Aggregate signature schema to aggregate messages in child block of B1

Could you please elaborate about this text from Lisk-BFT scientific paper: “ Using an aggregate signature scheme, these messages could be efficiently aggregated in the network and only added to any child block of Bl by the next proposer. Applying the properties of the general consensus framework analogous to the case of the Lisk-Bft protocol, it is possible to show that a block can be finalized already after two additional blocks are added as descendants.`

Researching about it I realized that exists aggregate signature schema that utilize the public key from a set like all top 101 block proposers requiring only 1 private key from such set of proposers, and the validation of such signature could be done in polynomial time. However, I would like to understand how it would work, even it superficially, and how it would allowing a block be finalized after two additional blocks added as descendants.