Research map / Label-efficient and robust learning / Semi-supervised learning
Learning by Association -- A Versatile Semi-Supervised Training Method for Neural Networks
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 unlabeled data through associative learning | Ours on SVHN with 500 labeled samples · SVHN test | error (%): 17.75 (0.65) -> 6.25 (0.32) | Helped |
| add unlabeled data through associative learning | Ours with only 20 labeled samples · SVHN test | error (%): 81.00 (3.01) -> 82.10 (1.91) | Hurt |
| add unlabeled data even when it duplicates labeled samples | Ours with all SVHN samples labeled · SVHN test | error (%): 3.09 (0.06) -> 2.69 (0.05) | Helped |
| add visit loss (ablation) | Ours on MNIST · MNIST test, 1000 labeled | error (%): 5.68 (0.53) -> 0.82 (0.12) | Helped |
| set visit loss weight to 0.25 rather than 0 (ablation) | Ours on SVHN · SVHN test, 1000 labeled | error (%): 7.91 (0.40) -> 6.31 (0.20) | Helped |
| increase visit loss weight from 0.25 to 1 (ablation) | Ours on SVHN · SVHN test, 1000 labeled | error (%): 6.31 (0.20) -> 6.43 (0.26) | Hurt |
| adapt source-trained model using unlabeled target-domain data | Ours, SVHN to MNIST · MNIST target test | error (%): 18.56 -> 0.51 | Helped |
| use associative training rather than Mutual Exclusivity + Transform. | Ours with 500 labeled samples · SVHN test | error (%): 9.62 (1.37) -> 6.25 (0.32) | Helped |
Closest papers by meaning
- Semi-supervised Learning with Multi-Head Co-Training (AAAI 2022 · similarity 0.87)
- Big Self-Supervised Models are Strong Semi-Supervised Learners (NeurIPS 2020 · similarity 0.86)
- Associated Learning: an Alternative to End-to-End Backpropagation that Works on CNN, RNN, and Transformer (ICLR 2022 · similarity 0.86)
- InterLUDE: Interactions between Labeled and Unlabeled Data to Enhance Semi-Supervised Learning (ICML 2024 · similarity 0.85)
- Scaling Up Semi-supervised Learning with Unconstrained Unlabelled Data (AAAI 2024 · similarity 0.84)
- CGMatch: A Different Perspective of Semi-supervised Learning (CVPR 2025 · similarity 0.84)
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.