A Survey on Hyper-Relational Knowledge Graph Embedding Models for Link Prediction Task

PhD Qualifying Examination


Title: "A Survey on Hyper-Relational Knowledge Graph Embedding Models for Link 
Prediction Task"

by

Mr. Yubo WANG


Abstract:

With the development of the Knowledge Graph (KG), KG embedding has also 
attracted a lot of attention, as these models can project entities and 
relations in KG into vector space, thereby carring out a variety of downstream 
tasks. However, KG's triplet structure is argued as insufficient to reflect 
real-world facts. Hence, a more advanced structure of KG, namely 
hyper-relational KG (HKG), was proposed. More recently, due to the rise of 
research on HKGs, researchers have begun to extend classic KG embedding models 
to HKGs.

Inspired by these studies, this survey is dedicated to introducing some KG 
embedding models that take different technical routes and their evolution to 
the more general HKG embedding models. Furthermore, this survey also introduces 
movie dels that were originally designed for HKG embedding. Lastly, this survey 
will give a systematic understanding of the essential difference between 
embedding models on KGs and HKGs and analyze HKG embedding models based on 
their performance on the Link Prediction task.


Date:                   Friday, 12 July 2024

Time:                   4:00pm - 6:00pm

Venue:                  Room 5506
                        Lifts 25/26

Committee Members:      Prof. Lei Chen (Supervisor)
                        Prof. Xiaofang Zhou (Chairperson)
                        Dr. Junxian He
                        Dr. Lei Li (HKUST-GZ)