Research map / Architectures and efficiency / RNNs and normalization
Hebbian and predictive coding
Develops Hebbian and predictive-coding learning algorithms and analyzes their dynamics.
52 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2018–2026) · an approach in RNNs and normalization, keyword cluster “brain · biological · feedback”
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.
14 papers in 2023–24 and 19 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic held steady, from 2.2% to 2.0%.
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
- 2018 · Learning Nonlinear Dynamics in Efficient, Balanced Spiking Networks Using Local Plasticity Rules (AAAI 2018 · 66 citations)
- 2019 · Selfless Sequential Learning (ICLR 2019 · 75 citations)
- 2020 · Predictive coding in balanced neural networks with noise, chaos and delays (NeurIPS 2020 · 19 citations)
- 2021 · Large Associative Memory Problem in Neurobiology and Machine Learning (ICLR 2021 · 55 citations)
- 2022 · Kernel Memory Networks: A Unifying Framework for Memory Modeling (NeurIPS 2022 · 2 citations)
- 2023 · Long Sequence Hopfield Memory (NeurIPS 2023 · 8 citations)
- 2024 · A Stable, Fast, and Fully Automatic Learning Algorithm for Predictive Coding Networks (ICLR 2024 · 8 citations)
- 2025 · Multi-View Incremental Learning with Structured Hebbian Plasticity for Enhanced Fusion Efficiency (AAAI 2025 · 3 citations)
- 2026 · Bidirectional Predictive Coding (ICLR 2026 · 5 citations)
Earliest papers here
- Learning Nonlinear Dynamics in Efficient, Balanced Spiking Networks Using Local Plasticity Rules (AAAI 2018 · 66 citations)
- Deep Predictive Coding Network for Object Recognition (ICML 2018 · 51 citations)
- Differentiable plasticity: training plastic neural networks with backpropagation (ICML 2018 · 17 citations)
What has been tried
36 papers here report 298 measured changes (a component added, removed or swapped, compared with a baseline): helped 65%, no clear effect 11%, hurt 18%, mixed 5%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Increase CNN width using multiplier C | CIFAR-10 · final accuracy | Mixed | A Stable, Fast, and Fully Automatic Learning Algorithm for Predictive Coding Networks (ICLR 2024 · 8 citations) | |
| Add structured Hebbian plasticity (C3) to C1+C2 | 100leaves, Animals, Flower17, NGs, NoisyMNIST_15000, YaleB_Extended · ACC | 90.79 (0.69) -> 91.36 (0.68); 84.01 (0.11) -> 84.42 (0.10); 56.21 (0.42) -> 57.79 (0.37); 94.37 (0.42) -> 95.04 (0.64); | Helped | Multi-View Incremental Learning with Structured Hebbian Plasticity for Enhanced Fusion Efficiency (AAAI 2025 · 3 citations) |
| Increase ELM memory units from d_m=10 to d_m=200 | Pathfinder · accuracy | 57.83 -> 72.98 | Helped | The Expressive Leaky Memory Neuron: an Efficient and Expressive Phenomenological Neuron Model Can Solve Long-Horizon Tasks. (ICLR 2024 · 3 citations) |
| add random noise pretraining before data training | MNIST · accuracy | 97.26 ± 0.07 -> 97.76 ± 0.07; \blacktriangle 0.49 ± 0.06 | Helped | Pretraining with Random Noise for Fast and Robust Learning without Weight Transport (NeurIPS 2024 · 2 citations) |
| Use only the original Oja loss rather than the full SPHeRe combination | CIFAR-10 · accuracy | 81.18 -> 73.8 | Hurt | Rethinking Hebbian Principle: Low-Dimensional Structural Projection for Unsupervised Learning (NeurIPS 2025 · 1 citation) |
| remove neuron-level models and synchronization as a representation | 2D mazes · maze-solving performance | Hurt | Continuous Thought Machines (NeurIPS 2025 · 1 citation) | |
| Use ViT encoder-decoder instead of ResNet | Google Street View (3×80×40) · FIDr | 48.05 -> 13.27 | Helped | Vector Quantization in the Brain: Grid-like Codes in World Models (NeurIPS 2025 · 0 citations) |
| decrease α from 1 to 0 | MNIST test · accuracy | 95.85\!±\!0.68 -> 98.46\!±\!0.07 | Helped | Two Tales of Single-Phase Contrastive Hebbian Learning (ICML 2024 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 9 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- CIFAR-10 · 4 of 9 papers
- MNIST · 3 of 9 papers
- CIFAR-100 · 3 of 9 papers
- ImageNet-1K · 2 of 9 papers
Baselines
- BP · 2 of 9 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 33 papers; from OpenAlex and the papers' affiliations).
- Peking University · 3 papers
- Tsinghua University · 2 papers
- University of Electronic Science and Technology of China · 2 papers
- University of Oxford · 2 papers
- Chalmers University · 2 papers
- Chalmers University of Technology · 2 papers
- Sakana AI · 2 papers
Use it
Download BibTeX of 35 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
- Brain-inspired network comparison (57 papers)
- Backprop-free credit assignment (44 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).