Compute-Optimal LLMs Provably Generalize Better With Scale

Why do larger language models generalize better? To investigate this question, we develop generalization bounds on the pretraining objective of large language models (LLMs) in the compute-optimal regime, as described by the Chinchilla scaling laws. We introduce a novel, fully empirical Freedman-type martingale concentration inequality that tightens existing bounds by accounting for the variance of the loss function. This generalization bound can be decomposed into three interpretable components: the number of parameters per token, the loss variance, and the quantization error at a fixed bitrate. As compute-optimal language models are scaled up, the number of parameters per data point remains constant; however, both the loss variance and the quantization error decrease, implying that larger models should have smaller generalization gaps. We examine why larger models tend to be more quantizable from an information theoretic perspective, showing that the rate at which they can integrate new information grows more slowly than their capacity on the compute-optimal frontier. From these findings we produce a scaling law for the generalization gap, with bounds that become predictably stronger with scale.

Partially Observed Trajectory Inference using Optimal Transport and a Dynamics Prior

Trajectory inference seeks to recover the temporal dynamics of a population from snapshots of its (uncoupled) temporal marginals, i.e. where observed particles are not tracked over time. Lavenant et al. arXiv:2102.09204 addressed this challenging problem under a stochastic differential equation (SDE) model with a gradient-driven drift in the observed space, introducing a minimum entropy estimator relative to the Wiener measure. Chizat et al. arXiv:2205.07146 then provided a practical grid-free mean-field Langevin (MFL) algorithm using Schr\"odinger bridges. Motivated by the overwhelming success of observable state space models in the traditional paired trajectory inference problem (e.g. target tracking), we extend the above framework to a class of latent SDEs in the form of observable state space models. In this setting, we use partial observations to infer trajectories in the latent space under a specified dynamics model (e.g. the constant velocity/acceleration models from target tracking). We introduce PO-MFL to solve this latent trajectory inference problem and provide theoretical guarantees by extending the results of arXiv:2102.09204 to the partially observed setting. We leverage the MFL framework of arXiv:2205.07146, yielding an algorithm based on entropic OT between dynamics-adjusted adjacent time marginals. Experiments validate the robustness of our method and the exponential convergence of the MFL dynamics, and demonstrate significant outperformance over the latent-free method of arXiv:2205.07146 in key scenarios.

k-Mixup Regularization for Deep Learning via Optimal Transport

Mixup is a popular regularization technique for training deep neural networks that can improve generalization and increase adversarial robustness. It perturbs input training data in the direction of other randomly-chosen instances in the training set. To better leverage the structure of the data, we extend mixup to \emph{$k$-mixup} by perturbing $k$-batches of training points in the direction of other $k$-batches using displacement interpolation, interpolation under the Wasserstein metric. We demonstrate theoretically and in simulations that $k$-mixup preserves cluster and manifold structures, and we extend theory studying efficacy of standard mixup. Our empirical results show that training with $k$-mixup further improves generalization and robustness on benchmark datasets.