Cross-Loss Influence Functions to Explain Deep Network Representations.

Andrew Silva, Rohit Chopra, Matthew C. Gombolay

Browse the full AISTATS paper archive.