Signals 4 · free daily AI digest

A Principled Approach to Unsupervised Anomaly Detection

Paper recorded by Signals 4 on 2026-09-18 in cs.CV. Abstract reproduced from arXiv; link to the original below.

Published 2026-09-18 on arXiv · recorded by Signals 4 on 2026-09-21

Category: cs.CV · 计算机视觉 · first seen 2026-09-21

Abstract

Traditional unsupervised anomaly detection (UAD) methods are designed to flag or localise deviations from a normative distribution, ignoring the underlying generative mechanisms of the anomalies. Yet the nature of an anomaly is often as important as its presence. We reformulate UAD as a Bayesian inverse problem, in which the objective is to infer the most probable corruption responsible for each o

Read on arXiv →

#32 most recent of 270 cs.CV papers we have recorded · ↑ newer: VideoReloc: Long-Term Indoor Video Relocalization against a Kilobyte-S · ↓ older: PointLAM: Local Attentive Mamba for Efficient Point-based 3D Object De
Cite this page: A Principled Approach to Unsupervised Anomaly Detection: the #32 most recent of 270 cs.CV papers we have recorded (as of 2026-09-18). Source: Signals 4 (Signals API) — https://data.jiangzhang.ca/signals4/t/papers/a-principled-approach-to-unsupervised-anomaly-detection.html
Free to quote with attribution to “Signals 4 (Signals API)”. Machine-readable: papers.json
Related: More cs.CV papers · arXiv signals · All papers · Today in AI
Get 4 AI signals a day by email — free.
Subscribe free → See all plans →
Get 4 AI signals a day by email — free
All models · All repos · By company · Daily editions