SpikeLite: Lightweight Spiking Neural Networks for Time-Series Forecasting

Bang Hu; Changze Lv; Mingjie Li; Xiaoqing Zheng; Wei cao; Fan Zhang

cs.LGSep 28, 2026

SpikeLite: Lightweight Spiking Neural Networks for Time-Series Forecasting

Authors: Bang Hu, Changze Lv, Mingjie Li, Xiaoqing Zheng, Wei cao, Fan Zhang

Organizations: School of Computer Science, Fudan University, Shanghai, China

Abstract

Spiking neural networks (SNNs) offer an energy-efficient paradigm for time-series forecasting through spike-driven computation. However, recent SNN forecasters often pursue higher accuracy through increasingly complex attention mechanisms, or specialized neuronal dynamics, weakening the lightweight motivation of SNNs. We introduce SpikeLite, a spiking forecasting framework built around two modules: a Frequency-Selective Spiking Encoder (FSSE) for frequency-sensitive temporal encoding and a Sparse Spiking Channel Attention (SSCA) module for selective cross-channel interaction. FSSE exploits the low-pass filtering behavior of LIF dynamics to reorganize each input sequence into frequency-sensitive components while collectively preserving the input at the decomposition stage. SSCA then learns a binary mask from encoded channel representations and uses it to selectively exchange information within spike-driven self-attention, retaining informative cross-channel interactions while suppressing redundant ones. When explicit channel interaction is unnecessary, SpikeLite uses the lighter FSSE-only channel-independent path. Experiments under the SeqSNN and SpikF protocols cover four standard multivariate and eight long-term forecasting benchmarks. SpikeLite achieves the best aggregate performance under both protocols, with an average $R^2$ of 0.790 and RSE of 0.440, and lowest average MSE/MAE of 0.343/0.345 in long-term forecasting. Moreover, evaluation on the ECL dataset shows that SpikeLite achieves the lowest reported energy consumption, further demonstrating its potential for energy-efficient time-series forecasting.

Figures & tables

Figure 1: Overview of SpikeLite. FSSE constructs frequency-sensitive representations for individual input channels, while SSCA selectively exchanges information across channels. The lower panel details mask generation and masked spiking attention: the mask is estimated from the encoded channel representations and applied to the attention scores before value aggregation.

Figure 2: LIF responses and the FSSE encoding process. (a) LIF responses under different decay factors. (b) Details of Frequency-Selective Spiking Encoder.

Method	Metric	METR-LA				PEMS-BAY				Solar				Electricity				Summary
		6	24	48	96	6	24	48	96	6	24	48	96	6	24	48	96	Avg.	Rank
ARIMA	$R^{2}\uparrow$	.687	.441	.282	.265	.741	.723	.692	.670	.951	.847	.725	.689	.963	.960	.914	.863	.713	9.69
	RSE $\downarrow$	.575	.742	.889	.902	.532	.548	.562	.612	.202	.365	.588	.589	.522	.534	.564	.599	.583	9.44
GP	$R^{2}\uparrow$	.685	.437	.265	.233	.732	.712	.689	.665	.944	.836	.711	.675	.962	.968	.912	.852	.705	11.03
	RSE $\downarrow$	.572	.738	.912	.925	.544	.532	.577	.592	.225	.388	.612	.575	.603	.612	.633	.642	.605	10.28
Autoformer	$R^{2}\uparrow$	.762	.548	.411	.282	.782	.711	.689	.668	.960	.852	.791	.701	.980	.977	.975	.963	.753	8.00

Table 1: Experimental results on standard multivariate forecasting under the SeqSNN protocol [ 12 ] . Bold and underlined values denote the best and second-best results, respectively.

\uparrow

\downarrow

indicates that higher/lower values are better.

