Research map / Learning theory and algorithms / Kernels and regression theory
Dropout Training, Data-dependent Regularization, and Generalization Bounds
Read the paperCheck 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 truthful dropout (unbiased variant) instead of no dropout | neural networks · MNIST and CIFAR-10 | test error: MNIST FC 1.78 to 1.60; shallow conv 1.16 to 0.99; CIFAR-10 deep conv 9.86 to 8.20 | Helped |
| truthful dropout instead of traditional dropout | fully-connected and shallow convolutional networks on MNIST · MNIST | test error: 1.53 to 1.60 (FC); 0.81 to 0.99 (shallow conv) | Hurt |
| truthful dropout instead of traditional dropout | deep convolutional network on CIFAR-10 · CIFAR-10 | test error: 8.39 to 8.20 (about 0.19 points) | Helped |
| Truthful dropout variance and dropout as data-dependent regularizer; framework bound for one-hidden-layer ReLU nets | one-hidden-layer ReLU networks · theoretical generalization bound | generalization error rate | Helped |
Closest papers by meaning
- Adaptive Dropout with Rademacher Complexity Regularization (ICLR 2018 · similarity 0.91)
- On the Regularization Properties of Structured Dropout (CVPR 2020 · similarity 0.89)
- On the Generalization of Models Trained with SGD: Information-Theoretic Bounds and Implications (ICLR 2022 · similarity 0.89)
- On the Inherent Regularization Effects of Noise Injection During Training (ICML 2021 · similarity 0.89)
- Generalizability of Neural Networks Minimizing Empirical Risk Based on Expressive Power (ICLR 2025 · similarity 0.88)
- Stochastic Modified Equations and Dynamics of Dropout Algorithm (ICLR 2024 · 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.