Sub-n-gram estimators are near-stationary points of the population cross-entropy loss for in-context n-gram learning, offering a theoretical explanation for stage-wise training plateaus.
Algorithmic Regularization in Model-free Overparametrized Asymmetric Matrix Factorization
1 Pith paper cite this work. Polarity classification is still indexing.
abstract
We study the asymmetric matrix factorization problem under a natural nonconvex formulation with arbitrary overparametrization. The model-free setting is considered, with minimal assumption on the rank or singular values of the observed matrix, where the global optima provably overfit. We show that vanilla gradient descent with small random initialization sequentially recovers the principal components of the observed matrix. Consequently, when equipped with proper early stopping, gradient descent produces the best low-rank approximation of the observed matrix without explicit regularization. We provide a sharp characterization of the relationship between the approximation error, iteration complexity, initialization size and stepsize. Our complexity bound is almost dimension-free and depends logarithmically on the approximation error, with significantly more lenient requirements on the stepsize and initialization compared to prior work. Our theoretical results provide accurate prediction for the behavior gradient descent, showing good agreement with numerical experiments.
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Learning In-context n-grams with Transformers: Sub-n-grams Are Near-stationary Points
Sub-n-gram estimators are near-stationary points of the population cross-entropy loss for in-context n-gram learning, offering a theoretical explanation for stage-wise training plateaus.