Reconciliation (as exemplified by OpenRefine) is "one to many": one tabular row is matched against many entities in a dataset or KG. (For efficiency, the protocol sends a batch of rows, but each one is treated separately.)
Entity matching (as a merging step needed in KG building) is "many to many": potentially ALL entities of the source dataset are considered, and entities in either the source or target dataset may be merged.
Is the scope of this group limited to Reconciliation only, or would it also include Matching?