DP-Muon adapts matrix-orthogonalized momentum optimization to differential privacy via per-matrix clipping and noise addition, with proofs of inherited privacy and optimization guarantees plus a bias-corrected version that improves private fine-tuning utility.
Title resolution pending
4 Pith papers cite this work. Polarity classification is still indexing.
4
Pith papers citing it
fields
cs.LG 4years
2026 4verdicts
UNVERDICTED 4representative citing papers
CCA does not compose autoregressively and retrofitting requires exponential query complexity under weak optimality.
Replaces determinant growth with generalized Rayleigh quotient for rare switching in private linear bandits to control worst-direction volume despite non-monotonic design matrices from noise.
Post-processing via random selection or linear combination of differentially private models allows meeting arbitrary target privacy parameters without additional training.
citing papers explorer
-
DP-Muon: Differentially Private Optimization via Matrix-Orthogonalized Momentum
DP-Muon adapts matrix-orthogonalized momentum optimization to differential privacy via per-matrix clipping and noise addition, with proofs of inherited privacy and optimization guarantees plus a bias-corrected version that improves private fine-tuning utility.
-
Barriers to Counterfactual Credit Attribution for Autoregressive Models
CCA does not compose autoregressively and retrofitting requires exponential query complexity under weak optimality.
-
When Determinants Are Not Enough: Private Rare Switching
Replaces determinant growth with generalized Rayleigh quotient for rare switching in private linear bandits to control worst-direction volume despite non-monotonic design matrices from noise.
-
Differentially Private Model Merging
Post-processing via random selection or linear combination of differentially private models allows meeting arbitrary target privacy parameters without additional training.