Challenge 1. Accurately model multi-level explicit features from noisy and sparse data Challenge 2.
Generate explanations that are easy for common users to understand from the multi-level structure Main idea 1. Build an initial network based on an explainable deep structure (e.g., knowledge graph) and improve accuracy by optimizing key variables in the explainable structure (e.g., node importance and relevance) Main idea 2.
To improve model accuracy, they propose an attentive multi-view learning...
