1 paper
Sonia Horchidan, Fabian Zeiher, Xiangyu Shi +4
Querying incomplete knowledge graphs with neural predictors is powerful but dangerous. Errors compound across multi-hop pipelines with no formal bound on the completeness of result…