Feature Request: Add Support for the BetaE Model to predict uncertainty

[PeterPirog]

Are there any plans to add the BetaE model to DGL-KE in the future ? This model use Beta distribution which gives great capabilities to make knowledge graphs embeddings.

This model can:

1. predict uncertainty of the nodes,
2. model symmetry,
3. model asymmetry,
4. model inversion,
5. model relation 1 to N,
6. support complex queries with complete FOL Logic (negation, conjunction, disjunction).

https://snap.stanford.edu/betae/
Github source: https://github.com/snap-stanford/KGReasoning