Dataset	iTransformer		RLinear		PatchTST		Crossformer		TimesNet		DLinear		SCINet		Autoformer		SpikF		SpikeLite
	MSE	MAE	MSE	MAE	MSE	MAE	MSE	MAE	MSE	MAE	MSE	MAE	MSE	MAE	MSE	MAE	MSE	MAE	MSE	MAE
ECL	.178	.270	.219	.298	.205	.290	.244	.334	.192	.295	.212	.300	.268	.365	.227	.338	.183	.275	.175	.263
Weather	.258	.278	.272	.291	.259	.281	.259	.315	.259	.287	.265	.317	.292	.363	.338	.382	.245	.265	.245	.265
ETTh1	.454	.447	.446	.434	.469	.454	.529	.522	.458	.450	.456	.452	.747	.647	.496	.487	.440	.428	.443	.429
ETTh2	.383	.407	.374	.398	.387	.407	.942	.684	.414	.427	.559	.515	.954	.723	.450	.459	.372	.394	.374	.394
ETTm1	.407	.410	.414	.407	.387	.400	.513	.496	.400	.406	.403	.407	.485	.481	.588	.517	.388	.385	.387	.389

Table 2: Long-term forecasting performance under the SpikF protocol [ 13 ] . Results are averaged across four prediction horizons (

96

192

336

, and

720

) and three random seeds. Lower MSE and MAE indicate better performance. Bold and underlined values indicate the best and second-best results, respectively.

Figure 3: Frequency-sweep analysis of FSSE: component-wise fundamental gain and branch-wise firing probability.

Figure 4: Ablation results on Traffic under the SpikF protocol. Lower values indicate better forecasting performance.

Model	Params	OPs	Energy ( $\mu$ J)	MSE@720
SpikF	1.2K	0.13G	117.66	.219
DLinear	0.14M	45M	205.19	.245
iTransformer	1.6M	0.72G	3289.29	.225
SpikeLite	67.2K	0.02G	95.30	.218

Table 3: Model complexity, estimated energy, and forecasting accuracy on ECL at a prediction horizon of 720.

Appendix figures & tables6 assets

Supplementary material from the paper’s appendix.

Appendix

Protocol	Dataset	Domain	Variables/nodes	Main temporal characteristics
SeqSNN	METR-LA	traffic sensors	207	Road-speed measurements with spatial correlation, daily repetition, and short-term congestion changes.
	PEMS-BAY	traffic sensors	325	Large sensor network with correlated locations, recurring traffic cycles, and rapidly changing local conditions.
	Solar	solar power	137	Diurnal structure with weather-driven local fluctuations and intermittent high-frequency changes.
	Electricity	electricity load	321	Heterogeneous consumption channels with periodic structure and channel-specific temporal profiles.
SpikF	ECL	electricity load	321	Long-horizon electricity demand with periodic and heterogeneous channel behavior.
	Weather	meteorological series	21	Smooth physical variables with seasonal trends and multi-scale local variation.

Appendix

Table 4: Datasets used by the two evaluation protocols. “Variables/nodes” denotes the input dimensionality after protocol preprocessing.

Dataset	Horizon	SpikeLite		SpikF		iTransformer		RLinear		PatchTST		Crossformer		TimesNet		DLinear		SCINet		Autoformer
		MSE	MAE	MSE	MAE	MSE	MAE	MSE	MAE	MSE	MAE	MSE	MAE	MSE	MAE	MSE	MAE	MSE	MAE	MSE	MAE
ECL	96	.145	.236	.156	.252	.148	.240	.201	.281	.181	.270	.219	.314	.168	.272	.197	.282	.247	.345	.201	.317
	192	.161	.250	.169	.262	.162	.253	.201	.283	.188	.274	.231	.322	.184	.289	.196	.285	.257	.355	.222	.334
	336	.175	.264	.188	.281	.178	.269	.215	.298	.204	.293	.246	.337	.198	.300	.209	.301	.269	.369	.231	.338
	720	.218	.300	.219	.306	.225	.317	.257	.331	.246	.324	.280	.363	.220	.320	.245	.333	.299	.390	.254	.361
ECL Avg.		.175	.263	.183	.275	.178	.270	.219	.298	.205	.290	.244	.334	.192	.295	.212	.300	.268	.365	.227	.338

Appendix

Table 5: Full horizon-wise long-term forecasting results under the SpikF protocol. Each entry reports MSE and MAE. Lower values are better. Bold and underlined values denote the best and second-best results, respectively.

