Stochastic gradient ascent with averaging learns Lagrangian multipliers for MILP at the minimax rate Θ(s/√N) and faster Θ(s/N) for warm-start, closing the gap between upper and lower bounds.
Title resolution pending
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
years
2026 3verdicts
UNVERDICTED 3representative citing papers
For binary classification in the NTK regime, LoRA rank r=1 suffices and is often optimal under cross-entropy loss, reducing the prior sufficient condition from r>=12.
Near-linear time algorithm for robust regression under Gaussian covariates achieves O(sqrt(ε κ)) error with Õ(d/ε⁴) samples when ε κ ≲ 1, plus SQ and low-degree lower bounds.
citing papers explorer
-
Provably Data-driven Lagrangian Relaxation for Mixed Integer Linear Programming
Stochastic gradient ascent with averaging learns Lagrangian multipliers for MILP at the minimax rate Θ(s/√N) and faster Θ(s/N) for warm-start, closing the gap between upper and lower bounds.
-
Rethinking the Rank Threshold for LoRA Fine-Tuning
For binary classification in the NTK regime, LoRA rank r=1 suffices and is often optimal under cross-entropy loss, reducing the prior sufficient condition from r>=12.
-
On efficient robust regression with subquadratic samples
Near-linear time algorithm for robust regression under Gaussian covariates achieves O(sqrt(ε κ)) error with Õ(d/ε⁴) samples when ε κ ≲ 1, plus SQ and low-degree lower bounds.