Research map / Label-efficient and robust learning / Domain adaptation and generalization
Selective Random Walk for Transfer Learning in Heterogeneous Label Spaces
Qiao Xiao, Yu Zhang, Qiang Yang
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 meta weight network to weigh transfer sequences (vs equal weights of 1.0) (ablation) | SRW · MiniImagenet tasks (within dataset) | accuracy | Improved |
| add meta weight network to weigh transfer sequences (vs random weights in [0,1]) (ablation) | SRW · MiniImagenet tasks (within dataset) | accuracy | Improved |
| add similarity loss l1 on neighboring points in transfer sequences (ablation) | SRW · MiniImagenet tasks and six cross-dataset tasks | accuracy | Improved |
| add LSTM reconstruction loss l2 for the target point in the sequence (ablation) | SRW · MiniImagenet tasks and six cross-dataset tasks | accuracy | Improved |
| use inner product instead of cosine similarity or euclidean distance for graph edges (ablation) | SRW · MiniImagenet tasks | accuracy | Improved |
| set gamma to 0.3 within range 0.1 to 1.0 (ablation) | SRW · six cross-dataset tasks | average accuracy | Improved |
| vary inner learning rate with outer rate fixed at 0.01 (ablation) | SRW · cross-dataset tasks | accuracy | No clear effect |
| increase labeled target instances (5 to 30) (ablation) | SRW · MiniImagenet and cross-dataset tasks | accuracy | Improved |
| forced domain alignment (DAN, DANN) or source transfer vs training on target only | DNN Dt baseline · heterogeneous label space tasks | accuracy | Hurt |
| evaluate on AwA unseen vs seen categories relative to ImageNet-1K | SRW · C/M to A_seen and A_unseen | accuracy | No clear effect |
Closest papers by meaning
- CGMatch: A Different Perspective of Semi-supervised Learning (CVPR 2025 · similarity 0.83)
- A Unified Framework for Heterogeneous Semi-supervised Learning (CVPR 2025 · similarity 0.83)
- Mixed-Sample SGD: an End-to-end Analysis of Supervised Transfer Learning (NeurIPS 2025 · similarity 0.83)
- A High-Dimensional Statistical Method for Optimizing Transfer Quantities in Multi-Source Transfer Learning (NeurIPS 2025 · similarity 0.83)
- Enhancing Semi-supervised Domain Adaptation via Effective Target Labeling (AAAI 2024 · similarity 0.83)
- Transfer Learning for Diffusion Models (NeurIPS 2024 · similarity 0.82)
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.
