perturbations
Small modifications or alterations applied to inputs or models to analyze their stability and robustness. In AI, perturbations can test information integrity or assess model performance under varied conditions.
- ALMGuard: Safety Shortcuts and Where to Find Them as Guardrails for Audio–Language Models
- Attention! Your Vision Language Model Could Be Maliciously Manipulated
- Boosting Adversarial Transferability with Spatial Adversarial Alignment
- CellCLIP - Learning Perturbation Effects in Cell Painting via Text-Guided Contrastive Learning
- CodeCrash: Exposing LLM Fragility to Misleading Natural Language in Code Reasoning
- Fast MRI for All: Bridging Access Gaps by Training without Raw Data
- INC: An Indirect Neural Corrector for Auto-Regressive Hybrid PDE Solvers
- Inv-Entropy: A Fully Probabilistic Framework for Uncertainty Quantification in Language Models
- MIBP-Cert: Certified Training against Data Perturbations with Mixed-Integer Bilinear Programs
- On the Stability of Graph Convolutional Neural Networks: A Probabilistic Perspective
- PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis
- Robust Equilibria in Continuous Games: From Strategic to Dynamic Robustness
- Robust Explanations of Graph Neural Networks via Graph Curvatures
- SNEAKDOOR: Stealthy Backdoor Attacks against Distribution Matching-based Dataset Condensation
- The Dual Nature of Plasticity Loss in Deep Continual Learning: Dissection and Mitigation
- Training-free Detection of AI-generated images via Cropping Robustness
- Unveiling m-Sharpness Through the Structure of Stochastic Gradient Noise