Fréchet ChemNet Distance Is Blind to Double-Bond Stereochemistry

Fréchet ChemNet Distance (FCD) cannot distinguish <i>E</i>/<i>Z</i> stereoisomers, because the ChemNet tokenizer maps the SMILES bond-direction symbols / and <i>\</i> to the same unknown token; inverting every double-bond configuration in MolT5-large’s ChEBI-20 output (384 molecules altered) changes FCD by 0.001 resampling standard deviations while conditional per-bond accuracy falls from 98.97% to 1.03%, and MOSES and GuacaMol cannot expose the defect because their benchmark data contain no stereochemistry.

A stereochemistry-aware correction is monotone under synthetic corruption yet prefers fully inverted real output, illustrating that a metric operating only on unpaired generated and reference embeddings cannot recover per-caption correctness. Across real generators FCD carries no dependable signal about a more-than-3-fold difference in <i>E</i>/<i>Z</i> recall: on the double bonds where MolT5-large and BioT5-base both recover the molecular skeleton, recall is 0.26 versus 0.997, yet FCD does not separate the two modelsits ordering is statistically borderline, with a bootstrap interval that crosses zero.