Search papers, labs, and topics across Lattice.
1
0
3
Training CBIR models on original, non-anonymized data yields more robust retrieval performance even after anonymization, challenging the assumption that anonymization-aware training is always necessary.
