English
Related papers

Related papers: Weight decay induces low-rank attention layers

200 papers

Regularization is typically understood as improving generalization by altering the landscape of local extrema to which the model eventually converges. Deep neural networks (DNNs), however, challenge this view: We show that removing…

Machine Learning · Computer Science 2019-06-03 Aditya Golatkar , Alessandro Achille , Stefano Soatto

This paper introduces novel theoretical approximation bounds for the output of quantized neural networks, with a focus on convolutional neural networks (CNN). By considering layerwise parametrization and focusing on the quantization of…

Information Theory · Computer Science 2025-10-13 Samy Houache , Jean François Aujol , Yann Traonmilin

Low-rank approximations, of the weight and feature space can enhance the performance of deep learning models, whether in terms of improving generalization or reducing the latency of inference. However, there is no clear consensus yet on…

Computation and Language · Computer Science 2024-05-24 Arnav Chavan , Nahush Lele , Deepak Gupta

Re-initializing a neural network during training has been observed to improve generalization in recent works. Yet it is neither widely adopted in deep learning practice nor is it often used in state-of-the-art training protocols. This…

Machine Learning · Computer Science 2023-04-04 Sheheryar Zaidi , Tudor Berariu , Hyunjik Kim , Jörg Bornschein , Claudia Clopath , Yee Whye Teh , Razvan Pascanu

How does scaling the number of parameters in large language models (LLMs) affect their core capabilities? We study two natural scaling techniques -- weight pruning and simply training a smaller or larger model, which we refer to as dense…

Computation and Language · Computer Science 2023-10-10 Tian Jin , Nolan Clement , Xin Dong , Vaishnavh Nagarajan , Michael Carbin , Jonathan Ragan-Kelley , Gintare Karolina Dziugaite

Regularization is a set of techniques that are used to improve the generalization ability of deep neural networks. In this paper, we introduce weight compander (WC), a novel effective method to improve generalization by reparameterizing…

Machine Learning · Computer Science 2023-06-30 Rinor Cakaj , Jens Mehnert , Bin Yang

We investigate the inherent bias of Stochastic Gradient Descent (SGD) toward learning low-rank weight matrices during the training of deep neural networks. Our results demonstrate that training with mini-batch SGD and weight decay induces a…

Machine Learning · Computer Science 2024-10-22 Tomer Galanti , Zachary S. Siegel , Aparna Gupte , Tomaso Poggio

In this paper, we first identify activation shift, a simple but remarkable phenomenon in a neural network in which the preactivation value of a neuron has non-zero mean that depends on the angle between the weight vector of the neuron and…

Neural and Evolutionary Computing · Computer Science 2024-03-22 Takuro Kutsuna

Deep Reinforcement Learning (Deep RL) has been receiving increasingly more attention thanks to its encouraging performance on a variety of control tasks. Yet, conventional regularization techniques in training neural networks (e.g., $L_2$…

Machine Learning · Computer Science 2021-11-30 Zhuang Liu , Xuanlin Li , Bingyi Kang , Trevor Darrell

Large Language Models (LLMs) exhibit strong general language capabilities. However, fine-tuning these models on domain-specific tasks often leads to catastrophic forgetting, where the model overwrites or loses essential knowledge acquired…

Computation and Language · Computer Science 2025-02-18 Shezheng Song , Hao Xu , Jun Ma , Shasha Li , Long Peng , Qian Wan , Xiaodong Liu , Jie Yu

Weight decay is a widely used technique for training Deep Neural Networks(DNN). It greatly affects generalization performance but the underlying mechanisms are not fully understood. Recent works show that for layers followed by…

Machine Learning · Computer Science 2021-03-30 Yucong Zhou , Yunxiao Sun , Zhao Zhong

Large language models (LLMs) deliver strong performance, but their high compute and memory costs make deployment difficult in resource-constrained scenarios. Weight-only post-training quantization (PTQ) is appealing, as it reduces memory…

Machine Learning · Computer Science 2026-02-09 Xianglong Yan , ChengZhu Bao , Zhiteng Li , Tianao Zhang , Shaoqiu Zhang , Ruobing Xie , Samm Sun , Yulun Zhang

Regularization techniques such as $\mathcal{L}_1$ and $\mathcal{L}_2$ regularizers are effective in sparsifying neural networks (NNs). However, to remove a certain neuron or channel in NNs, all weight elements related to that neuron or…

Machine Learning · Computer Science 2023-05-31 Ali Haisam Muhammad Rafid , Adrian Sandu

L2 regularization for weights in neural networks is widely used as a standard training trick. However, L2 regularization for gamma, a trainable parameter of batch normalization, remains an undiscussed mystery and is applied in different…

Computer Vision and Pattern Recognition · Computer Science 2022-05-17 Bum Jun Kim , Hyeyeon Choi , Hyeonah Jang , Dong Gu Lee , Wonseok Jeong , Sang Woo Kim

Weight-sharing is ubiquitous in deep learning. Motivated by this, we propose a "weight-sharing regularization" penalty on the weights $w \in \mathbb{R}^d$ of a neural network, defined as $\mathcal{R}(w) = \frac{1}{d - 1}\sum_{i > j}^d |w_i…

Machine Learning · Computer Science 2024-03-12 Mehran Shakerinava , Motahareh Sohrabi , Siamak Ravanbakhsh , Simon Lacoste-Julien

We provide a function space characterization of the inductive bias resulting from minimizing the $\ell_2$ norm of the weights in multi-channel convolutional neural networks with linear activations and empirically test our resulting…

Machine Learning · Computer Science 2022-07-12 Meena Jagadeesan , Ilya Razenshteyn , Suriya Gunasekar

The large capacity of neural networks enables them to learn complex functions. To avoid overfitting, networks however require a lot of training data that can be expensive and time-consuming to collect. A common practical approach to…

Machine Learning · Computer Science 2020-03-10 Majed El Helou , Frederike Dümbgen , Sabine Süsstrunk

We show the formal equivalence of linearised self-attention mechanisms and fast weight controllers from the early '90s, where a ``slow" neural net learns by gradient descent to program the ``fast weights" of another net through sequences of…

Machine Learning · Computer Science 2021-06-10 Imanol Schlag , Kazuki Irie , Jürgen Schmidhuber

As the context window expands, self-attention increasingly dominates the transformer's inference time. Therefore, accelerating attention computation while minimizing performance degradation is essential for the efficient deployment of Large…

Computation and Language · Computer Science 2025-03-14 Eli Sason , Darya Frolova , Boris Nazarov , Felix Goldberd

Over-parameterized deep networks trained using gradient-based optimizers are a popular choice for solving classification and ranking problems. Without appropriately tuned $\ell_2$ regularization or weight decay, such networks have the…

Machine Learning · Computer Science 2021-08-13 Aman Gupta , Rohan Ramanath , Jun Shi , Anika Ramachandran , Sirou Zhou , Mingzhou Zhou , S. Sathiya Keerthi