English
Related papers

Related papers: Depth-Aware Initialization for Stable and Efficien…

200 papers

Successfully training deep neural networks often requires either batch normalization, appropriate weight initialization, both of which come with their own challenges. We propose an alternative, geometrically motivated method for training.…

Machine Learning · Computer Science 2019-10-08 Aram-Alexandre Pooladian , Chris Finlay , Adam M Oberman

Theoretical and empirical evidence indicates that the depth of neural networks is crucial for their success. However, training becomes more difficult as depth increases, and training of very deep networks remains an open problem. Here we…

Machine Learning · Computer Science 2015-11-24 Rupesh Kumar Srivastava , Klaus Greff , Jürgen Schmidhuber

Despite Deep Learning's (DL) empirical success, our theoretical understanding of its efficacy remains limited. One notable paradox is that while conventional wisdom discourages perfect data fitting, deep neural networks are designed to do…

Machine Learning · Computer Science 2024-02-06 Oria Gruber , Haim Avron

We introduce the concept of dynamically growing a neural network during training. In particular, an untrainable deep network starts as a trainable shallow network and newly added layers are slowly, organically added during training, thereby…

Neural and Evolutionary Computing · Computer Science 2015-11-24 Leslie N. Smith , Emily M. Hand , Timothy Doster

In recent years, newly developed methods to train spiking neural networks (SNNs) have rendered them as a plausible alternative to Artificial Neural Networks (ANNs) in terms of accuracy, while at the same time being much more energy…

Neural and Evolutionary Computing · Computer Science 2023-05-17 Nicolas Perez-Nieves , Dan F. M Goodman

In this work, we propose a multi-stage training strategy for the development of deep learning algorithms applied to problems with multiscale features. Each stage of the pro-posed strategy shares an (almost) identical network structure and…

Numerical Analysis · Mathematics 2020-09-25 Eric Chung , Wing Tat Leung , Sai-Mang Pun , Zecheng Zhang

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

Before training a neural net, a classic rule of thumb is to randomly initialize the weights so the variance of activations is preserved across layers. This is traditionally interpreted using the total variance due to randomness in both…

Machine Learning · Computer Science 2019-08-07 Kyle Luther , H. Sebastian Seung

Predictable adaptation of network depths can be an effective way to control inference latency and meet the resource condition of various devices. However, previous adaptive depth networks do not provide general principles and a formal…

Computer Vision and Pattern Recognition · Computer Science 2024-10-01 Woochul Kang , Hyungseop Lee

During the last decade, several research works have focused on providing novel deep learning methods in many application fields. However, few of them have investigated the weight initialization process for deep learning, although its…

Machine Learning · Computer Science 2021-02-16 Wadii Boulila , Maha Driss , Mohamed Al-Sarem , Faisal Saeed , Moez Krichen

Deep neural networks (DNNs) are often trained on the premise that the complete training data set is provided ahead of time. However, in real-world scenarios, data often arrive in chunks over time. This leads to important considerations…

Machine Learning · Computer Science 2023-03-21 Vijaya Raghavan T. Ramkumar , Elahe Arani , Bahram Zonooz

As overparameterized models become increasingly prevalent, training loss alone offers limited insight into generalization performance. While smoothness has been linked to improved generalization across various settings, directly enforcing…

Machine Learning · Computer Science 2025-09-30 Yifan Hao , Yanxin Lu , Hanning Zhang , Xinwei Shen , Tong Zhang

Deep equilibrium networks (DEQs) are a promising way to construct models which trade off memory for compute. However, theoretical understanding of these models is still lacking compared to traditional networks, in part because of the…

Machine Learning · Computer Science 2022-07-20 Atish Agarwala , Samuel S. Schoenholz

Weight initialization remains decisive for neural network optimization, yet existing methods are largely layer-agnostic. We study initialization for deeply-supervised architectures with auxiliary classifiers, where untrained auxiliary heads…

Machine Learning · Computer Science 2026-01-06 Hyunjun Kim

Memory-aware network scheduling is becoming increasingly important for deep neural network (DNN) inference on resource-constrained devices. However, due to the complex cell-level and network-level topologies, memory-aware scheduling becomes…

Machine Learning · Computer Science 2023-08-29 Shuzhang Zhong , Meng Li , Yun Liang , Runsheng Wang , Ru Huang

Proper weight initialization prior to training has historically been one of the key factors that helped kick off the deep learning revolution. Initialization is even more crucial in "reservoir computing", where the weights of a readout…

Machine Learning · Computer Science 2026-05-12 Tommaso Fioratti , Riccardo Marcaccioli , Francesco Casola

Existing methods for estimating uncertainty in deep learning tend to require multiple forward passes, making them unsuitable for applications where computational resources are limited. To solve this, we perform probabilistic reasoning over…

Machine Learning · Statistics 2020-12-08 Javier Antorán , James Urquhart Allingham , José Miguel Hernández-Lobato

Initialising the synaptic weights of artificial neural networks (ANNs) with orthogonal matrices is known to alleviate vanishing and exploding gradient problems. A major objection against such initialisation schemes is that they are deemed…

Neural and Evolutionary Computing · Computer Science 2023-03-23 Nikolay Manchev , Michael Spratling

Researchers have proposed various activation functions. These activation functions help the deep network to learn non-linear behavior with a significant effect on training dynamics and task performance. The performance of these activations…

Computer Vision and Pattern Recognition · Computer Science 2020-01-07 Pravendra Singh , Munender Varshney , Vinay P. Namboodiri

Appropriate weight initialization has been of key importance to successfully train neural networks. Recently, batch normalization has diminished the role of weight initialization by simply normalizing each layer based on batch statistics.…

Computer Vision and Pattern Recognition · Computer Science 2022-08-03 Pedro Hermosilla , Michael Schelling , Tobias Ritschel , Timo Ropinski