Lightweight Inference-Time Personalization for Frozen Knowledge Graph Embeddings
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866908734250287104 |
|---|---|
| author | Oguztuzun, Ozan Oguztuzun, Cerag |
| author_facet | Oguztuzun, Ozan Oguztuzun, Cerag |
| contents | Foundation models for knowledge graphs (KGs) achieve strong cohort-level performance in link prediction, yet fail to capture individual user preferences; a key disconnect between general relational reasoning and personalized ranking. We propose GatedBias, a lightweight inference-time personalization framework that adapts frozen KG embeddings to individual user contexts without retraining or compromising global accuracy. Our approach introduces structure-gated adaptation: profile-specific features combine with graph-derived binary gates to produce interpretable, per-entity biases, requiring only ${\sim}300$ trainable parameters. We evaluate GatedBias on two benchmark datasets (Amazon-Book and Last-FM), demonstrating statistically significant improvements in alignment metrics while preserving cohort performance. Counterfactual perturbation experiments validate causal responsiveness; entities benefiting from specific preference signals show 6--30$\times$ greater rank improvements when those signals are boosted. These results show that personalized adaptation of foundation models can be both parameter-efficient and causally verifiable, bridging general knowledge representations with individual user needs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_22398 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Lightweight Inference-Time Personalization for Frozen Knowledge Graph Embeddings Oguztuzun, Ozan Oguztuzun, Cerag Artificial Intelligence Foundation models for knowledge graphs (KGs) achieve strong cohort-level performance in link prediction, yet fail to capture individual user preferences; a key disconnect between general relational reasoning and personalized ranking. We propose GatedBias, a lightweight inference-time personalization framework that adapts frozen KG embeddings to individual user contexts without retraining or compromising global accuracy. Our approach introduces structure-gated adaptation: profile-specific features combine with graph-derived binary gates to produce interpretable, per-entity biases, requiring only ${\sim}300$ trainable parameters. We evaluate GatedBias on two benchmark datasets (Amazon-Book and Last-FM), demonstrating statistically significant improvements in alignment metrics while preserving cohort performance. Counterfactual perturbation experiments validate causal responsiveness; entities benefiting from specific preference signals show 6--30$\times$ greater rank improvements when those signals are boosted. These results show that personalized adaptation of foundation models can be both parameter-efficient and causally verifiable, bridging general knowledge representations with individual user needs. |
| title | Lightweight Inference-Time Personalization for Frozen Knowledge Graph Embeddings |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2512.22398 |