Research map / Label-efficient and robust learning / Self-supervised and contrastive
Understanding Negative Samples in Instance Discriminative Self-supervised Representation Learning
Kento Nozawa, Issei Sato
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 |
|---|---|---|---|
| Increase the number of negatives from K+1=32 to K+1=512 | SimCLR with ResNet-18 · CIFAR-10 validation | mean classifier accuracy: 72.75 -> 80.12 | Improved |
| Increase the number of negatives from K+1=32 to K+1=512 | SimCLR with ResNet-18 · CIFAR-10 validation | linear classifier accuracy: 77.13 -> 84.13 | Improved |
| Increase the number of negatives from K+1=32 to K+1=512 | SimCLR with ResNet-18, projection head discarded for evaluation · CIFAR-10 validation | linear classifier accuracy: 82.02 -> 87.66 | Improved |
| Increase the number of negatives from K+1=128 to K+1=1024 | SimCLR with ResNet-18 · CIFAR-100 validation | mean classifier accuracy: 32.67 -> 37.44 | Improved |
| Increase the number of negatives from K+1=128 to K+1=1024 | SimCLR with ResNet-18 · CIFAR-100 validation | linear classifier accuracy: 41.95 -> 46.57 | Improved |
| Increase the number of negatives from K+1=128 to K+1=1024 | SimCLR with ResNet-18, projection head discarded for evaluation · CIFAR-100 validation | linear classifier accuracy: 57.92 -> 59.46 | Improved |
Closest papers by meaning
- Disentangling Sampling and Labeling Bias for Learning in Large-output Spaces (ICML 2021 · similarity 0.86)
- Difficulty-Based Sampling for Debiased Contrastive Representation Learning (CVPR 2023 · similarity 0.85)
- Solving Inefficiency of Self-Supervised Representation Learning (ICCV 2021 · similarity 0.85)
- Does Negative Sampling Matter? a Review With Insights Into its Theory and Applications (TPAMI 2024 · similarity 0.85)
- Collapse-Proof Non-Contrastive Self-Supervised Learning (ICML 2025 · similarity 0.83)
- Measuring Self-Supervised Representation Quality for Downstream Classification Using Discriminative Features (AAAI 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.
