Objective | Problem Statement:
Explaining Neural Models is an unsolved research problem. Especially explaining a Graph Neural Model has attracted lot of interest in the research community. Self Explaining AI idea introduced by [Elton, 2020] can be used to explain Link Prediction GNN models. A desired solution will have human explainable predictions rather than just interpreting what the model is doing.
Approach: Design a Mutual Information based solution with creative user interfaces to evaluate the explanations.
Related Work: 1. Ganesan, Balaji, Gayatri Mishra, Srinivas Parkala, Neeraj R. Singh, Hima Patel, and Somashekar Naganna. "Link Prediction using Graph Neural Networks for Master Data Management." arXiv preprint arXiv:2003.04732 (2020). https://arxiv.org/abs/2003.04732
2. Elton, Daniel C. "Self-explaining AI as an alternative to interpretable AI." arXiv preprint arXiv:2002.05149 (2020). https://arxiv.org/abs/2002.05149 |