Deterministic versus stochastic dynamical classifiers: opposing random adversarial attacks with noise

doi:10.1088/2632-2153/ae0244

2025-09-30

PRE

DOI：10.1088/2632-2153/ae0244

原文链接

原文求助

摘要

En 中文

连续变量发放率（CVFR）模型在神经科学中被广泛用于描述兴奋性生物神经元复杂的动力学特性，此处被训练和测试为一种动力学分类器。为此，通过其谱分解，模型被赋予一组先验嵌入在节点耦合矩阵中的吸引子。学习过程旨在调整剩余参数，以塑造一条非平衡路径，连接输入（待分类数据）和输出（目标记忆槽）。所施加的吸引子在训练过程中保持不变，这使得事后比较最终得以进行，例如关于它们相关吸引盆的大小。连续变量发放率（CVFR）模型的一个随机变体也被研究，并发现其对非目标对抗攻击具有鲁棒性，这些攻击通过随机扰动破坏待最终分类的项目。总体而言，本研究表明，一个基于耦合常微分方程（ODEs）编写的生物学上合理的模型家族能够有效应对非平凡的分类任务。

Keyword:

neural networks

dynamical systems

continuous-time Hopfield model

classification task

stochastic processes

期刊

Machine Learning-Science and Technology

IF:

4.6

论文数:

1.1K

被引数:

3.4K

机构

istituto nazionale di fisica nucleare (infn)

学者数:

3.0W

论文数: 1.2W

被引数: 14

university of florence

学者数:

4.2W

论文数: 3.1W

被引数: 42

引用论文

Searching for exotic particles in high-energy physics with deep learning 用深度学习寻找高能物理中的奇异粒子

NATURE COMMUNICATIONS

IF15.7

2014-07-02

836

PREAI

Baldi, P.; Sadowski, P.; Whiteson, D.

Long Short-Term Memory 长短期记忆

Neural Computation

IF0

1997-11-01

PREAI

Sepp Hochreiter; Jürgen Schmidhuber

Stable attractors for neural networks classification via ordinary differential equations (SA-nODE)通过常微分方程 (SA节点) 进行神经网络分类的稳定吸引子

MACHINE LEARNING-SCIENCE AND TECHNOLOGY

IF4.6

2024-10-01

OAAI

Marino, Raffaele; Buffoni, Lorenzo; Chicchi, Lorenzo; Giambagli, Lorenzo; Fanelli, Duccio

Dense Associative Memory Is Robust to Adversarial Inputs

Neural Computation

IF0

2018-12-01

OAAI

Dmitry Krotov; John Hopfield

Liquid Time-constant Networks

Proceedings of the AAAI Conference on Artificial Intelligence

IF0

2021-05-18

OAAI

Ramin Hasani; Mathias Lechner; Alexander Amini; Daniela Rus; Radu Grosu

Highly accurate protein structure prediction with AlphaFold 基于AlphaFold的高精度蛋白质结构预测

Nature

IF0

2021-07-15

OAAI

John Jumper; Richard Evans; Alexander Pritzel; Tim Green; Michael Figurnov; Olaf Ronneberger; Kathryn Tunyasuvunakool; Russ Bates; Augustin Žídek; Anna Potapenko; Alex Bridgland; Clemens Meyer; Simon A. A. Kohl; Andrew J. Ballard; Andrew Cowie; Bernardino Romera-Paredes; Stanislav Nikolov; Rishub Jain; Jonas Adler; Trevor Back; Stig Petersen; David Reiman; Ellen Clancy; Michal Zielinski; Martin Steinegger; Michalina Pacholska; Tamas Berghammer; Sebastian Bodenstein; David Silver; Oriol Vinyals; Andrew W. Senior; Koray Kavukcuoglu; Pushmeet Kohli; Demis Hassabis

The MNIST Database of Handwritten Digit Images for Machine Learning Research [Best of the Web]用于机器学习研究的手写数字图像MNIST数据库 [Best of The Web]

IEEE Signal Processing Magazine

IF0

2012-11-01

PREAI

Li Deng

Frontiers of thermobarometry: GAIA, a novel Deep Learning-based tool for volcano plumbing systems

Earth and Planetary Science Letters

IF0

2023-10-01

PREAI

Lorenzo Chicchi; Luca Bindi; Duccio Fanelli; Simone Tommasini