Fixed-clock optimizer memory turns equal-multiset data shuffle order into an O(η) source of fine-tuning noise, larger than the O(η²) effect in memoryless cases, with a fit-free sizing method derived.
arXiv preprint arXiv:2501.15556 (2025)
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
fields
cs.LG 3verdicts
UNVERDICTED 3representative citing papers
Lie brackets of gradient fields provide a geometric score for optimal transfer order in sequential learning, enabling efficient ranking of many sources via tournament methods.
An RL agent learns domain re-weighting policies from evaluation feedback to improve balanced performance in continual pre-training of LLMs across source and target domains.
citing papers explorer
-
Optimizer Memory Makes Shuffle Order a First-Order Source of Fine-Tuning Noise
Fixed-clock optimizer memory turns equal-multiset data shuffle order into an O(η) source of fine-tuning noise, larger than the O(η²) effect in memoryless cases, with a fit-free sizing method derived.
-
The Geometry of Sequential Learning: Lie-Bracket Prediction of Transfer Order
Lie brackets of gradient fields provide a geometric score for optimal transfer order in sequential learning, enabling efficient ranking of many sources via tournament methods.
-
Data Mixing Agent: Learning to Re-weight Domains for Continual Pre-training
An RL agent learns domain re-weighting policies from evaluation feedback to improve balanced performance in continual pre-training of LLMs across source and target domains.