mitigation strategies
These are approaches designed to reduce the negative impacts of biases or errors in AI systems. Mitigation strategies can include techniques like data balancing, model auditing, or implementing fairness constraints.
- Adaptive Distraction: Probing LLM Contextual Robustness with Automated Tree Search
- Adjusting Initial Noise to Mitigate Memorization in Text-to-Image Diffusion Models
- Non-exchangeable Conformal Prediction with Optimal Transport: Tackling Distribution Shift with Unlabeled Data
- Risk Management for Mitigating Benchmark Failure Modes: BenchRisk
- Shape it Up! Restoring LLM Safety during Finetuning
- Shortcuts and Identifiability in Concept-based Models from a Neuro-Symbolic Lens