Decomposing Uncertainty in Probabilistic Knowledge Graph Embeddings: Why Entity Variance Is Not Enough
Chorok Lee
TL;DR
The paper demonstrates a fundamental limitation of probabilistic KG embeddings: entity-level variances are relation-agnostic and fail to detect novel relational contexts, as shown by an impossibility result. It distinguishes two OOD types—emerging entities and novel contexts—and proves their signals are complementary. To address this, the authors propose Coverage-Augmented GP-KGE (CAGP), which combines semantic uncertainty from entity variance with structural uncertainty from entity-relation co-occurrence via an explicit binary coverage matrix, with a learnable mixing weight. Empirically, CAGP achieves strong temporal OOD detection (0.94–0.99 AUROC) across FB15k-237, WN18RR, and YAGO3-10, substantially outperforming relation-agnostic baselines (0.52–0.64 AUROC) and showing clear complementary behavior between the two signals. The approach is architecture-agnostic, scalable, and provides practical guidance for deploying robust uncertainty estimation in evolving knowledge graphs.
Abstract
Probabilistic knowledge graph embeddings represent entities as distributions, using learned variances to quantify epistemic uncertainty. We identify a fundamental limitation: these variances are relation-agnostic, meaning an entity receives identical uncertainty regardless of relational context. This conflates two distinct out-of-distribution phenomena that behave oppositely: emerging entities (rare, poorly-learned) and novel relational contexts (familiar entities in unobserved relationships). We prove an impossibility result: any uncertainty estimator using only entity-level statistics independent of relation context achieves near-random OOD detection on novel contexts. We empirically validate this on three datasets, finding 100 percent of novel-context triples have frequency-matched in-distribution counterparts. This explains why existing probabilistic methods achieve 0.99 AUROC on random corruptions but only 0.52-0.64 on temporal distribution shift. We formalize uncertainty decomposition into complementary components: semantic uncertainty from entity embedding variance (detecting emerging entities) and structural uncertainty from entity-relation co-occurrence (detecting novel contexts). Our main theoretical result proves these signals are non-redundant, and that any convex combination strictly dominates either signal alone. Our method (CAGP) combines semantic and structural uncertainty via learned weights, achieving 0.94-0.99 AUROC on temporal OOD detection across multiple benchmarks, a 60-80 percent relative improvement over relation-agnostic baselines. Empirical validation confirms complete frequency overlap on three datasets (FB15k-237, WN18RR, YAGO3-10). On selective prediction, our method reduces errors by 43 percent at 85 percent answer rate.