SOURCE-LINKED INTELLIGENCE

What an odour descriptor corpus can and cannot measure: valence, attenuation, and the ceiling of the public record

arXiv · AI, language, vision and robotics · article · Sep 11, 2026 · UTC

Machine olfaction trains on pooled public descriptor corpora, but whether a shared descriptor word measures the same thing across corpora has not been tested, nor has the ceiling of what any of them can measure. We audit four corpora from Pyrfume. Conditioning on the molecule makes McNemar's test the exact conditional test of the corpus effect. Corpora disagree heterogeneously across descriptors ($I^2 = 80\%$) and non-uniformly with labelling breadth ($z = 17.2$), so no single offset repairs pooling. Median tetrachoric agreement is 0.795 against median $κ$ of 0.413: sources largely concur on w

Read original source ↗ Open in workspace

recordType

paper

region

Global