Jove
Visualize
联系我们
JoVE
x logofacebook logolinkedin logoyoutube logo
关于 JoVE
概览领导团队博客JoVE 帮助中心
作者
出版流程编辑委员会范围与政策同行评审常见问题投稿
图书馆员
用户评价订阅访问资源图书馆顾问委员会常见问题
研究
JoVE JournalMethods CollectionsJoVE Encyclopedia of Experiments存档
教育
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab Manual教师资源中心教师网站
使用条款与条件
隐私政策
政策

相关概念视频

Understanding Memory01:19

Understanding Memory

1.7K
Memory is the retention of information or experiences over time, facilitated through three main processes: encoding, storage, and retrieval. Encoding is the process of inputting information into the memory system. For instance, when listening to a lecture, watching a play, reading a book, or having a conversation, the brain is actively encoding information. This initial stage involves transforming sensory input into a form that can be processed and stored by the brain. Various factors, such as...
1.7K

您也可能阅读

相关文章

通过共同作者、期刊和引用图与本文相关的文章。

排序
Same author

Hybrid ferroelectric-ionic memristive hardware for high scalability in-memory computing.

Nature communications·2026
Same author

CMOS-compatible ferroelectric tunnel junctions integrate stochastic sampling and deterministic computing for image generation.

Nature communications·2026
Same author

Analysis of vertical AND flash memory for energy-efficient, scalable, fast CIM beyond vertical NAND flash memory.

Nano convergence·2026
Same author

Repurposing Si CMOS nonidealities for stochastic and analog image processing.

Science advances·2026
Same author

Noise Fingerprints as a Quantitative Order Parameter for Polarization- and Defect-Mediated Switching in Hafnia Ferroelectrics.

Advanced science (Weinheim, Baden-Wurttemberg, Germany)·2026
Same author

Gate insulator stack engineering for fully CMOS-compatible reservoir computing.

Nano convergence·2026

相关实验视频

Updated: Feb 28, 2026

Assembly and Characterization of Biomolecular Memristors Consisting of Ion Channel-doped Lipid Membranes
08:07

Assembly and Characterization of Biomolecular Memristors Consisting of Ion Channel-doped Lipid Membranes

Published on: March 9, 2019

8.4K

使用垂直NAND闪存的物理非克隆性神经网络.

Sung-Ho Park1, Ryun-Han Koo1, Jonghyun Ko1

  • 1Department of Electrical and Computer Engineering and Inter-university Semiconductor Research Center, Seoul National University, Seoul, Republic of Korea.

Advanced science (Weinheim, Baden-Wurttemberg, Germany)
|February 27, 2026
PubMed
概括

我们在闪存上开发了一个安全的神经网络,使其能够抵抗克隆和数据盗窃. 这种物理不可克隆的神经网络 (PUNN) 保护隐私关键应用中的敏感信息.

关键词:
在V-NAND闪存内存.神经网络的神经网络的神经网络安全的安全的安全的安全的安全.

更多相关视频

In Situ Transmission Electron Microscopy with Biasing and Fabrication of Asymmetric Crossbars Based on Mixed-Phased a-VOx
09:49

In Situ Transmission Electron Microscopy with Biasing and Fabrication of Asymmetric Crossbars Based on Mixed-Phased a-VOx

Published on: May 13, 2020

4.4K
Author Spotlight: Deciphering Memory and Learning Through Neural Implants for Multi-Region Brain Studies
08:51

Author Spotlight: Deciphering Memory and Learning Through Neural Implants for Multi-Region Brain Studies

Published on: April 26, 2024

2.0K

相关实验视频

Last Updated: Feb 28, 2026

Assembly and Characterization of Biomolecular Memristors Consisting of Ion Channel-doped Lipid Membranes
08:07

Assembly and Characterization of Biomolecular Memristors Consisting of Ion Channel-doped Lipid Membranes

Published on: March 9, 2019

8.4K
In Situ Transmission Electron Microscopy with Biasing and Fabrication of Asymmetric Crossbars Based on Mixed-Phased a-VOx
09:49

In Situ Transmission Electron Microscopy with Biasing and Fabrication of Asymmetric Crossbars Based on Mixed-Phased a-VOx

Published on: May 13, 2020

4.4K
Author Spotlight: Deciphering Memory and Learning Through Neural Implants for Multi-Region Brain Studies
08:51

Author Spotlight: Deciphering Memory and Learning Through Neural Implants for Multi-Region Brain Studies

Published on: April 26, 2024

2.0K

科学领域:

  • 计算机科学 计算机科学
  • 电气工程 电气工程
  • 密码学 密码学 密码学 密码学

背景情况:

  • 神经网络在敏感应用中越来越多地被使用,这需要加强数据和模型的保护.
  • 现有的架构容易受到模型倒置和克隆攻击,危及隐私和知识产权.

研究的目的:

  • 介绍一种在商业垂直NAND (V-NAND) 闪存上实现的新型反转电阻物理非克隆神经网络 (PUNN).
  • 为了证明硬件根基的安全性,防止模型克隆和模型倒置攻击,同时保持高精度.

主要方法:

  • 在V-NAND闪存中使用弱门诱导的排水泄漏清除实现了PUNN,以创建独特的,不可重复的设备级导电性模式.
  • 使用前向前向 (FF) 算法进行训练,与V-NAND的共同源线结构兼容,消除了向后传播.
  • 通过将训练重量转移到芯片之间来评估模型非克隆性,并在MIT-BIH心电图数据集上评估隐私保护.

主要成果:

  • 由于固有的随机性,V-NAND FF-PUNN表现出硬件根植的对模型克隆的抗性,推断准确性在重量转移到不同的芯片时崩.
  • 该系统在分类心电图数据方面取得了很高的准确性,同时完全阻止了通过模型反转攻击来重建数据.
  • 在只向前学习的情况下,PUNN保持了竞争性分类准确性.

结论:

  • 建立了一个可扩展的框架,用于直接在商业闪存上安全,节能和保护隐私的神经计算.
  • 在现实应用中,V-NAND FF-PUNN为保护敏感数据和模型完整性提供了实用的解决方案.
  • 这种方法提供了内在的非克隆性和强大的防御对复杂的网络威胁.