ADVERSARIAL MACHINE LEARNINGPROPOSED STUDY
The small changes that fool a model
An experiment in model robustness, controlled perturbations, and the limits of accuracy.
PLANNED LAB · NOT YET COMPLETEDThis is a proposed lab brief. Findings, evidence, and reproducible steps will be added after the work is completed.
The question
How can a small, controlled change to an input alter a model’s decision?
The lab
An open model and a public or synthetic dataset inside an isolated research environment.
What the work will produce
- A baseline evaluation and threat model
- A reproducible robustness experiment
- A comparison of defenses and their trade-offs
This is a proposed lab brief. Findings, evidence, and reproducible steps will appear after the work is completed.
END OF LAB BRIEFBACK TO RESEARCH ↗