-
精確檢索 : 冠狀病毒
-
模糊檢索 : 冠狀病毒
-
冠狀病毒感染冠狀病毒疾病
-
查詢出版品: 冠狀病毒
Read this ↓
- Proceedings
- OpenAccess
Deep Neural Network-based Data Forgery Detection in Transportation System
Abstracts
Intelligent Transportation System highly relies on sensory data from a large number of instruments. The advantage of having this data is that it will lead to better transportation policies formulated by the traffic authority. In order to become useful, this data should be in the original condition. Forgery attacks by adversaries are very critical in this situation, can result in a bad decision or wrong conclusion that harms traveler's needs. We propose two approaches to detecting attacks. We adopt Generative Adversarial Network to generate fake data to perform the attacks. The first method is based on Long Short-Term Memory (LSTM) classifier, which unfortunately, only performs better in the same data distribution. Our second method is LSTM regression combined with Manhattan Similarity to separate real data from the fake one. The last model performs very promising, it has an excellent performance on the same data distribution and on unknown data distribution.