ICACT20250142 Question.1
Questioner: jianms@nfu.edu.tw 2025-02-18 오후 8:00:52 |
ICACT20250142 Answer.1
Answer by Auhor bbgupta@asia.edu.tw 2025-02-18 오후 8:00:52
Chrome  Click!! |
| authors provide the high accuracy methods for smishing detection.
Considering the possible "updating" of smishing, how about the training time and computing resources required by the proposed method? |
Computing resources are relatively less as compared to traditional deep learning models |
ICACT20250142 Question.2
Questioner: mamadouba92@gmail.com 2025-02-18 오후 8:01:47 |
ICACT20250142 Answer.2
Answer by Auhor bbgupta@asia.edu.tw 2025-02-18 오후 8:01:47
Chrome Click!! |
| Why did the researchers choose to combine TinyBERT and Aquila Optimization specifically for Android devices, and what unique challenges does this platform present? |
TinyBERT model is selected because it is lightweight. |
ICACT20250142 Question.3
Questioner: mamadouba92@gmail.com 2025-02-18 오후 8:04:18 |
ICACT20250142 Answer.3
Answer by Auhor bbgupta@asia.edu.tw 2025-02-18 오후 8:04:18
Chrome Click!! |
| What specific contextual features of SMS messages does the model analyze to differentiate between legitimate and malicious content? |
We analyze 16 different features for this model. |
ICACT20250142 Question.4
Questioner: haun620@kyonggi.ac.kr 2025-02-18 오후 8:05:59 |
ICACT20250142 Answer.4
Answer by Auhor bbgupta@asia.edu.tw 2025-02-18 오후 8:05:59
Chrome Click!! |
| For real-time processing, inference time seems to be important, how long does the proposed method take? |
This paper is focus on the development of model, in future work we will focus on the real-time processing. |
ICACT20250142 Question.6
Questioner: benjikone@yahoo.fr 2025-02-18 오후 8:07:25 |
ICACT20250142 Answer.6
Answer by Auhor bbgupta@asia.edu.tw 2025-02-18 오후 8:07:25
Chrome Click!! |
| Why did you choose TinyBERT over other lightweight transformer models for smishing detection? Were there specific advantages in terms of performance, efficiency, or compatibility with Android devices? |
TinyBERT offers several advantages, especially when dealing with resource-constrained environments. Due to its smaller size, TinyBERT achieves much faster inference speeds compared to the original BERT model. TinyBERT requires significantly less memory, making it suitable for environments with limited computational resources, such as embedded systems or IoT devices. |
