Box-To-Box Transformations for Modeling Joint Hierarchies
Open accessAbstract
Learning representations of entities and relations in structured knowledge bases is an active area of research, with much emphasis placed on choosing the appropriate geometry to capture the hierarchical structures exploited in, for example, ISA or HASPART relations.Box embeddings (Vilnis et al., 2018;Li et al., 2019; Dasgupta et al., 2020), which represent concepts as n-dimensional hyperrectangles, are capable of embedding hierarchies when training on a subset of the transitive closure.In Patel et al. (2020), the authors demonstrate that only the transitive reduction is required and further extend box embeddings to capture joint hierarchies by augmenting the graph with new nodes.While it is possible to represent joint hierarchies with this method, the parameters for each hierarchy are decoupled, making generalization between hierarchies infeasible.In this work, we introduce a learned box-to-box transformation that respects the structure of each hierarchy.We demonstrate that this not only improves the capability of modeling cross-hierarchy compositional edges but is also capable of generalizing from a subset of the transitive reduction.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The method in [23] is not scalable because it needs to learn a quadratic number of transformation functions between all pairs of hierarchical relations.
“However, the method is not scalable since the model needs to learn a quadratic number of transformation functions between all pairs of hierarchical relations.”
From Modeling Heterogeneous Hierarchies with Relation-specific Hyperbolic Cones (2021) · cites this paper as [23]
Section 4.1 describes learning a box-to-box transformation to model a second hierarchical relation alongside the first. It does not say that extending the method requires transformations between every pair of relations, that their number grows quadratically, or that this makes the method unscalable. The Conclusion says extension beyond two relations is future work. (LLM, conf=0.78)
Experimental — accuracy not yet measuredThe method in [23] builds upon box embedding and models joint (two) hierarchies using two boxes as entity embeddings.
“Recently, [23] builds upon the box embedding and further models joint (two) hierarchies using two boxes as entity embeddings.”
From Modeling Heterogeneous Hierarchies with Relation-specific Hyperbolic Cones (2021) · cites this paper as [23]
BEUrRE makes use of a learned transform for each relation, differs from Ren et al. in that entities themselves are boxes, and differs from both Ren et al. and Dasgupta et al. in the structure of the learned transform.
“Similarly to Ren et al. (2020) and Dasgupta et al. (2021), we also make use of a learned transform for each relation, however we differ from Ren et al. (2020) in that entities themselves are boxes, and differ from both in the structure of the learned transform.”
From Probabilistic Box Embeddings for Uncertain Knowledge Graph Reasoning (2021) · cites this paper as Dasgupta et al., 2021
Patel et al. demonstrate the capability of box embeddings to jointly model two hierarchical relations, which is improved upon using a learned transform by Dasgupta et al.
“In (Patel et al., 2020), the authors demonstrate the capability of box embeddings to jointly model two hierarchical relations, which is improved upon using a learned transform in (Dasgupta et al., 2021).”
From Probabilistic Box Embeddings for Uncertain Knowledge Graph Reasoning (2021) · cites this paper as Dasgupta et al., 2021
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
WordNet: a lexical database for English
Miller, G. A. (1995) · Communications of the ACM · doi:10.1145/219717.219748 · Cited by 14,015 · see citing works
(2013) · Cited by 5,726 · see citing works
Attention, similarity, and the identification-categorization relationship.
Nosofsky, R. M. (1986) · Journal of Experimental Psychology General · doi:10.1037//0096-3445.115.1.39 · Cited by 2,207 · see citing works
Complex Embeddings for Simple Link Prediction
Trouillon, T., Welbl, J., Riedel, S., Gaussier, É., & Bouchard, G. (2016) · arXiv · doi:10.48550/arxiv.1606.06357 · Cited by 2,100 · see citing works
5.7 The WORDNET lexical database of English
Ooi, V. B. Y. (2019) · Edinburgh University Press eBooks · doi:10.1515/9781474471459-050 · Cited by 2,034 · see citing works
RotatE: Knowledge Graph Embedding by Relational Rotation in Complex Space
Sun, Z., Deng, Z.-H., Nie, J.-Y., & Tang, J. (2019) · arXiv · doi:10.48550/arxiv.1902.10197 · Cited by 1,414 · see citing works
Poincaré Embeddings for Learning Hierarchical Representations
Nickel, M., & Kiela, D. (2017) · arXiv · doi:10.48550/arxiv.1705.08039 · Cited by 667 · see citing works
Low-Dimensional Hyperbolic Knowledge Graph Embeddings
Chami, I., Wolf, A., Juan, D.-C., Sala, F., Ravi, S., & Ré, C. (2020) · Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/2020.acl-main.617 · Cited by 384 · see citing works
Combining prototypes: A selective modification model
Smith, E. E. (1988) · Cognitive Science · doi:10.1016/0364-0213(88)90011-0 · Cited by 344 · see citing works
Order-Embeddings of Images and Language
Vendrov, I., Kiros, R., Fidler, S., & Urtasun, R. (2015) · arXiv · doi:10.48550/arxiv.1511.06361 · Cited by 296 · see citing works
Hyperbolic Neural NetworksPreprint
Ganea, O.-E., Bécigneul, G., & Hofmann, T. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1805.09112 · Cited by 290 · see citing works
Word Representations via Gaussian EmbeddingPreprint
Vilnis, L., & McCallum, A. (2014) · arXiv (Cornell University) · doi:10.48550/arxiv.1412.6623 · Cited by 282 · see citing works
Self-citationQuery2box: Reasoning over Knowledge Graphs in Vector Space using Box Embeddings
Ren, H., Hu, W., & Leskovec, J. (2020) · arXiv · doi:10.48550/arxiv.2002.05969 · Cited by 135 · see citing works
Multi-relational Poincaré Graph Embeddings
Balažević, I., Allen, C., & Hospedales, T. (2019) · arXiv · doi:10.48550/arxiv.1905.09791 · Cited by 123 · see citing works
Hyperbolic Entailment Cones for Learning Hierarchical Embeddings
Ganea, O.-E., Bécigneul, G., & Hofmann, T. (2018) · arXiv · doi:10.48550/arxiv.1804.01882 · Cited by 115 · cited 3× · see citing works
“Hyperbolic space has been shown to be efficient in representing hierarchical relations, but also encounters difficulties in training (Nickel and Kiela, 2017; Ganea et al., 2018b; Chamberlain et al., 2017).”
