RASP: Read-All, Sparse-Plasticity Continual Learning in a Stationary-Size LoRA
cs.LG
Submitted: 2026-11-02
Updated: 2026-11-02
Terminology
Sources
- NeuroLoRA: Context-Aware Neuromodulation for Parameter-Efficient Multi-Task Adaptation
- Decomposing and Composing: Towards Efficient Vision-Language Continual Learning via Rank-1 Expert Pool in a Single LoRA
- C-LoRA: Continual Low-Rank Adaptation for Pre-trained Visual Models
- SPACE-LoRA: Allocating Activation-Subspace Protection for Continual Learning
- MLoRA:Continual Low-Rank Adaptation via Relative-Response Subspace Selection and Momentum Orthogonalization
- LoRA-Based Continual Learning with Constraints on Critical Parameter Changes
- One Adapter, Many Tasks: Task-Conditioned Feature Transformations for Continual Learning
- How Many Ranks Does a Token Need? A Noise-Floor Squelch for Rank-One LoRA Experts
- Routing Is Not Enough: Diagnosing Intra-Adapter Subspace Contention in MoE+LoRA Fine-Tuning
- ImageNet-21K Pretraining for the Masses
- Bilinear Optimization Divergence: Diagnosing Factor-Constrained LoRA Continual Learning
- A Dynamical Theory of LoRA in Continual Learning
- The Intruder Threshold: A Spectral Law for LoRA Fine-Tuning
- Structural features of the fly olfactory circuit mitigate the stability-plasticity dilemma in continual learning
Related papers
- A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis