Research map / Architectures and efficiency / RNNs and normalization
Flipout: Efficient Pseudo-Independent Weight Perturbations on Mini-Batches
Yeming Wen, Paul Vicol, Jimmy Ba, Dustin Tran, Roger Grosse
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 |
|---|---|---|---|
| Use Flipout instead of shared multiplicative Gaussian perturbations | LSTM · PTB character-level test | bits-per-character: 1.230 -> 1.227 | Improved |
| Use Flipout-based DropConnect instead of shared-mask DropConnect | LSTM · PTB word-level test | perplexity: 75.71 -> 73.20 | Improved |
| Use Flipout instead of shared perturbations | Bayes by Backprop · FC and ConvLe, mini-batch size 8192 | iterations to converge: about 3 times fewer iterations | Improved |
| Use Flipout instead of shared perturbations | Bayes by Backprop · FC and ConvLe, mini-batch size 8192 | training speedup: 1.5x speedup | Improved |
| Use Flipout pseudo-independent perturbations instead of fully independent perturbations | Evolution strategies · MNIST | performance | No clear effect |
| Use Flipout to perturb each example pseudo-independently rather than share one perturbation across the batch | Weight-perturbed neural networks · FC, ConVGG, and LSTM variance experiments | gradient variance | Improved |
Closest papers by meaning
- Swapout: Learning an ensemble of deep architectures (NeurIPS 2016 · similarity 0.87)
- Rotational Equilibrium: How Weight Decay Balances Learning Across Neural Networks (ICML 2024 · similarity 0.86)
- Exact, Tractable Gauss-Newton Optimization in Deep Reversible Architectures Reveal Poor Generalization (NeurIPS 2024 · similarity 0.86)
- Understanding the Generalization of Stochastic Gradient Adam in Learning Neural Networks (NeurIPS 2025 · similarity 0.85)
- On Double Descent in Reinforcement Learning with LSTD and Random Features (ICLR 2024 · similarity 0.85)
- Improving Deep Learning Speed and Performance Through Synaptic Neural Balance (AAAI 2025 · similarity 0.85)
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.
