Search papers, labs, and topics across Lattice.
Tevian Moscow
1
0
2
Pairwise ranking metrics can mislead model evaluation, with high scores not guaranteeing correct object assignments in multi-view tasks.