Research map / Continual learning / Continual fine-tuning of LLMs
Sparse Tuning Enhances Plasticity in PTM-based Continual Learning
Huan Zhang, Shenghua Fan, Shuyu Dong, Yujin Zheng, Dingwen Wang, Fan Lyu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add MIST pre-adaptation | SimpleCIL · Cars-196 | A_T: 27.8 -> 43.5 | Improved |
| add MIST pre-adaptation | L2P · Cars-196 | A_T: 39.6 -> 52.7 | Improved |
| add MIST pre-adaptation | RanPAC · ImageNet-R | A_T: 77.9 -> 81.0 | Improved |
| add MIST pre-adaptation | SLCA · Cars-196 | A_T: 67.7 -> 74.6 | Improved |
| add MIST pre-adaptation | SimpleCIL · ImageNet-R | A_T: 54.3 -> 72.2 | Improved |
| replace MI-based selection with full fine-tuning pre-adaptation (ablation) | RanPAC · ImageNet-R | A_T: 77.9 -> 36.7 | Hurt |
| replace MI-based selection with random or L2 selection (ablation) | RanPAC · Cars-196 | A_T: 74.6 -> 10.9 (Rand) | Hurt |
| add MI loss to MI-based sparse selection (ablation) | RanPAC + MS · Cars-196 | A_T: 66.2 -> 72.1 | Improved |
| add gradient dropout to MS+ML (ablation) | RanPAC · ImageNet-R | A_T: 76.7 -> 81.0 | Improved |
| use MI-based selection instead of gradient-magnitude selection (both with MI loss and dropout) (ablation) | RanPAC · Cars-196 | A_T: 42.0 -> 76.4 | Improved |
| add MIST | RanPAC · CIFAR-100 | A_T: 90.8 -> 92.4 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- RanPAC — RanPAC: Random Projections and Pre-trained Models for Continual Learning (NeurIPS 2023) · better on 2 of 2 shared results
- SLCA — SLCA: Slow Learner with Classifier Alignment for Continual Learning on a Pre-trained Model (ICCV 2023) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- SAFE: Slow and Fast Parameter-Efficient Tuning for Continual Learning with Pre-Trained Models (NeurIPS 2024 · similarity 0.90)
- Model Sensitivity Aware Continual Learning (NeurIPS 2024 · similarity 0.90)
- Self-Expansion of Pre-trained Models with Mixture of Adapters for Continual Learning (CVPR 2025 · similarity 0.89)
- InfLoRA: Interference-Free Low-Rank Adaptation for Continual Learning (CVPR 2024 · similarity 0.88)
- Estimating and Orthogonalizing Unknown Pre-training Gradients for Continual Fine-tuning of Large Language Models (NeurIPS 2026 · similarity 0.88)
- Train with Perturbation, Infer after Merging: A Two-Stage Framework for Continual Learning (NeurIPS 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
