Federated personalization of foundation models creates hard-to-detect trustworthiness failures due to privacy constraints, and existing benchmarks cannot adequately evaluate them.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Silent Failures in Federated Personalization of Foundation Models
Federated personalization of foundation models creates hard-to-detect trustworthiness failures due to privacy constraints, and existing benchmarks cannot adequately evaluate them.