TRIX: A More Expressive Model for Zero-Shot Domain Transfer in Knowledge Graphs

Yucheng Zhang, Beatrice Bevilacqua, Mikhail Galkin, Bruno Ribeiro
Proceedings of the Third Learning on Graphs Conference, PMLR 269:12:1-12:28, 2025.

Abstract

Fully inductive knowledge graph models can be trained on multiple domains and subsequently perform zero-shot knowledge graph completion (KGC) in new unseen domains. This is an important capability towards the goal of having foundation models for knowledge graphs. In this work, we introduce a more expressive and capable fully inductive model, dubbed TRIX, which not only yields strictly more expressive triplet embeddings (head entity, relation, tail entity) compared to state-of-the-art methods, but also introduces a new capability: directly handling both entity and relation prediction tasks in inductive settings. Empirically, we show that TRIX outperforms the state-of-the-art fully inductive models in zero-shot entity and relation predictions in new domains, and outperforms large-context LLMs in out-of-domain predictions. The source code is available at https://github.com/yuchengz99/TRIX.

Cite this Paper

BibTeX
@InProceedings{pmlr-v269-zhang25a, title = {TRIX: A More Expressive Model for Zero-Shot Domain Transfer in Knowledge Graphs}, author = {Zhang, Yucheng and Bevilacqua, Beatrice and Galkin, Mikhail and Ribeiro, Bruno}, booktitle = {Proceedings of the Third Learning on Graphs Conference}, pages = {12:1--12:28}, year = {2025}, editor = {Wolf, Guy and Krishnaswamy, Smita}, volume = {269}, series = {Proceedings of Machine Learning Research}, month = {26--29 Nov}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v269/main/assets/zhang25a/zhang25a.pdf}, url = {https://proceedings.mlr.press/v269/zhang25a.html}, abstract = {Fully inductive knowledge graph models can be trained on multiple domains and subsequently perform zero-shot knowledge graph completion (KGC) in new unseen domains. This is an important capability towards the goal of having foundation models for knowledge graphs. In this work, we introduce a more expressive and capable fully inductive model, dubbed TRIX, which not only yields strictly more expressive triplet embeddings (head entity, relation, tail entity) compared to state-of-the-art methods, but also introduces a new capability: directly handling both entity and relation prediction tasks in inductive settings. Empirically, we show that TRIX outperforms the state-of-the-art fully inductive models in zero-shot entity and relation predictions in new domains, and outperforms large-context LLMs in out-of-domain predictions. The source code is available at https://github.com/yuchengz99/TRIX.} }
Endnote
%0 Conference Paper %T TRIX: A More Expressive Model for Zero-Shot Domain Transfer in Knowledge Graphs %A Yucheng Zhang %A Beatrice Bevilacqua %A Mikhail Galkin %A Bruno Ribeiro %B Proceedings of the Third Learning on Graphs Conference %C Proceedings of Machine Learning Research %D 2025 %E Guy Wolf %E Smita Krishnaswamy %F pmlr-v269-zhang25a %I PMLR %P 12:1--12:28 %U https://proceedings.mlr.press/v269/zhang25a.html %V 269 %X Fully inductive knowledge graph models can be trained on multiple domains and subsequently perform zero-shot knowledge graph completion (KGC) in new unseen domains. This is an important capability towards the goal of having foundation models for knowledge graphs. In this work, we introduce a more expressive and capable fully inductive model, dubbed TRIX, which not only yields strictly more expressive triplet embeddings (head entity, relation, tail entity) compared to state-of-the-art methods, but also introduces a new capability: directly handling both entity and relation prediction tasks in inductive settings. Empirically, we show that TRIX outperforms the state-of-the-art fully inductive models in zero-shot entity and relation predictions in new domains, and outperforms large-context LLMs in out-of-domain predictions. The source code is available at https://github.com/yuchengz99/TRIX.
APA
Zhang, Y., Bevilacqua, B., Galkin, M. & Ribeiro, B.. (2025). TRIX: A More Expressive Model for Zero-Shot Domain Transfer in Knowledge Graphs. Proceedings of the Third Learning on Graphs Conference, in Proceedings of Machine Learning Research 269:12:1-12:28 Available from https://proceedings.mlr.press/v269/zhang25a.html.

Related Material