Skip to content
conftrace_
2026 EMNLP EMNLP 2026

Wasl at KnowledgeGraphEval 2026: Multi-Head Nested BIO Tagging and Constraint-Aware Relation Classification for Arabic Knowledge Graph Construction

AbstractWe describe Team Wasl’s systems for the KnowledgeGraphEval 2026 shared task at ArabicNLP 2026, which decomposes Arabic knowledge-graph construction into two evaluated stages: nested NER under domain shift and relation extraction between given mentions. For Subtask 1 (AdaptNER), we train a multi-head BIO tagger (one three-way O/B/I head per entity type over a shared pretrained Arabic encoder) on Wojood and combine three random seeds by span-level majority vote. For Subtask 2, forced-choice classification over the 40 WojoodRelations relation types, we fine-tune a typed-marker classifier with type-constraint decoding, ensembling five grouped folds by logit averaging. Our submissions scored 0.7483 micro-F1 on AdaptNER’s blind Konooz slice (ten unseen domains), 4th of the eight officially ranked teams and 11.8 points above the official baseline, and 0.9571 micro-F1 on relation extraction with the no-relation class masked at decoding, 2nd at the results release; re-scored by the organizers over all 41 classes, the same ensemble reaches 0.9328, 4th in the organizers’ final ranking. We analyze where the 18.6-point domain-shift tax on NER falls: chiefly on recall. We further show that type-constraint decoding and ensembling are partially redundant (+0.9 points on a single RE model, +0.1 on the ensemble) and that the surviving errors share an argument-type signature that no type constraint can separate.

More from EMNLP 2026

All EMNLP 2026 papers →