DeepGuard: A DeepBillboard Attack Detection Technique against Connected and Autonomous Vehicles
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Artificial intelligence technology is leading the innovations in connected and automation vehicles (CAVs). This technology revolution mainly relies on embedded smart devices and high-tech sensors along with deep learning-based modules to provide the data and intellect necessary for automated decisions and responses. Generally, imagery data captured from dash cameras are fed into deep neural network models to identify street signs, traffic lights, and surrounding obj ects to augment steering decisions. Such neural networks are proven to be vulnerable to a wide range of adversarial attacks. Despite the emergence of adversarial manipulations, there has been a dramatic increase in the sophisticated methods of these attacks. One of these methods is the DeepBillBoard attack which uses machine-generated imagery applied to roadside billboards to induce errors to the steering model with the capability to dictate whether this error should cause the vehicle to veer to the left or the right. As the sheer risk of such attacks continues to grow, the safety, security, and reliability concerns grow even more. Such concerns cannot be tolerated given the safety-critical environment where CA V s operate. This paper proposes a novel approach, DeepGuard, to detect, counter, and neutralize DeepBillBoard attacks against CA V s. DeepGuard leverages advanced deep learning techniques to boost its generalization capabilities for detecting adversarial patterns used in DeepBillboard attacks. Experimental evaluation is conducted using existing driving datasets that reflect dynamic real-life scenarios. The evaluation results demonstrate that our solution achieves high detection effectiveness and computational efficiency.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it