Search papers, labs, and topics across Lattice.
University of Sydney
1
0
1
Closing the gap in multiclass PAC learning reveals that optimal excess risk can be achieved without prior knowledge of oracle risk, fundamentally shifting our understanding of classifier performance.
