Research map / Label-efficient and robust learning / Domain adaptation and generalization
Learning Systems Expansion with Efficient Heterogeneity-aware Knowledge Transfer
Gaole Dai, Huatao Xu, Yifan Yang, Rui Tan, Mo Li
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 Feature-based Coarse Selection (ablation) | HaT · HARBox (OTSE) | accuracy (%): 73.86 -> 80.57 | Improved |
| add Feature-based Coarse Selection (ablation) | HaT · ImageNet-R (OTSE) | accuracy (%): 56.86 -> 57.87 | Improved |
| add Centroids-Accuracy Joint Selection (ablation) | HaT · HARBox (OTSE) | accuracy (%): 79.92 -> 80.57 | Improved |
| add Centroids-Accuracy Joint Selection (ablation) | HaT · ImageNet-R (OTSE) | accuracy (%): 56.45 -> 57.87 | Improved |
| add Sample-wise Knowledge Fusion (ablation) | HaT · HARBox (OTSE) | accuracy (%): 69.05 -> 80.57 (11.5% improvement) | Improved |
| add Sample-wise Knowledge Fusion (ablation) | HaT · AD (OTSE) | accuracy (%): 60.17 -> 67.33 | Improved |
| add Adaptive Knowledge Injection (ablation) | HaT · AD (OTSE) | accuracy (%): 48.50 -> 67.33 | Improved |
| add Adaptive Knowledge Injection (ablation) | HaT · HARBox (OTSE) | accuracy (%): 66.73 -> 80.57 | Improved |
| add Quality Verification of the knowledge dictionary (ablation) | HaT · AD (OTSE) | accuracy (%): 54.33 -> 67.33 | Improved |
| add Quality Verification of the knowledge dictionary (ablation) | HaT · HARBox (OTSE) | accuracy (%): 76.66 -> 80.57 | Improved |
| use multi-round expansion instead of one-round | HaT · HARBox | accuracy (%): 67.51 -> 75.42 (DistillWeighted); 70.96 -> 79.27 (HaT) | Improved |
| add Cost-effective Adaptation (freeze encoders, cache features) (ablation) | HaT · ImageNet-R | memory and training time: 2.0x memory reduction and 2.3x training time reduction | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- LEAD — LEAD: Exploring Logit Space Evolution for Model Selection (CVPR 2024) · better on 5 of 5 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
- MergeNet: Knowledge Migration Across Heterogeneous Models, Tasks, and Modalities (AAAI 2025 · similarity 0.87)
- TAID: Temporally Adaptive Interpolated Distillation for Efficient Knowledge Transfer in Language Models (ICLR 2025 · similarity 0.83)
- Fantastic Gains and Where to Find Them: On the Existence and Prospect of General Knowledge Transfer between Any Pretrained Model (ICLR 2024 · similarity 0.83)
- When Models Know More Than They Can Explain: Quantifying Knowledge Transfer in Human-AI Collaboration (NeurIPS 2025 · similarity 0.83)
- Overcoming Generic Knowledge Loss with Selective Parameter Update (CVPR 2024 · similarity 0.83)
- Twin-Merging: Dynamic Integration of Modular Expertise in Model Merging (NeurIPS 2024 · similarity 0.83)
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.
