Baselines

Organiser-run baseline systems for both competitions, published before the evaluation window. These are organiser baselines — participant submissions are not added to these tables. Participant results live on the CodaBench leaderboards and appear on the results page once the window opens (12 July 2026).

Track 1 · Subtrack 1 — global equivalence alignment

Repaired coherence-aware P/R/F1

Macro-averaged over the three pairs (NCIT–DOID, SNOMED–FMA, SNOMED–NCIT); sorted by the headline column. See evaluation metrics.

System F1 (repaired)P (repaired)R (repaired)
LogMapLt 0.625 0.628 0.635
AML 0.535 0.451 0.766
LogMap 0.523 0.392 0.803
BERTMap 0.516 0.821 0.401
BERTMapLt 0.508 0.785 0.403
Naive-Lexical 0.161 0.424 0.126

Track 1 · Subtrack 2 — local equivalence ranking

MRR and Hits@k

Macro-averaged over the three pairs (NCIT–DOID, SNOMED–FMA, SNOMED–NCIT); sorted by the headline column. See evaluation metrics.

System MRRHits@1Hits@5Hits@10
Naive-Lexical 0.731 0.656 0.814 0.863
SapBERT 0.716 0.638 0.809 0.850
BERTMapLt 0.583 0.524 0.636 0.692
BERTMap-ss 0.496 0.463 0.510 0.539
BERTMap 0.445 0.421 0.447 0.470

About these baselines

Organiser baselines (macro over the three pairs). Global = semi-supervised test-slice (option a): repaired coherence-aware F1 headline. Participant results appear on the CodaBench leaderboards. The global-alignment headline is the repaired coherence-aware F1; local ranking is headlined by MRR. Full data: leaderboard.json. Participant results appear on the results page, fed by the CodaBench leaderboards.

Reference coherence

Reference coherence — before vs after repair

The Task-1 reference alignment is grounded in UMLS/Mondo and can be logically incoherent; the repair procedure removes or weakens the incoherence-inducing correspondences. This table reports the coherence of each pair's reference before and after repair — the number of unsatisfiable named classes in the merged ontology and the corresponding incoherence degree. All three repaired references are coherent (0 unsatisfiable classes).

Pair Unsat. (before) Incoherence degree (before) Unsat. (after) Incoherence degree (after)
NCIT–DOID 17,850† 7.7% 0 0%
SNOMED–FMA 0 0%
SNOMED–NCIT 466,306† 78% 0 0%

Unsatisfiable classes are counted in the merged ontology (Osrc ∪ Otgt ∪ R) under the ELK reasoner; the incoherence degree is that count over the merged named-class signature. † denotes an ELK lower bound (a sound under-approximation; the exact figure is pending). The all-equivalence standard SNOMED–FMA merge does not classify under ELK even at a 128 GB heap, so its before-repair figure is unavailable (—). Machine-readable data: reference_coherence.json.

Editions

Baselines and results by edition

OAEI Bio-ML 2026 is the current edition. As the track matures, each edition's final baselines and participant results will be archived here, and changes between editions are recorded in the changelog.

EditionStatusBaselinesMachine-readable snapshot
2026 current — evaluation & leaderboard open 12 July 2026 this page leaderboard.json