Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
Activity Pruning for Efficient Spiking Neural Networks
Authors: Tong Bu, Xinyu Shi, Zhaofei Yu
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Through extensive experiments on multiple datasets, we demonstrate that our approach achieves significant reductions in average firing rates and synaptic operations without sacrificing much accuracy. Furthermore, we show that our method complements weight-based pruning techniques and successfully trains an SNN with only 0.06 average firing rate and 2.22M parameters on Image Net, highlighting its potential for building highly efficient and scalable SNN models. Code is available at https://github.com/putshua/Activity-Pruning-SNN. The paper includes a dedicated section "4 Experiments" which contains tables comparing accuracy, average firing rate, and synaptic operations of the proposed method against existing methods and an ablation study. |
| Researcher Affiliation | Academia | Tong Bu Institution for Artificial Intelligence School of Computer Science Peking University EMAIL Xinyu Shi Institution for Artificial Intelligence School of Computer Science Peking University EMAIL Zhaofei Yu Institution for Artificial Intelligence School of Computer Science Peking University EMAIL |
| Pseudocode | Yes | The full derivation and pseudo-code of the overall algorithm are provided in the supplementary material. |
| Open Source Code | Yes | Code is available at https://github.com/putshua/Activity-Pruning-SNN. Also, in the NeurIPS checklist, Question 5 states: "Does the paper provide open access to the data and code, with sufficient instructions to faithfully reproduce the main experimental results, as described in supplemental material?" Answer: [Yes] Justification: We will provide open access to the code while the data we use are all publicly available. |
| Open Datasets | Yes | Through extensive experiments on multiple datasets, we demonstrate that our approach achieves significant reductions in average firing rates and synaptic operations without sacrificing much accuracy... on Image Net. The paper uses CIFAR-10, Image Net-100, and DVS-CIFAR10 datasets, which are all well-known public datasets. Also, in the NeurIPS checklist, Question 5 states: "Does the paper provide open access to the data and code, with sufficient instructions to faithfully reproduce the main experimental results, as described in supplemental material?" Answer: [Yes] Justification: We will provide open access to the code while the data we use are all publicly available. |
| Dataset Splits | Yes | In section 4.1 "General experimental setting", the paper states: "We provide the general hyper-parameters and the neuron parameter setting in here and the supplementary material for better reproducibility." Additionally, in the NeurIPS checklist, Question 6 states: "Does the paper specify all the training and test details (e.g., data splits, hyperparameters, how they were chosen, type of optimizer, etc.) necessary to understand the results?" Answer: [Yes] Justification: We provide all the detail in the training and evaluating process in the supplementary material. |
| Hardware Specification | Yes | In the NeurIPS checklist, Question 8 states: "For each experiment, does the paper provide sufficient information on the computer resources (type of compute workers, memory, time of execution) needed to reproduce the experiments?" Answer: [Yes] Justification: The information of computer resources for each experiments can be found in the supplementary material. |
| Software Dependencies | Yes | In section 4.1 "General experimental setting", the paper states: "We provide the general hyper-parameters and the neuron parameter setting in here and the supplementary material for better reproducibility." Additionally, in the NeurIPS checklist, Question 6 states: "Does the paper specify all the training and test details (e.g., data splits, hyperparameters, how they were chosen, type of optimizer, etc.) necessary to understand the results?" Answer: [Yes] Justification: We provide all the detail in the training and evaluating process in the supplementary material. |
| Experiment Setup | Yes | In section 4.1 "General experimental setting", the paper states: "We provide the general hyper-parameters and the neuron parameter setting in here and the supplementary material for better reproducibility." Table 1 also explicitly lists "T" (simulation time-steps), "λs" (spike activity regularization coefficient), and "λθ" (threshold adaptation coefficient) as parameters used in the experiments. |