Pith. sign in

REVIEW 7 cited by

The Benefits of Reusing Batches for Gradient Descent in Two-Layer Networks: Breaking the Curse of Information and Leap Exponents

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2402.03220 v3 pith:BZGJMBWN submitted 2024-02-05 stat.ML cs.LG

classification stat.MLcs.LG
keywords functionsgradientbatchesdescenttargetabbedynamicalexponent
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
read the original abstract

We investigate the training dynamics of two-layer neural networks when learning multi-index target functions. We focus on multi-pass gradient descent (GD) that reuses the batches multiple times and show that it significantly changes the conclusion about which functions are learnable compared to single-pass gradient descent. In particular, multi-pass GD with finite stepsize is found to overcome the limitations of gradient flow and single-pass GD given by the information exponent (Ben Arous et al., 2021) and leap exponent (Abbe et al., 2023) of the target function. We show that upon re-using batches, the network achieves in just two time steps an overlap with the target subspace even for functions not satisfying the staircase property (Abbe et al., 2021). We characterize the (broad) class of functions efficiently learned in finite time. The proof of our results is based on the analysis of the Dynamical Mean-Field Theory (DMFT). We further provide a closed-form description of the dynamical process of the low-dimensional projections of the weights, and numerical experiments illustrating the theory.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 7 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Online Learning of Neural Networks

    stat.ML 2025-05 conditional novelty 8.0 of 10

    Online mistake bounds for sign-activation feedforward networks are governed by a totally-separable packing number of the first layer margin, with dimension-free bounds under multi-index or all-layer margin assumptions.

  2. Approximate Message Passing with Random Initialization for Phase Retrieval

    math.ST 2026-08 conditional novelty 7.0 of 10

    Randomly initialized Bayes-optimal AMP provably achieves the weak-recovery threshold δ=1/2 and arbitrarily accurate recovery for δ>1.13 in proportional-regime noiseless phase retrieval.

  3. Low-dimensional Functions are Efficiently Learnable under Randomly Biased Distributions

    cs.LG 2025-02 conditional novelty 7.0 of 10

    A random shift of Gaussian inputs forces the first Hermite coefficient of any non-linear target to be large, yielding near-linear sample complexity independent of the target's information exponent, and a similar resul...

  4. Gradient dynamics for low-rank fine-tuning beyond kernels

    cs.LG 2024-11 accept novelty 7.0 of 10

    In a student-teacher model with Gaussian inputs and a rank-1 teacher perturbation, online SGD converges to the teacher in d k^{O(1)} iterations, independent of the activation's Hermite information exponent.

  5. Dropout and Random Gradient Masking Are Asymptotically Equivalent in Large ResNets

    stat.ML 2026-07 conditional novelty 6.0 of 10

    As ResNets grow deep and wide with fixed dropout rate, dropout training and random-gradient-masking training converge to the same limiting dynamics, and the common masking variants collapse to one limit.

  6. Learning Hierarchical Polynomials of Multiple Nonlinear Features with Three-Layer Networks

    cs.LG 2024-11 conditional novelty 6.0 of 10

    A three-layer network with layerwise gradient descent provably recovers the span of multiple quadratic features in O~(d^4) samples and then learns any polynomial link in the features.

  7. Learning Gaussian Multi-Index Models with Gradient Flow: Time Complexity and Directional Convergence

    cs.LG 2024-11 conditional novelty 6.0 of 10

    For orthogonal hidden directions, gradient flow provably sends each neuron to the nearest direction and a log-factor overparameterization suffices, but for equiangular directions with overlap above beta_c = (p*-2)/(k+...

Pith tools