Knowledge Graph Completion for Action Prediction on Situational Graphs -- A Case Study on Household Tasks

TL;DR

This paper explores the challenges of completing situational knowledge graphs for household tasks, highlighting the limitations of existing link prediction algorithms in this context.

Contribution

The study identifies unique characteristics of household action knowledge graphs that hinder standard link prediction methods, proposing a need for specialized approaches.

Findings

Standard link prediction algorithms underperform on household action graphs.

Many algorithms cannot outperform simple baselines in this domain.

Situational graphs have unique features affecting completion methods.

Abstract

Knowledge Graphs are used for various purposes, including business applications, biomedical analyses, or digital twins in industry 4.0. In this paper, we investigate knowledge graphs describing household actions, which are beneficial for controlling household robots and analyzing video footage. In the latter case, the information extracted from videos is notoriously incomplete, and completing the knowledge graph for enhancing the situational picture is essential. In this paper, we show that, while a standard link prediction problem, situational knowledge graphs have special characteristics that render many link prediction algorithms not fit for the job, and unable to outperform even simple baselines.