Skip to main content
zenodorestricted

Supplementary materials of the paper entitled: "Generating Natural Language Requirements via Adversarial Examples in Deep Learning"

<p>Supplementary materials of the paper entitled:</p> <p>&quot;Generating Natural Language Requirements via&nbsp;Adversarial Examples in Deep Learning&quot;</p> <p>File A: Datasets (.txt)</p> <p>&nbsp;&nbsp; &nbsp;A1:&nbsp; Webex</p> <p>&nbsp;&nbsp; &nbsp;A2: Zoom</p> <p>&nbsp;&nbsp; &nbsp;A3: Teams</p> <p>&nbsp;&nbsp; &nbsp;A4: Word</p> <p>&nbsp;&nbsp; &nbsp;A5: Power Point</p> <p>&nbsp;&nbsp; &nbsp;A6: Excel</p> <p>&nbsp;</p> <p>File B: Python Code (Both Ours and Baseline)</p> <p>&nbsp;&nbsp; &nbsp;B1: adversarial_samples.ipynb</p> <p>&nbsp;&nbsp; &nbsp;B2: Baseline.ipynb</p> <p>&nbsp;</p> <p>FIle C: Result Tables (.xlsx)</p> <p>&nbsp;&nbsp; &nbsp;C1: Table of Perturbed outputs in WebEx&nbsp;</p> <p>&nbsp;&nbsp; &nbsp;C2: Table of Perturbed outputs in Zoom</p> <p>&nbsp;&nbsp; &nbsp;C3: Table of Perturbed outputs in Teams</p> <p>&nbsp;&nbsp; &nbsp;C4: Table of Perturbed outputs in Word</p> <p>&nbsp;&nbsp; &nbsp;C5: Table of Perturbed outputs in Power Point</p> <p>&nbsp;&nbsp; &nbsp;C6: Table of Perturbed outputs in Excel (Excel)</p> <p>&nbsp;</p> <p>File&nbsp; D: Trend of Adversarial Shifts (Graphs)</p> <p>&nbsp;&nbsp; &nbsp;D1:&nbsp; Adversarial shifts of office suit (LSTM)</p> <p>&nbsp;&nbsp; &nbsp;D2: Adversarial shifts of video conferencing suit (LSTM)</p> <p>&nbsp;&nbsp; &nbsp;D3: Non-Adversarial shifts of office suit (LSTM)</p> <p>&nbsp;&nbsp; &nbsp;D4: Non-Adversarial shifts of video conferencing suit(LSTM)</p> <p>&nbsp;&nbsp; &nbsp;D5:&nbsp; Adversarial shifts of office suit (GRU)</p> <p>&nbsp;&nbsp; &nbsp;D6: Adversarial shifts of video conferencing suit (GRU)</p> <p>&nbsp;&nbsp; &nbsp;D7: Non-Adversarial shifts of office suit (GRU)</p> <p>&nbsp;&nbsp; &nbsp;D8: Non-Adversarial shifts of video conferencing suit(GRU)</p> <p>&nbsp;&nbsp; &nbsp;D9:&nbsp; Adversarial shifts of office suit (Bi-LSTM)</p> <p>&nbsp;&nbsp; &nbsp;D10: Adversarial shifts of video conferencing suit (Bi-LSTM)</p> <p>&nbsp;&nbsp; &nbsp;D11: Non-Adversarial shifts of office suit (Bi-LSTM)</p> <p>&nbsp;&nbsp; &nbsp;D12: Non-Adversarial shifts of video conferencing suit(Bi-LSTM)</p> <p>&nbsp;</p> <p>File E: Adversarial Examples (.pdf)</p> <p>&nbsp;&nbsp; &nbsp;E1: Adversarial vs Original in WebEx</p> <p>&nbsp;&nbsp; &nbsp;E2: Adversarial vs Original in Zoom</p> <p>&nbsp;&nbsp; &nbsp;E3: Adversarial vs Original in Teams</p> <p>&nbsp;&nbsp; &nbsp;E4: Adversarial vs Original in Word</p> <p>&nbsp;&nbsp; &nbsp;E5: Adversarial vs Original in Power Point</p> <p>&nbsp;&nbsp; &nbsp;E6: Adversarial vs Original in Excel</p> <p>&nbsp;</p> <p>File F: Questionnaire</p> <p><br> <br> <br> <br> <br> <br> &nbsp;</p>

ShareScore

8/100

Overall dataset sharing score

Score breakdown

These five areas show where the dataset supports — or may limit — practical reuse.

Stewardship
4
Harmonization
4
Access
0
Reuse readiness
0
Engagement
0