Ethical and Sustainability considerations for Knowledge Graphs based Machine Learning
2022 IEEE Fifth International Conference on Artificial Intelligence and Knowledge Engineering (AIKE), Carsten Felix Draschner, Hajira Jabeen, Jens Lehmann
TL;DR ⏱️
- Ethical and sustainability considerations in KG-based ML
- Responsible AI check, technical system setup
- Data insights, machine learning, training & evaluation, deployment
This paper presents ethical and sustainability considerations in KG-based ML. These are presented along the classical R&D life-cycle, including responsible AI check, technical system setup, data insights, machine learning, training & evaluation, and finally, deployment. Each of these steps is introduced by multiple sub-dimension. These are structured by definition, explaining challenges, providing examples, stating research questions, and offering recommendations. We also investigate data validity, bias, and privacy-related concerns of KG data in ML pipelines. The ethical and explainability dimension, as well as the sustainability aspect of processing power-intensive KG-based ML, are presented.