PulseAugur
EN
LIVE 13:16:52

New SPUNA framework detects covariate shift using weaker supervision

By PulseAugur Editorial · [2 sources] ·

Researchers have developed a new framework called Spectral PU Neighborhood Annotation (SPUNA) to detect covariate shift in computer vision systems. This geometry-aware approach uses Positive Unlabeled (PU) learning, requiring weaker supervision than traditional methods. SPUNA leverages the local manifold structure of visual features to progressively identify shifted data, achieving state-of-the-art performance and matching fully supervised methods. AI

IMPACT Introduces a novel method for improving the reliability of vision systems by detecting and adapting to data shifts.

RANK_REASON The cluster contains an academic paper detailing a new research framework.

Read on arXiv cs.LG →

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

New SPUNA framework detects covariate shift using weaker supervision

COVERAGE [2]

  1. arXiv cs.LG TIER_1 English(EN) · Firas Gabetni, Alexandre Rocchi Henry, Nacim Belkhir, Ziyi Liu, Gianni Franchi ·

    From Local Geometry to Global Pseudo Labeling for Robust Positive Unlabeled Learning under Covariate Shift

    arXiv:2605.31187v1 Announce Type: cross Abstract: Detecting covariate shift is critical for building reliable vision systems. While most prior work focuses on improving robustness to shift, explicitly detecting covariate shift remains underexplored. Existing approaches typically …

  2. arXiv cs.CV TIER_1 English(EN) · Gianni Franchi ·

    From Local Geometry to Global Pseudo Labeling for Robust Positive Unlabeled Learning under Covariate Shift

    Detecting covariate shift is critical for building reliable vision systems. While most prior work focuses on improving robustness to shift, explicitly detecting covariate shift remains underexplored. Existing approaches typically rely on fully supervised training, requiring label…

RELATED ENTITIES

RELATED TOPICS