Adversarial Attacks on Face Detectors using Neural Net based Constrained\n Optimization
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
Adversarial attacks involve adding, small, often imperceptible, perturbations\nto inputs with the goal of getting a machine learning model to misclassifying\nthem. While many different adversarial attack strategies have been proposed on\nimage classification models, object detection pipelines have been much harder\nto break. In this paper, we propose a novel strategy to craft adversarial\nexamples by solving a constrained optimization problem using an adversarial\ngenerator network. Our approach is fast and scalable, requiring only a forward\npass through our trained generator network to craft an adversarial sample.\nUnlike in many attack strategies, we show that the same trained generator is\ncapable of attacking new images without explicitly optimizing on them. We\nevaluate our attack on a trained Faster R-CNN face detector on the cropped\n300-W face dataset where we manage to reduce the number of detected faces to\n$0.5\\%$ of all originally detected faces. In a different experiment, also on\n300-W, we demonstrate the robustness of our attack to a JPEG compression based\ndefense typical JPEG compression level of $75\\%$ reduces the effectiveness of\nour attack from only $0.5\\%$ of detected faces to a modest $5.0\\%$.\n
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 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.002 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.002 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.004 | 0.003 |
| Research integrity | 0.001 | 0.003 |
| Insufficient payload (model declined to judge) | 0.001 | 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