High-quality gene/disease embedding in a multi-relational heterogeneous graph after a joint matrix/tensor decomposition

The framework of the gene/disease embedding computation in a multi-relational heterogeneous graph. (The left part shows the heterogeneous data representation both for (gene/disease, uni-relation, other entity) and (gene, multi-relation, disease) triples. The middle part depicts the data structure of matrix and tensor for uni- or multi-relational triple. Meanwhile, in the right part, a joint decomposition of the heterogeneous matrix/tensor data is illustrated.)