Peter Slater Piazza's Highlights on 'Learning heterogeneous graph embedding for Chinese legal document similarity' | Glasp