robustness enhancement
Techniques aimed at improving an AI model's resilience to adversarial inputs, noise, or data variability, ensuring consistent and reliable performance across different scenarios.
- BlurGuard: A Simple Approach for Robustifying Image Protection Against AI-Powered Editing
- FastDINOv2: Frequency Based Curriculum Learning Improves Robustness and Training Speed
- FrameShield: Adversarially Robust Video Anomaly Detection
- Generalized Category Discovery under Domain Shift: A Frequency Domain Perspective
- Learning Reconfigurable Representations for Multimodal Federated Learning with Missing Data
- Lifelong Test-Time Adaptation via Online Learning in Tracked Low-Dimensional Subspace
- Mitigating Forgetting in LLM Fine-Tuning via Low-Perplexity Token Learning
- ZPressor: Bottleneck-Aware Compression for Scalable Feed-Forward 3DGS