Explainable Benchmarking

Quan Nian Zhang
Research Theme
R3 Sustainability & Efficiency
Explainable Benchmarking Knowledge Graphs

This project aims to explain benchmarking results of machine learning algorithms to support developers in improving their systems. The first step focuses on explaining evaluation results of question answering systems. This is done by representing the input data of the question answering system as a knowledge graph, and then we further populate this knowledge graph and apply class expression learning to identify the question groups.