Faithful Embeddings for $$\mathcal{E}\mathcal{L}^{++}$$ Knowledge Bases

Xiong, Bo; Potyka, Nico; Tran, Trung-Kien; Nayyeri, Mojtaba; Staab, Steffen

doi:10.1007/978-3-031-19433-7_2

Cited by 10 publications

(6 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Geometric relational embeddings encode real-world relational knowledge by geometric objects such as convex regions like n-balls (Kulmanov et al, 2019), convex cones (Zhang et al, 2021;, axis-parallel boxes (Vilnis et al, 2018;Xiong et al, 2022c;Ren et al, 2020) and non-Euclidean manifold components (Xiong et al, 2022a). A key advantage of these geometric embeddings is that they nicely model the set-theoretic semantics that can be used to capture logical rules of KGs (Abboud et al, 2020), ontological axioms (Kulmanov et al, 2019;Xiong et al, 2022c), transitive closure (Vilnis et al, 2018), and logical query for multi-hop reasoning (Ren et al, 2020). Different from all previous work, ShrinkE is the first geometric embedding that aims at modeling inference patterns for hyper-relational KGs.…”

Section: A Supplemental Related Workmentioning

confidence: 99%

“…Link prediction on knowledge graphs (KGs) is a central problem for many KG-based applications Lukovnikov et al, 2017;Lu et al, 2023;Xiong et al, 2022b;Chen et al, 2022). Existing works (Sun et al, 2019;Bordes et al, 2013) have mostly studied link prediction on binary relational KGs, wherein each fact is represented by a triple, e.g., (Einstein, educated_at, University of Zurich).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Shrinking Embeddings for Hyper-Relational Knowledge Graphs

Xiong¹,

Nayyeri²,

Pan³

et al. 2023

Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)

View full text Add to dashboard Cite

Link prediction on knowledge graphs (KGs) has been extensively studied on binary relational KGs, wherein each fact is represented by a triple. A significant amount of important knowledge, however, is represented by hyperrelational facts where each fact is composed of a primal triple and a set of qualifiers comprising a key-value pair that allows for expressing more complicated semantics. Although some recent works have proposed to embed hyper-relational KGs, these methods fail to capture essential inference patterns of hyperrelational facts such as qualifier monotonicity, qualifier implication, and qualifier mutual exclusion, limiting their generalization capability. To unlock this, we present ShrinkE, a geometric hyper-relational KG embedding method aiming to explicitly model these patterns. ShrinkE models the primal triple as a spatial-functional transformation from the head into a relation-specific box. Each qualifier "shrinks" the box to narrow down the possible answer set and, thus, realizes qualifier monotonicity. The spatial relationships between the qualifier boxes allow for modeling core inference patterns of qualifiers such as implication and mutual exclusion. Experimental results demonstrate ShrinkE's superiority on three benchmarks of hyper-relational KGs.

show abstract

Section: A Supplemental Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Shrinking Embeddings for Hyper-Relational Knowledge Graphs

Xiong¹,

Nayyeri²,

Pan³

et al. 2023

Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)

View full text Add to dashboard Cite

show abstract

“…This simplification cannot fully express the logical structure of axioms. To overcome these limitations, Xiong et al proposed BoxEL [54] for embedding EL ++ ontologies. They modeled the concepts in the ontology as boxes (axis-aligned hyperrectangles).…”

Section: Description Logic Elmentioning

confidence: 99%

Neuro-Symbolic Techniques for Description Logic Reasoning (Student Abstract)

Singh

Mondal²,

Bhatia

et al. 2021

AAAI

View full text Add to dashboard Cite

With the goal to find scalable reasoning approaches, neuro-symbolic techniques have gained significant attention. However, the existing approaches do not take into account the inference capabilities of ontology languages that are based on expressive description logic (such as OWL 2). To fill this gap, we propose two approaches: an ontology-based embedding model for theories in EL++ description logic and a reinforcement learning-based solution for efficient tableau-based reasoning on description logic. We describe promising initial results of our efforts towards these directions and lay down the direction for future work.

show abstract

“…However, these methods fail to capture the rich semantics conveyed in entities and terms ( e.g ., synonyms, definitions, types), which are essential to handle the inconsistent naming conventions from multi-sources. Another line of work leverage neural embedding models [ 9 , 19 , 20 , 38 , 46 ] to represent entities as dense vectors using semantic attributes, structural properties, and alignment supervisions. These models perform better than unsupervised models when sufficient training samples are available.…”

Section: Introductionmentioning

confidence: 99%

HiPrompt: Few-Shot Biomedical Knowledge Fusion via Hierarchy-Oriented Prompting

Lü

Shen

Xiong

et al. 2023

Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval

View full text Add to dashboard Cite

Medical decision-making processes can be enhanced by comprehensive biomedical knowledge bases, which require fusing knowledge graphs constructed from different sources via a uniform index system. The index system often organizes biomedical terms in a hierarchy to provide the aligned entities with fine-grained granularity. To address the challenge of scarce supervision in the biomedical knowledge fusion (BKF) task, researchers have proposed various unsupervised methods. However, these methods heavily rely on ad-hoc lexical and structural matching algorithms, which fail to capture the rich semantics conveyed by biomedical entities and terms. Recently, neural embedding models have proved effective in semantic-rich tasks, but they rely on sufficient labeled data to be adequately trained. To bridge the gap between the scarce-labeled BKF and neural embedding models, we propose HiPrompt, a supervision-efficient knowledge fusion framework that elicits the few-shot reasoning ability of large language models through hierarchy-oriented prompts. Empirical results on the collected KG-Hi-BKF benchmark datasets demonstrate the effectiveness of HiPrompt.

show abstract

Faithful Embeddings for $$\mathcal{E}\mathcal{L}^{++}$$ Knowledge Bases

Cited by 10 publications

References 17 publications

Shrinking Embeddings for Hyper-Relational Knowledge Graphs

Shrinking Embeddings for Hyper-Relational Knowledge Graphs

Neuro-Symbolic Techniques for Description Logic Reasoning (Student Abstract)

HiPrompt: Few-Shot Biomedical Knowledge Fusion via Hierarchy-Oriented Prompting

Contact Info

Product

Resources

About