Search papers, labs, and topics across Lattice.
This paper addresses the issue of candidate-set interference in continual knowledge graph embedding, where the admission of new entities can negatively impact the ranking of historical answers despite their preserved scores. The authors introduce Matched Excess-Outranker Regularization (MEOR), a novel host-level objective that mitigates this interference by penalizing excessive newcomer competition against structurally matched old references. Experimental results demonstrate that MEOR improves the mean reciprocal rank (MRR) of historical answers and effectively reduces candidate-set interference across multiple settings, highlighting its potential to enhance continual learning systems without altering their foundational architectures.
Candidate admission can lead to significant rank loss for historical answers, but MEOR effectively mitigates this interference, improving MRR without changing the underlying embedding architecture.
Continual knowledge graph embedding updates entity and relation representations as a graph grows. Existing methods primarily address catastrophic forgetting, but entity admission also changes the candidate universe of every compatible query. A historical answer can therefore lose rank even when its score and its ordering among old entities are preserved. We formalize this effect as candidate-set interference and introduce Matched Excess-Outranker Regularization (MEOR), a host-level objective that compares smooth answer-relative newcomer pressure with score-blind, structurally matched old references. Its one-sided penalty acts only when newcomer competition exceeds the matched reference, preserving the host learner's signal for legitimate new entities. Across eight paired runs on ENTITY-ComplEx, MEOR improves historical current-universe mean reciprocal rank (MRR) by 0.0057 over replay and reduces candidate-set interference by 0.0055, with one-sided 95% lower bounds of 0.0052 and 0.0051, respectively. It satisfies the preservation criteria for old-universe ranking and newcomer acquisition and improves historical current-universe MRR over persistent calibration, matched maximum regularizer (MMR), and unmatched old regularizer (UOR). Direct ablations support each component of its reference construction and aggregation. Adding MEOR also improves historical ranking in all ten reported FBInc-S and FBInc-L host and backbone settings, with every paired 95% confidence interval excluding zero. These results establish candidate admission as a distinct source of continual rank loss and show that it can be controlled without replacing the underlying embedding architecture or continual learner.
