Preprint
Oct 2026
End-to-End Guarantees for Masked Diffusion Models under Conditional Mixing
The empirical risk minimizer trained under any masking distribution attains per-token excess $\varepsilon'$ on that distribution from $\varepsilon'^{-\mathcal{O}(\tau_c\log|\mathcal{V}|)}$ training sequences, independent of sequence length $N.
F. Sheriff
· 0 citations