Framework Overview (IMAGE)
Caption
The new method works in two phases: offline and online. In the offline phase, data graph embeddings are created and stored in a vector database with an index for fast search. In the online phase, streaming query graphs are embedded and matched through k-nearest neighbor (KNN) search. The inverted file (IVF) and product quantization (PQ) indexing methods speed up similarity search and improve scalability.
Credit
Liuyi Chen et al.
Usage Restrictions
Credit must be given to the creator.
License
CC BY