Research map / Architectures and efficiency / Spiking networks
Sparse efficient spiking sequence models
Builds sparse, multiplication-free, and state-space spiking networks for sequence modeling.
26 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2018–2026) · an approach in Spiking networks, keyword cluster “neurons · biological · neuronal”
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 5 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 19.2% to 3.2%.
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 · Long short-term memory and Learning-to-learn in networks of spiking neurons (NeurIPS 2018 · 53 citations)
- 2020 · Biologically Plausible Sequence Learning with Spiking Neural Networks (AAAI 2020 · 2 citations)
- 2021 · Learning to live with Dale's principle: ANNs with separate excitatory and inhibitory units (ICLR 2021 · 40 citations)
- 2022 · Biologically plausible solutions for spiking networks with efficient coding (NeurIPS 2022 · 7 citations)
- 2023 · Complex Dynamic Neurons Improved Spiking Transformer Network for Efficient Automatic Speech Recognition (AAAI 2023 · 28 citations)
- 2024 · Towards efficient deep spiking neural networks construction with spiking activity based pruning (ICML 2024 · 8 citations)
- 2025 · SpikingSSMs: Learning Long Sequences with Sparse and Parallel Spiking State Space Models (AAAI 2025 · 10 citations)
- 2026 · LongSpike: Fractional Order Spiking State Space Models for Efficient Long Sequence Learning (NeurIPS 2026 · 0 citations)
Earliest papers here
- Long short-term memory and Learning-to-learn in networks of spiking neurons (NeurIPS 2018 · 53 citations)
- Biologically Plausible Sequence Learning with Spiking Neural Networks (AAAI 2020 · 2 citations)
- Learning to live with Dale's principle: ANNs with separate excitatory and inhibitory units (ICLR 2021 · 40 citations)
What has been tried
21 papers here report 182 measured changes (a component added, removed or swapped, compared with a baseline): helped 70%, no clear effect 8%, hurt 17%, mixed 4%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| use trainable instead of fixed threshold | LRA AVG · accuracy | 84.30 -> 84.33 | Helped | SpikingSSMs: Learning Long Sequences with Sparse and Parallel Spiking State Space Models (AAAI 2025 · 10 citations) |
| Add channel regrowth to pruning | CIFAR10 · test accuracy | about a 2\% decrease | Helped | Towards efficient deep spiking neural networks construction with spiking activity based pruning (ICML 2024 · 8 citations) |
| Add Frequency Encoding without Evolutionary Leak factor | CIFAR10, T=4, white-box PGD · accuracy | 15.59 -> 21.56 | Helped | FEEL-SNN: Robust Spiking Neural Networks with Frequency Encoding and Evolutionary Leak Factor (NeurIPS 2024 · 6 citations) |
| use sawtooth dilation instead of no dilation | Pixel CIFAR10 · accuracy | 66.21 -> 68.96; 72.97 -> 76.82 | Helped | Multiplication-Free Parallelizable Spiking Neurons with Efficient Spatio-Temporal Dynamics (NeurIPS 2025 · 2 citations) |
| Add synapse-level heterogeneous time constants rather than using homogeneous LIF dynamics | Pattern generation · MSE | Helped | HetSyn: Versatile Timescale Integration in Spiking Neural Networks via Heterogeneous Synapses (NeurIPS 2025 · 1 citation) | |
| remove temporal circuit | Moving MNIST · MSE | 47.0 -> 62.5 | Hurt | Autaptic Synaptic Circuit Enhances Spatio-temporal Predictive Learning of Spiking Neural Networks (ICML 2024 · 1 citation) |
| add long-range feedback connections to a UCF101-trained spiking network | Movie1 and Movie2 · TSRSA | +15.2% (Movie1); +201.1% (Movie2) | Helped | Long-Range Feedback Spiking Network Captures Dynamic and Static Representations of the Visual Cortex under Movie Stimuli (NeurIPS 2024 · 1 citation) |
| Remove the reset mechanism (PSN$_0$ rather than LIF) | Sequential CIFAR10; Sequential CIFAR100 · accuracy (%) | 81.50 -> 79.80; 55.45 -> 53.12 | Hurt | Efficient Parallel Training Methods for Spiking Neural Networks with Constant Time Complexity (ICML 2025 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 5 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- sMNIST · 3 of 5 papers
- psMNIST · 2 of 5 papers
- WikiText-103 · 2 of 5 papers
- DVS128 Gesture · 2 of 5 papers
- Speech Commands · 2 of 5 papers
- Sequential CIFAR10 · 2 of 5 papers
- Sequential CIFAR100 · 2 of 5 papers
Baselines
- S4D-Lin · 2 of 5 papers
- Transformer · 2 of 5 papers
- LMUFormer · 2 of 5 papers
- Spiking LMUFormer · 2 of 5 papers
- Binary S4D · 2 of 5 papers
- SpikeGPT · 2 of 5 papers
- BPTT · 2 of 5 papers
- SLTT · 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 19 papers; from OpenAlex and the papers' affiliations).
- Peking University · 6 papers
- Zhejiang University · 3 papers
- Chinese Academy of Sciences · 2 papers
- University of Chinese Academy of Sciences · 2 papers
- Peng Cheng Laboratory · 2 papers
- Georgia Institute of Technology · 2 papers
- Dalian University of Technology · 2 papers
- School of Computer Science and Technology · 2 papers
- CNRS · 2 papers
- Peng Cheng Lab · 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
- Bio-inspired spiking neuron models (22 papers)
Elsewhere in the topic
- Training deep spiking networks (141 papers)
- Spiking transformers (87 papers)
- Event-camera spiking networks (13 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).