Prediction of Adverse Biological Effects of Chemicals Using Knowledge Graph Embeddings

Tracking #: 2658-3872

This paper is currently under review
Erik B. Myklebust
Ernesto Jimenez Ruiz
Jiaoyan Chen
Raoul Wolf
Knut Erik Tollefsen

Responsible editor: 
Guest Editors DeepL4KGs 2021

Submission type: 
Full Paper
Semantic web technologies enable the interoperability of disparate data sources. We have created a knowledge graph based on major data sources used in ecotoxicological risk assessment. This facilities the use of the extensive library of semantic web tools. We have applied this knowledge graph to a important task in risk assessment, namely chemical effect prediction. Our extensive evaluation shows that by using knowledge graph embeddings we can increase the accuracy of effect prediction over a simple baseline. Furthermore, we have implemented a fine-tuning architecture which adapts the knowledge graph embeddings to the effect prediction task and leads to a better performance.
Full PDF Version: 
Under Review