Research map / Label-efficient and robust learning / Semi-supervised learning
Label Propagation for Deep Semi-Supervised Learning
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 |
|---|---|---|---|
| Use label propagation rather than prediction-based pseudo-labels | Semi-supervised learning · CIFAR-10 | performance: 7-9% | Helped |
| Use the proposed method rather than TMMF | CIFAR-10 semi-supervised learning · CIFAR-10 | performance: 1-5% | Helped |
| Add label propagation to Mean Teacher | Mean Teacher · CIFAR-100 | performance: +1.5% | Helped |
| Use the proposed method rather than Mean Teacher | Semi-supervised learning · miniImageNet | top1: +2.5% | Helped |
| Use the proposed method rather than temporal ensemble and the Π-model | Semi-supervised learning · CIFAR-100, 10000 labels | performance: 2.7% | Helped |
| Add the proposed weighting and graph update to GTG | GTG · CIFAR-10, one split, 1000 labels | error: 40.33 -> 23.53 | Helped |
| Use the proposed propagation rather than GTG with weighting and graph update | Label Propagation for Deep Semi-Supervised Learning · CIFAR-10, one split, 1000 labels | error: 23.53 -> 21.45 | Helped |
| Vary the number of nearest neighbors k from 5 to 200 (ablation) | Proposed method · CIFAR-10, single split | error: [19.74, 21.22] | No clear effect |
| Use normalized feature vectors after average pooling rather than other feature representations (ablation) | Proposed method · CIFAR-10 | performance: more than 4% | Helped |
| Use convolutional-layer representations rather than normalized features after average pooling (ablation) | Proposed method · CIFAR-10 | performance: 20% or more | Hurt |
Closest papers by meaning
- LEARNING TO PROPAGATE LABELS: TRANSDUCTIVE PROPAGATION NETWORK FOR FEW-SHOT LEARNING (ICLR 2019 · similarity 0.89)
- Label Propagation with Weak Supervision (ICLR 2023 · similarity 0.88)
- All Labels Are Not Created Equal: Enhancing Semi-Supervision via Label Grouping and Co-Training (CVPR 2021 · similarity 0.88)
- Meta Propagation Networks for Graph Few-shot Semi-supervised Learning (AAAI 2022 · similarity 0.88)
- Why Propagate Alone? Parallel Use of Labels and Features on Graphs (ICLR 2022 · similarity 0.88)
- Contrastive Credibility Propagation for Reliable Semi-supervised Learning (AAAI 2024 · similarity 0.86)
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.