arXiv:2606. 23942v1 Announce Type: new Abstract: We present a large-scale empirical study isolating the contributions of the Derivative Regularization penalty (DREG).
arXiv:2610.00751v1 Announce Type: cross Abstract: Recent theoretical work identified fundamental properties of representation geometry that shape inference ability of deep neural networks. These incl...
arXiv:2606. 27321v1 Announce Type: cross Abstract: Sparse autoencoders (SAEs) have become a leading tool for interpreting the representations of vision foundation models, decomposing their polysemantic activations into a larger set of sparse, more monosemantic features.
arXiv:2606. 07908v1 Announce Type: new Abstract: Derivative-controlled networks based on ChainzRule (CR) combine cubic polynomial layers with a lightweight forward-mode per-layer Jacobian penalty (DREG).
arXiv:2607. 13380v1 Announce Type: new Abstract: Predictive Coding (PC) offers a biologically motivated alternative to backpropagation via local weight updates, yet routing error between layers still relies on an autograd Jacobian-transpose ($J^\top$) product - the last non-local operation in PC.
The paper investigates why diffusion models, unlike typical deep learning models, exhibit catastrophic overfitting when overparameterized. Through experiments on U‑Nets trained on CelebA and a random‑features theoretical analysis, it shows that the interpolation peak occurs at a model size proportional to the product of training samples and noise realizations, but the test loss starts to rise already at the number of samples, leading to memorization of the empirical score. Regularization techniques such as ridge penalties or early stopping can still make large models outperform smaller, unregularized ones.
We investigate how each component of the Transformer feedforward block architecture design determines how much rank survives across depth at initialization. We reinterpret skip connections and normalization, long understood as controlling magnitude, as mechanisms for preserving gradient rank across depth, since the very matrix multiplications and nonlinear activations that make the network expressive also reduce the rank.
Conventional wisdom in deep learning holds that overparameterization---having more parameters $p$ than training samples $n$---is benign: larger models generalize better and, even without regularizatio...
arXiv:2606. 06888v1 Announce Type: new Abstract: Classical scaling laws for language model pretraining balance model size against training dataset size under a fixed compute budget, assuming abundant data and a single pass over the corpus.
arXiv:2607. 14466v1 Announce Type: new Abstract: Noise injection is a well-known technique in stochastic optimization.
arXiv:2607. 08754v1 Announce Type: cross Abstract: Low-rank factorization is widely used to compress neural networks, but modern models are often not naturally amenable to aggressive factorization without significant accuracy loss.
arXiv:2607. 14018v1 Announce Type: cross Abstract: We investigate how each component of the Transformer feedforward block architecture design determines how much rank survives across depth at initialization.
The paper introduces Sparsity-Adaptive Sharpness-Aware Minimization (SA‑SAM), a method that adjusts the perturbation radius in sharpness-aware training to remain consistent as model sparsity increases. It also evaluates a Magnitude‑Weighted Hessian (MWH) importance metric derived from second‑order analysis. Experiments on CIFAR‑10‑C, CIFAR‑100‑C, and ImageNet‑100‑C show that SA‑SAM improves corruption robustness at 80–90% sparsity while maintaining clean accuracy, and the study reports inference throughput at deployment‑relevant sparsity levels.