Metraipolis › archive › dc_271d1a2520ea

HiTS-CL: A Continual Learning Framework for Long-Horizon Temporal Knowledge Graph Extrapolation — arXiv:2609.36559

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.36559
License (as archived)
CC0-1.0
Archived
2026-10-07T16:26:03.931Z
SHA-256 of the archived text
df17f4038028262c3a82cfe18278ec410301c67beeaa2de4e13d9200c58f9268
Stable URL of this page
https://metraipolis.com/documento/dc_271d1a2520ea

How to cite

HiTS-CL: A Continual Learning Framework for Long-Horizon Temporal Knowledge Graph Extrapolation — arXiv:2609.36559 — arXiv (en), https://arxiv.org/abs/2609.36559 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_271d1a2520ea (Metraipolis archive dc_271d1a2520ea, sha256 df17f4038028262c3a82cfe18278ec410301c67beeaa2de4e13d9200c58f9268)
BibTeX
@misc{dc271d1a2520ea,
  title = {HiTS-CL: A Continual Learning Framework for Long-Horizon Temporal Knowledge Graph Extrapolation — arXiv:2609.36559},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.36559},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_271d1a2520ea; sha256 df17f4038028262c3a82cfe18278ec410301c67beeaa2de4e13d9200c58f9268},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.36559. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Extrapolative temporal knowledge graph reasoning (TKGR) predicts future facts from historical snapshots. Most existing methods train once on an early prefix of the timeline and then use a frozen model for all future timestamps. We argue that this fixed-prefix protocol is misaligned with extrapolation. It learns from a static prefix, whereas the target stream is non-stationary: new entities and facts emerge, temporal dependencies shift across regimes, and recurring historical signals must be refreshed online. As a result, models trained only on early snapshots become outdated and degrade over long horizons. We address this mismatch by formulating extrapolative TKGR as continual learning over streaming snapshots. Under this view, effective extrapolation must jointly handle current dynamics, stable knowledge, and recurring historical evidence. Based on these requirements, we propose History-enhanced Two-Step Continual Learning (HiTS-CL), a backbone-agnostic continual learning framework for extrapolative TKGR. HiTS-CL tracks current dynamics via continual fine-tuning, preserves stable knowledge via multi-teacher adaptive distillation, and retains recurring historical evidence via a selective memory of recent and frequent facts. We integrate HiTS-CL into five representative TKGR backbones and evaluate it on four benchmark datasets. HiTS-CL consistently improves extrapolation accuracy, reduces long-horizon degradation, and outperforms strong continual-learning baselines, including a recent method for temporal knowledge graphs. Source code and data are available at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-30 in cs.AI).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.36559 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-07T15:54:53.859Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).