Research map / Architectures and efficiency / RNNs and normalization
Batch normalization improvements
Analyzes and improves batch normalization through decorrelation, stable statistics, and regularization insight.
42 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in RNNs and normalization, keyword cluster “normalization · bn · mini”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
3 papers in 2023–24 and 3 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 0.5% to 0.3%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2016 · Normalization Propagation: A Parametric Technique for Removing Internal Covariate Shift in Deep Networks (ICML 2016 · 85 citations)
- 2017 · Batch Renormalization: Towards Reducing Minibatch Dependence in Batch-Normalized Models (NeurIPS 2017 · 397 citations)
- 2018 · In-Place Activated BatchNorm for Memory-Optimized Training of DNNs (CVPR 2018 · 351 citations)
- 2019 · Understanding the Disharmony Between Dropout and Batch Normalization by Variance Shift (CVPR 2019 · 312 citations)
- 2020 · Filter Response Normalization Layer: Eliminating Batch Dependence in the Training of Deep Neural Networks (CVPR 2020 · 126 citations)
- 2021 · High-Performance Large-Scale Image Recognition Without Normalization (ICML 2021 · 252 citations)
- 2022 · Delving Into the Estimation Shift of Batch Normalization in a Network (CVPR 2022 · 25 citations)
- 2023 · Tanh Works Better with Asymmetry (NeurIPS 2023 · 3 citations)
- 2024 · Batch Normalization Is Blind to the First and Second Derivatives of the Loss (AAAI 2024 · 1 citation)
- 2025 · MERIT: Maximum-normalized Element-wise Ratio for Language Model Large-batch Training (ICML 2025 · 0 citations)
- 2026 · The BatchNorm Illusion: Diagnosing Normalization Artifacts in Machine Unlearning Evaluation (NeurIPS 2026 · 0 citations)
Earliest papers here
- Normalization Propagation: A Parametric Technique for Removing Internal Covariate Shift in Deep Networks (ICML 2016 · 85 citations)
- Batch Renormalization: Towards Reducing Minibatch Dependence in Batch-Normalized Models (NeurIPS 2017 · 397 citations)
- Normalizing the Normalizers: Comparing and Extending Network Normalization Schemes (ICLR 2017 · 72 citations)
What has been tried
39 papers here report 460 measured changes (a component added, removed or swapped, compared with a baseline): helped 67%, no clear effect 8%, hurt 22%, mixed 3%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| replace batch normalization with layer normalization | MNIST image generation · log-likelihood separation of real and interpolated images | Helped | Batch Normalization Is Blind to the First and Second Derivatives of the Loss (AAAI 2024 · 1 citation) | |
| Remove element-wise update clipping | OpenWebText, GPT-2 large-batch training · convergence | Hurt | MERIT: Maximum-normalized Element-wise Ratio for Language Model Large-batch Training (ICML 2025 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 5 papers of this approach from 2024–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- CIFAR-10 · 2 of 5 papers
- C4 · 2 of 5 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 6 papers; from OpenAlex and the papers' affiliations).
- National University of Singapore · 2 papers
Use it
Download BibTeX of 25 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Normalization layer theory (37 papers)
- Weight normalization and decay interplay (32 papers)
Elsewhere in the topic
- Efficient network compression and training (229 papers)
- Theory of depth and width (223 papers)
- Representation geometry and collapse (169 papers)
- Dataset distillation and data-efficient training (158 papers)
- Neural scaling laws (155 papers)
- Distributed LLM inference systems (147 papers)
- Layer-wise analysis of language models (145 papers)
- Dataset condensation and flow estimation (126 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).