InfoResearchPreprint
Conv-Reluplex: A Verification Framework For Convolution Neural Networks
- Published
- Record updated
Summary
Conv-Reluplex is a verification framework that extends the Reluplex tool to convolutional neural networks (CNNs), covering their convolutional and pooling layers. It generates adversarial examples when classification is misguided. In MNIST experiments, the original CNN improved when retrained with these adversarial examples, and classification precision rose significantly.
Topics
Related items
- MediumGemini AI agents repeatedly bypassed safeguards to attack real websiteSimilar attack · BleepingComputer
- InfoSTA: A Split Temporal Adversary Method for Evaluating Robustness of Multimodal RL-based Driving PoliciesSimilar attack · Arxiv (cs.RO + cs.CV security)
- InfoSmoothShift: Disrupting Visual Adversarial Attacks on LVLMs at Test TimeSimilar attack · Arxiv (cs.RO + cs.CV security)
- InfoCracks in Embeddings: Robustness and Stability of JEPA-Based Chest X-Ray RepresentationsSimilar attack · Arxiv (cs.RO + cs.CV security)
- LowAdversarial Tracking: Attacks on Image Sensors for Object Tracking in Autonomous VehiclesSimilar attack · Arxiv (cs.RO + cs.CV security)