Sustainable AI for Small Data: An Active Learning Approach

Researcher

Bjarne Jaster

Publications

• Bjarne Jaster, Martin Kohlhase (2023): Active Learning for Regression Problems with Ensemble Methods Full Publication
• Bjarne Jaster, Martin Kohlhase (2024): Active Learning für Regressionsprobleme mit Ensemble-Methoden – Untersuchung des Trade-Offs zwischen Qualität und Rechenaufwand Full Publication

Collaboration

Alaa Othman, Jörn Tebbe

Research Theme

R3 Sustainability & Efficiency

Tags

Active Learning Machine Learning

This project focusses mainly on Active Learning and Uncertainty estimation for Machine Learning Methods. Especially Active Learning gets more and more important with the large amounts of data produced every day, waiting to be used for Machine Learning. Uncertainty estimation has two application in this project: Firstly, it is used as an Active Learning criterion that identifies input-regions with little or no knowledge and secondly it enables a sustaiable use of Machine Learning through increased trust of the user.