InfoResearchPreprint
Quantum-inspired orthogonal adapters for certifiably robust medical image classification
- Published
- Record updated
Summary
This paper proposes fine-tuning certified medical image classifiers with quantum-inspired orthogonal butterfly layers, which keep the Lipschitz bounds of pretrained networks while using only O(n log n) trainable parameters. On MNIST and MedMNIST the approach improves certified robustness over conventional alternatives, and on trichoscopic alopecia images it reaches roughly twice the certified accuracy of conventional fine-tuning and LoRA at comparable classification accuracy.
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)