Prompting Creative Requirements via Traceable and Adversarial Examples in Deep Learning
收藏NIAID Data Ecosystem2026-03-14 收录
下载链接：
https://zenodo.org/record/7714144
下载链接
链接失效反馈官方服务：
资源简介：
File A: Datasets (.txt)
A1: Webex
A2: Zoom
A3: Teams
A4: Word
A5: PowerPoint
A6: Excel
File B: Python Code (Both ours and baseline)
B1: pert_class.ipynb
B2: Baseline.ipynb
FIle C: Result Tables (.xlsx)
C1: Table of perturbed outputs in Webex
C2: Table of perturbed outputs in Zoom
C3: Table of perturbed outputs in Teams
C4: Table of perturbed outputs in Word
C5: Table of perturbed outputs in PowerPoint
C6: Table of perturbed outputs in Excel
File D: Trend of Adversarial Shifts (Graphs)
D1: Webex Adversarial Shifts
D2: Zoom Adversarial Shifts
D3: Teams Adversarial Shifts
D4: Word Adversarial Shifts
D5: Powerpoint Adversarial Shifts
D6: Excel Adversarial Shifts
File E: Trend of Non-Adversarial Shifts (Graphs)
E1: Webex Non-Adversarial Shifts
E2: Zoom Non-Adversarial Shifts
E3: Teams Non-Adversarial Shifts
E4: Word Non-Adversarial Shifts
E5: Powerpoint Non-Adversarial Shifts
E6: Excel Non-Adversarial Shifts
File F: Adversarial Examples (.pdf)
F1: Adversarial vs original in Webex
F2: Adversarial vs original in Zoom
F3: Adversarial vs original in Teams
F4: Adversarial vs original in Word
F5: Adversarial vs original in PowerPoint
F6: Adversarial vs original in Excel
File G: Questionnaire
创建时间：
2023-03-10
