中文
相关论文

相关论文: How the Optimizer Shapes Learned Solutions in Equi…

200 篇论文

Designing machine learning architectures for processing neural networks in their raw weight matrix form is a newly introduced research direction. Unfortunately, the unique symmetry structure of deep weight spaces makes this design very…

机器学习 · 计算机科学 2023-06-02 Aviv Navon , Aviv Shamsian , Idan Achituve , Ethan Fetaya , Gal Chechik , Haggai Maron

Visual imitation learning with 3D point clouds has advanced robotic manipulation by providing geometry-aware, appearance-invariant observations. However, point cloud-based policies remain highly sensitive to sensor noise, pose…

机器人学 · 计算机科学 2026-01-27 Zhiyuan Zhang , Yu She

The rapid scaling of large language models (LLMs) has made low-precision training essential for reducing memory, improving efficiency, and enabling larger models and datasets. Existing convergence theories for adaptive optimizers, however,…

机器学习 · 计算机科学 2026-03-03 Xuan Tang , Jichu Li , Difan Zou

Deep neural networks are traditionally trained using human-designed stochastic optimization algorithms, such as SGD and Adam. Recently, the approach of learning to optimize network parameters has emerged as a promising research topic.…

机器学习 · 计算机科学 2018-11-26 Shipeng Wang , Jian Sun , Zongben Xu

To overcome the restriction of identical distribution assumption, invariant representation learning for unsupervised domain adaptation (UDA) has made significant advances in computer vision and pattern recognition communities. In UDA…

计算机视觉与模式识别 · 计算机科学 2024-07-16 You-Wei Luo , Chuan-Xian Ren , Xiao-Lin Xu , Qingshan Liu

A core issue with learning to optimize neural networks has been the lack of generalization to real world problems. To address this, we describe a system designed from a generalization-first perspective, learning to update optimizer…

机器学习 · 计算机科学 2021-06-09 Diogo Almeida , Clemens Winter , Jie Tang , Wojciech Zaremba

Many successful deep learning architectures are equivariant to certain transformations in order to conserve parameters and improve generalization: most famously, convolution layers are equivariant to shifts of the input. This approach only…

机器学习 · 计算机科学 2021-03-31 Allan Zhou , Tom Knowles , Chelsea Finn

In this paper, we elucidate how representations in deep neural networks (DNNs) evolve during training. Our focus is on overparameterized learning settings where the training continues much after the trained DNN starts to perfectly fit its…

机器学习 · 计算机科学 2025-02-04 Yuval Sharon , Yehuda Dar

Descent methods for deep networks are notoriously capricious: they require careful tuning of step size, momentum and weight decay, and which method will work best on a new benchmark is a priori unclear. To address this problem, this paper…

神经与进化计算 · 计算机科学 2021-09-21 Yang Liu , Jeremy Bernstein , Markus Meister , Yisong Yue

Composite materials with different microstructural material symmetries are common in engineering applications where grain structure, alloying and particle/fiber packing are optimized via controlled manufacturing. In fact these…

材料科学 · 物理学 2024-04-30 Ravi Patel , Cosmin Safta , Reese E. Jones

Equivariant neural networks enforce symmetry within the structure of their convolutional layers, resulting in a substantial improvement in sample efficiency when learning an equivariant or invariant function. Such models are applicable to…

机器人学 · 计算机科学 2022-03-10 Dian Wang , Robin Walters , Robert Platt

Optimization is an integral part of modern deep learning. Recently, the concept of learned optimizers has emerged as a way to accelerate this optimization process by replacing traditional, hand-crafted algorithms with meta-learned…

机器学习 · 计算机科学 2023-12-13 Jan Sobotka , Petr Šimánek , Daniel Vašata

Adversarial examples reveal critical vulnerabilities in deep neural networks by exploiting their sensitivity to imperceptible input perturbations. While adversarial training remains the predominant defense strategy, it often incurs…

机器学习 · 计算机科学 2025-11-04 Longwei Wang , Ifrat Ikhtear Uddin , KC Santosh , Chaowei Zhang , Xiao Qin , Yang Zhou

Equivariant deep learning architectures exploit symmetries in learning problems to improve the sample efficiency of neural-network-based models and their ability to generalise. However, when modelling real-world data, learning problems are…

机器学习 · 统计学 2024-11-12 Matthew Ashman , Cristiana Diaconu , Adrian Weller , Wessel Bruinsma , Richard E. Turner

Machine learning pipelines that include a combinatorial optimization layer can give surprisingly efficient heuristics for difficult combinatorial optimization problems. Three questions remain open: which architecture should be used, how…

机器人学 · 计算机科学 2023-02-07 Axel Parmentier

We survey the mathematical foundations of geometric deep learning, focusing on group equivariant and gauge equivariant neural networks. We develop gauge equivariant convolutional neural networks on arbitrary manifolds $\mathcal{M}$ using…

The choice of optimizer significantly impacts the training efficiency and computational costs of large language models (LLMs). Recently, the Muon optimizer has demonstrated promising results by orthogonalizing parameter updates, improving…

机器学习 · 计算机科学 2025-10-08 Zichong Li , Liming Liu , Chen Liang , Weizhu Chen , Tuo Zhao

The crucial role played by the underlying symmetries of high energy physics and lattice field theories calls for the implementation of such symmetries in the neural network architectures that are applied to the physical system under…

高能物理 - 格点 · 物理学 2022-02-16 Srinath Bulusu , Matteo Favoni , Andreas Ipp , David I. Müller , Daniel Schuh

We introduce Equivariant Isomorphic Networks (EquIN) -- a method for learning representations that are equivariant with respect to general group actions over data. Differently from existing equivariant representation learners, EquIN is…

The underspecification of most machine learning pipelines means that we cannot rely solely on validation performance to assess the robustness of deep learning systems to naturally occurring distribution shifts. Instead, making sure that a…