BackgroundMethodProbabilistic Embedding of Knowledge Graphs with Box Lattice Measures
Vilnis, L., Li, X., Murty, S., & McCallum, A. (2018) · Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/p18-1025 · Cited by 111 · see citing works
Self-citationSmoothing the Geometry of Probabilistic Box Embeddings
Li, X., Vilnis, L., Zhang, D., Boratko, M., & McCallum, A. (2018) · International Conference on Learning Representations · Cited by 56 · see citing works
Self-citationRepresenting words as regions in vector space
Erk, K. (2009) · Conference on Computational Natural Language Learning · doi:10.3115/1596374.1596387 · Cited by 53 · see citing works
Neural Embeddings of Graphs in Hyperbolic Space
Chamberlain, B. P., Clough, J. R., & Deisenroth, M. P. (2017) · Spiral (Imperial College London) · Cited by 44 · see citing works
Learning to Predict Denotational Probabilities For Modeling Entailment
Lai, A., & Hockenmaier, J. (2017) · Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 1, Long Papers · doi:10.18653/v1/e17-1068 · Cited by 33 · see citing works
Improving Local Identifiability in Probabilistic Box Embeddings
Dasgupta, S. S., Boratko, M., Zhang, D., Vilnis, L., Li, X. L., & McCallum, A. (2020) · arXiv · doi:10.48550/arxiv.2010.04831 · Cited by 33 · see citing works
Self-citationHierarchical Density Order EmbeddingsPreprint
Athiwaratkun, B., & Wilson, A. G. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1804.09843 · Cited by 27 · see citing works
(2020) · Automated Knowledge Base Construction · Cited by 10 · see citing works
Self-citation
Flagged first, then by citations
8 of the 31 references couldn’t be matched to a known paper.
Research works printed but not linked: 4
- Ralph Abboud, İsmail İlkan Ceylan, Thomas Lukasiewicz, and Tommaso Salvatori. 2020. Boxe: A box embedding model for knowledge base completion. In Proceedings of the 34th Annual Conference on Neural Information Processing Systems NeurIPS.
- Ivana Balazevic, Carl Allen, and Timothy Hospedales. 2019b. TuckER: Tensor factorization for knowledge graph completion. In Proceedings of the 2019 Con- ference on Empirical Methods in Natural Language Processing and the 9th International Joint Confer- ence on Natural Language Processing.
- James A Hampton. 1991. The combination of proto- type concepts. The psychology of word meanings, pages 91-116.
- Robert M Nosofsky. 1986. Attention, similar- ity, and the identification-categorization relation- ship. Journal of experimental psychology: General, 115(1):39.
Cited by
Sorted by citations
Probabilistic Box Embeddings for Uncertain Knowledge Graph Reasoning
Chen, X., Boratko, M., Chen, M., Dasgupta, S. S., Li, X. L., & McCallum, A. (2021) · Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies · doi:10.18653/v1/2021.naacl-main.68 · Cited by 47 · see citing works
Self-citationModeling Heterogeneous Hierarchies with Relation-specific Hyperbolic Cones
Bai, Y., Ying, R., Ren, H., & Leskovec, J. (2021) · arXiv · doi:10.48550/arxiv.2110.14923 · Cited by 31 · see citing works
“Recently, [23] builds upon the box embedding and further models joint (two) hierarchies using two boxes as entity embeddings.”
MethodBinder: Hierarchical Concept Representation through Order Embedding of Binary Vectors
Gyurek, C., Talukder, N., & Hasan, M. A. (2024) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3637528.3671793 · Cited by 2 · see citing works
Box embeddings for extending ontologies: a data-driven and interpretable approach
Memariani, A., Glauer, M., Flügel, S., Neuhaus, F., Hastings, J., & Mossakowski, T. (2025) · Journal of Cheminformatics · doi:10.1186/s13321-025-01086-1 · Cited by 1 · see citing works
- Authors
- Dasgupta, S. S., Li, X. L., Boratko, M., Zhang, D., & McCallum, A.
- Affiliations
- University of Massachusetts Amherst (US)
- Venue
- Proceedings of the 6th Workshop on Representation Learning for NLP (RepL4NLP-2021)
- Published
- 2021
- Pages
- 277-288
- OpenAlex
- W3132675136
- Cited by
- 2
- References
- 31
- Citations by year
- 2025: 1 · 2024: 1
- FWCI
- 0.12
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- Locations
- Workshop on Representation Learning for NLP (RepL4NLP) (publishedVersion, cc-by)
- Funders
- National Science Foundation (1763618)
- University of Southern California
- Chan Zuckerberg Initiative
- Advanced Research Projects Agency
- Office of Naval Research
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Advanced Graph Neural Networks
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Physical Sciences › Computer Science › Artificial Intelligence › Semantic Web and Ontologies
- Keywords
- box embeddings, transitive reduction, knowledge base embedding
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
