When the Merge Coefficient Stops Mattering: Proximity Regularized Merging for Continual LoRA Adaptation
cs.LG
Submitted: 2026-09-26
Updated: 2026-09-26
Code: https://github.com/limei0818/BECAME
Terminology
Sources
- Sharpness-Aware Minimization for Efficiently Improving Generalization
- Editing Models with Task Arithmetic
- Averaging Weights Leads to Wider Optima and Better Generalization
- BECAME: BayEsian Continual Learning with Adaptive Model MErging
- Decoupled Weight Decay Regularization
- Linear Mode Connectivity in Multitask and Continual Learning
- Gradient Projection Memory for Continual Learning
- How to train your ViT? Data, Augmentation, and Regularization in Vision Transformers
- SD-LoRA: Scalable Decoupled Low-Rank Adaptation for Class Incremental Learning
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks