomaly detection so that the significance of detected anomalies is tested conditional on the selection event, yielding valid p-values despite the data-driven detection procedure.

Experimental setup

Builds on: deep kNN anomaly detection selective inference

Closest papers by meaning

• One-to-Normal: Anomaly Personalization for Few-shot Anomaly Detection (NeurIPS 2024 · similarity 0.85)
• Beyond Normal References: Discriminative Few-Shot Anomaly Detection (NeurIPS 2026 · similarity 0.85)
• RealNet: A Feature Selection Network with Realistic Synthetic Anomaly for Anomaly Detection (CVPR 2024 · similarity 0.85)
• Online Adaptive Anomaly Thresholding with Confidence Sequences (ICML 2024 · similarity 0.85)
• We Need to Rethink Benchmarking in Anomaly Detection (NeurIPS 2026 · similarity 0.85)
• An Evidence-Based Post-Hoc Adjustment Framework for Anomaly Detection Under Data Contamination (NeurIPS 2025 · similarity 0.84)