Improving Inductive Link Prediction Using Hyper-relational Facts

Ali, Mehdi; Berrendorf, Max; Galkin, Mikhail; Thost, Veronika; Ma, Tengfei; Tresp, Volker; Lehmann, Jens

doi:10.1007/978-3-030-88361-4_5

Cited by 11 publications

(15 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this survey, we also regard logical rules of different forms, such as Horn clause, Datalog rules and SWRL 4 rules, as well as their soft or fuzzy extensions (i.e., weighted rules) [135], within the scope of KG. This is because many of these rules can be transformed into equivalent relational facts and ontological knowledge, and vice versa [77,94].…”

Section: Definition and Scopementioning

confidence: 99%

“…It predicts the object of a triple, and can explain the prediction by finding out paths from the subject to the object. Ali et al [4] aimed at predicting relations between seen entities and unseen entities (semi-inductive setting), and between unseen entities (fully-inductive setting), utilizing not only the triples but also their Wikidata qualifiers, each of which is composed of a relation and an entity for describing the triple. For fully-inductive setting, they initialized the entity embeddings by entities' textual information using Sentence BERT [146], and then propagated to update the entities' embeddings by a graph encoder named StarE [61].…”

Section: Embeddingmentioning

confidence: 99%

“…Besides complex neural networks, some simpler propagation operations have also been considered for embedding unseen entities. Ali et al [4] used a linear projection from the entity embedding to the relation embedding for the semi-inductive setting. Dai et al [43] used two modules: an estimator which calculates a candidate set of embeddings for an unseen entity according to its all associated triples using the translation operation of TransE [20] or RotatE [167],…”

Section: Embeddingmentioning

confidence: 99%

“…The current methods often aim to fully utilize these triples, mainly by methods of the propagation-based paradigm [3,4,15,43,68,184,225], the transfer-based paradigm [29,107,153,172] and the optimization-based paradigm [8].…”

Section: Kg Completion With Unseen Entitiesmentioning

confidence: 99%

See 3 more Smart Citations

Zero-shot and Few-shot Learning with Knowledge Graphs: A Comprehensive Survey

Chen¹,

Geng²,

Chen³

et al. 2021

Preprint

View full text Add to dashboard Cite

Machine learning methods especially deep neural networks have achieved great success but many of them often rely on a number of labeled samples for training. In real-world applications, we often need to address sample shortage due to e.g., dynamic contexts with emerging prediction targets and costly sample annotation. Therefore, low-resource learning, which aims to learn robust prediction models with no enough resources (especially training samples), is now being widely investigated. Among all the low-resource learning studies, many prefer to utilize some auxiliary information in the form of Knowledge Graph (KG), which is becoming more and more popular for knowledge representation, to reduce the reliance on labeled samples. In this survey, we very comprehensively reviewed over 90 papers about KG-aware research for two major low-resource learning settings -zero-shot learning (ZSL) where new classes for prediction have never appeared in training, and few-shot learning (FSL) where new classes for prediction have only a small number of labeled samples that are available. We first introduced the KGs used in ZSL and FSL studies as well as the existing and potential KG construction solutions, and then systematically categorized and summarized KG-aware ZSL and FSL methods, dividing them into different paradigms such as the mapping-based, the data augmentation, the propagation-based and the optimization-based. We next presented different applications, including not only KG augmented prediction tasks in Computer Vision and Natural Language Processing (e.g., image classification, visual question answering, text classification and knowledge extraction), but also tasks for KG curation (e.g., inductive KG completion), and some typical evaluation resources for each task. We eventually discussed some challenges and future directions on aspects such as new learning and reasoning paradigms, and the construction of high quality KGs.

show abstract

Section: Definition and Scopementioning

confidence: 99%

Section: Embeddingmentioning

confidence: 99%

Section: Embeddingmentioning

confidence: 99%

Section: Kg Completion With Unseen Entitiesmentioning

confidence: 99%

See 2 more Smart Citations

Zero-shot and Few-shot Learning with Knowledge Graphs: A Comprehensive Survey

Chen¹,

Geng²,

Chen³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…For instance (see Fig. 1 in [Ali et al, 2021a]), the directorgenre pattern from the seen graph allows to predict a missing genre link for The Martian in the unseen subgraph.…”

Section: Introductionmentioning

confidence: 99%

Improving Inductive Link Prediction Using Hyper-Relational Facts (Extended Abstract)

Choudhury

Gupta

Kochenderfer

2022

Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence

Self Cite

View full text Add to dashboard Cite

We present a scalable planning algorithm for multi-agent sequential decision problems that require dynamic collaboration. Teams of agents need to coordinate decisions in many domains, but naive approaches fail due to the exponential growth of the joint action space with the number of agents. We circumvent this complexity through an anytime approach that allows us to trade computation for approximation quality and also dynamically coordinate actions. Our algorithm comprises three elements: online planning with Monte Carlo Tree Search (MCTS), factorizing local agent interactions with coordination graphs, and selecting optimal joint actions with the Max-Plus method. On the benchmark SysAdmin domain with static coordination graphs, our approach achieves comparable performance with much lower computation cost than the MCTS baselines. We also introduce a multi-drone delivery domain with dynamic, i.e., state-dependent coordination graphs, and demonstrate how our approach scales to large problems on this domain that are intractable for other MCTS methods.

show abstract