Keep Moving Forward
Watermark Stealing in Large Language Models 阅读笔记 Watermark Stealing in Large Language Models 阅读笔记
论文研读笔记（2-3页） 论文题目：Watermark Stealing in Large Language Models 发表时间：2024 发表期刊/会议：ICML （CCF A） 1. 内容简介（简要概括文章，不可以直接翻译
2024-09-21 wangxh
论文笔记 大语言模型 水印窃取
Reinforcement Learning-Based Black-Box Model Inversion Attacks 阅读笔记 Reinforcement Learning-Based Black-Box Model Inversion Attacks 阅读笔记
Reinforcement Learning-Based Black-Box Model Inversion AttacksMetadata Type: ConferencePaper Title: Reinforcement Learni
2024-01-18 wangxh
机器学习 论文笔记 强化学习 模型逆向 GAN
CADE: Detecting and Explaining Concept Drift Samples for Security Applications 阅读笔记 CADE: Detecting and Explaining Concept Drift Samples for Security Applications 阅读笔记
CADE: Detecting and Explaining Concept Drift Samples for Security ApplicationsAbstractConcept drift poses a critical cha
2023-10-12 wangxh
机器学习 论文笔记 概念漂移
代码执行日志记录 代码执行日志记录
忘了在哪个开源项目里看到了一个执行日志系统。它会在每次运行时备份代码。对于我这种随时会忘掉实验结果和实验设置的人来说，这种东西应该会很有用……吧。凭印象和自己的喜好复现了一下，具体如下。 首先，项目应该保持如下的结构。code路径下存放所有
2023-10-05 wangxh
日志系统 执行日志
2023安全顶会AI安全相关汇总 2023安全顶会AI安全相关汇总
NDSS对抗样本防御 Adversarial Robustness for Tabular Data through Cost and Utility Awareness BARS: Local Robustness Certificati
2023-09-06 wangxh
顶会 汇总 AI安全 机器学习
SplineCam 阅读笔记 SplineCam 阅读笔记
SplineCam: Exact Visualization and Characterization of Deep Network Geometry and Decision BoundariesMetadata Type: Journ
2023-08-07 wangxh
机器学习 论文笔记 样条理论 SplineCam 可解释性
3 / 4