L3S Best Paper of the Quarter (Q2/2024)
Category: Knowledge Graphs and Causal Models
CauseKG: A Framework Enhancing Causal Inference With Implicit Knowledge Deduced From Knowledge Graphs
Authors: Hao Huang, Maria-Esther Vidal
Published in IEEE Access 12: 61810-61827 (2024)
https://ieeexplore.ieee.org/document/10510291
The paper in a nutshell:
This paper presents CauseKG, a framework for asking and answering causal questions using knowledge graphs (KGs).
For example, CauseKG can be used to answer questions such as:
(1) How much does prestige affect the acceptance of a paper?
(2) To what extent does smoking cause lung cancer?
KGs store data in a semi-structured graph format and represent the meaning of the data in a machine-understandable way. CauseKG uses both the data and its meaning in KGs to answer causal questions more accurately.
Which problem do you solve with your research?
This paper solves the problem of accurately estimating causal effects (i.e., answering causal questions) from data.
What is the potential impact of your findings?
Our findings show that CauseKG can improve decision-making in various areas, such as healthcare, social sciences, and economies.
For example, it can help patients choose more effective treatments; it can help business managers understand the effect of a marketing strategy on sales.
What is new about the research?
CauseKG is a novel framework that considers the semantics of data and exploits the implicit facts inferred from KG to improve causal effect estimation.
Existing methods often neglect these implicit facts, which can lead to incorrect causal conclusions.
Paper link:
Huang, H. and Vidal, M.E., 2024. CauseKG: A Framework Enhancing Causal Inference With Implicit Knowledge Deduced From Knowledge Graphs. IEEE Access.
