linear mode connectivity
Linear mode connectivity refers to the phenomenon where models can connect different local minima through linear paths in weight space, suggesting that multiple models can achieve similar performance. This has implications for understanding optimization landscapes in deep learning.
- CodeMerge: Codebook-Guided Model Merging for Robust Test-Time Adaptation in Autonomous Driving
- From Dormant to Deleted: Tamper-Resistant Unlearning Through Weight-Space Regularization
- Generalized Linear Mode Connectivity for Transformers
- Generalized Linear Mode Connectivity for Transformers
- On Linear Mode Connectivity of Mixture-of-Experts Architectures
- On Linear Mode Connectivity of Mixture-of-Experts Architectures
- RoMa: A Robust Model Watermarking Scheme for Protecting IP in Diffusion Models