Research map / Label-efficient and robust learning / Domain adaptation and generalization
ME-SFDA: Marginal Exploration with Pyramidal Atkinson-Shiffrin Memory for Source-Free Domain Adaptation
Chunzhi Liu, Yuwu Lu
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 two-step segmentation (ablation) | ME-SFDA · Hyperparameter / two-step ablation avg of three datasets (SFDA) | mean accuracy (%): 79.4 -> 84.7 (lambda2=1.6) | Improved |
| add separation loss alone to source-only (ablation) | ME-SFDA · Ablation avg of Office/Office-Home/VisDA (SFDA) | mean accuracy (%): 62.3 -> 75.9 | Improved |
| add clustering loss alone to source-only (ablation) | ME-SFDA · Ablation avg of Office/Office-Home/VisDA (SFDA) | mean accuracy (%): 62.3 -> 74.6 | Improved |
| add contrastive loss alone to source-only (ablation) | ME-SFDA · Ablation avg of Office/Office-Home/VisDA (SFDA) | mean accuracy (%): 62.3 -> 64.1 | Improved |
| combine clustering and separation without contrastive loss (ablation) | ME-SFDA · Ablation avg of Office/Office-Home/VisDA (SFDA) | mean accuracy (%): 75.9 -> 66.7 | Hurt |
| add contrastive loss on top of clustering and separation (ablation) | ME-SFDA · Ablation avg of Office/Office-Home/VisDA (SFDA) | mean accuracy (%): 66.7 -> 84.9 | Improved |
| raise lambda2 from 0.5 to 1.6 (ablation) | ME-SFDA · Hyperparameter / two-step ablation avg of three datasets (SFDA) | mean accuracy (%): 83.5 -> 84.7 | Improved |
| raise lambda2 from 1.6 to 2 (ablation) | ME-SFDA · Hyperparameter / two-step ablation avg of three datasets (SFDA) | mean accuracy (%): 84.7 -> 84.5 | No clear effect |
| use ME-SFDA instead of C-SFDA | ME-SFDA · Office-Home SFDA (12 scenarios mean) | mean accuracy (%): 73.5 -> 74.0 | Improved |
| use ME-SFDA instead of AaD on truck class | ME-SFDA · VisDA SFDA (truck class) | accuracy (%): 64.7 -> 41.3 | Hurt |
| extend to SSDA with 3 labeled shots (ME-SSDA) vs ProML | ME-SFDA · Office-Home SSDA 3-shot (mean) | accuracy (%): 78.1 -> 78.5 | 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.
- C-SFDA — C-SFDA: A Curriculum Learning Aided Self-Training Framework for Efficient Source Free Domain Adaptation (CVPR 2023) · better on 2 of 2 shared results
- SF(DA)2 — SF(DA)$^2$: Source-free Domain Adaptation Through the Lens of Data Augmentation (ICLR 2024) · better on 1 of 1 shared result
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
- Generalized Source-Free Domain Adaptation (ICCV 2021 · similarity 0.91)
- When Source-Free Domain Adaptation Meets Learning with Noisy Labels (ICLR 2023 · similarity 0.89)
- Understanding and Improving Source-free Domain Adaptation from a Theoretical Perspective (CVPR 2024 · similarity 0.89)
- Revisiting Source-Free Domain Adaptation: Insights into Representativeness, Generalization, and Variety (CVPR 2025 · similarity 0.88)
- Dynamic Target Distribution Estimation for Source-Free Open-Set Domain Adaptation (AAAI 2025 · similarity 0.88)
- Diffusion-Driven Progressive Target Manipulation for Source-Free Domain Adaptation (NeurIPS 2025 · similarity 0.87)
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.