Dataset	Metric	96	192	336	720	Avg.
ECL	MSE	$.1453\!\pm\!.0006$	$.1610\!\pm\!.0000$	$.1750\!\pm\!.0000$	$.2187\!\pm\!.0012$	$.1750\!\pm\!.0003$
ECL	MAE	$.2363\!\pm\!.0006$	$.2500\!\pm\!.0000$	$.2640\!\pm\!.0000$	$.3007\!\pm\!.0012$	$.2628\!\pm\!.0003$
Weather	MSE	$.1587\!\pm\!.0012$	$.2090\!\pm\!.0000$	$.2660\!\pm\!.0000$	$.3460\!\pm\!.0000$	$.2449\!\pm\!.0003$
Weather	MAE	$.1977\!\pm\!.0012$	$.2440\!\pm\!.0000$	$.2853\!\pm\!.0006$	$.3380\!\pm\!.0000$	$.2663\!\pm\!.0003$
ETTm2	MSE	$.1753\!\pm\!.0006$	$.2400\!\pm\!.0000$	$.3023\!\pm\!.0006$	$.3950\!\pm\!.0000$	$.2782\!\pm\!.0003$
ETTm2	MAE	$.2540\!\pm\!.0000$	$.2950\!\pm\!.0000$	$.3360\!\pm\!.0000$	$.3900\!\pm\!.0000$	$.3188\!\pm\!.0000$

Appendix

Table 6: Three-seed stability under the SpikF protocol. Each entry reports the mean

\pm

sample standard deviation over three independent runs. The Avg. column is computed within each run over the four prediction horizons.

Dataset	Metric	Horizon 6	Horizon 24	Horizon 48	Horizon 96
PEMS-BAY	$R^{2}\!\uparrow$	$.899\pm.001$	$.781\pm.001$	$.715\pm.000$	$.639\pm.003$
PEMS-BAY	$RSE\downarrow$	$.361\pm.000$	$.506\pm.001$	$.577\pm.000$	$.649\pm.002$
METR-LA	$R^{2}\!\uparrow$	$.847\pm.001$	$.622\pm.000$	$.452\pm.000$	$.316\pm.001$
METR-LA	$RSE\downarrow$	$.414\pm.001$	$.649\pm.000$	$.781\pm.002$	$.873\pm.002$
Solar	$R^{2}\!\uparrow$	$.962\pm.000$	$.885\pm.001$	$.810\pm.001$	$.767\pm.001$
Solar	$RSE\downarrow$	$.201\pm.000$	$.348\pm.000$	$.446\pm.000$	$.496\pm.000$

Appendix

Table 7: Three-seed stability under the standard multivariate protocol. Each entry reports the mean

\pm

sample standard deviation over three independent runs.

Dataset	Variant	96		192		336		720
		MSE	MAE	MSE	MAE	MSE	MAE	MSE	MAE
Traffic	SpikeLite	.453	.285	.468	.287	.481	.293	.514	.311
	w/o FSSE	.463	.291	.479	.292	.491	.298	.524	.315
	w/o SSCA	.615	.345	.581	.327	.589	.330	.624	.349
Weather	SpikeLite	.158	.197	.209	.244	.266	.285	.346	.338
	w/o FSSE	.172	.210	.223	.256	.278	.297	.358	.350

Appendix

Table 8: Horizon-wise ablations under the long-term protocol. Lower MSE and MAE are better.

Figure 5: Frequency diagnostics for FSSE. Panels (a)–(c): real-data component spectra. Panel (d): spectral centroids. Panel (e): initialized and trained decay factors. Panel (f): branch-order violations and decay–centroid concordance. Panels (g)–(h): PEMS sinusoidal fundamental gain and firing probability.

Explore similar work

SpikF-GO: Spiking Fourier Graph Operators for Multivariate Time Series Forecasting
Jun 11, 2026Jafar Bakhshaliyev, Niels LandwehrMultivariate Time Series Forecasting Spiking Neural Networks
Data Science Group, University of Hildesheim, Hildesheim, Germany
LongSpike: Fractional Order Spiking State Space Models for Efficient Long Sequence Learning
Jun 11, 2026Xinrui He, Qiyu Kang, Xuhao Li +1Spiking Neural Networks Time-To-First-Spike
Wuhan University · University of Science and Technology of China · Anhui University
SAFformer:Improving Spiking Transformer via Active Predictive Filtering
May 8, 2026Zequan Xie, Weiming Zeng, Yunhua Chen +3Spiking Neural Networks Transformer Architectures
School of Computer Science and Technology, Guangdong University of Technology, Guangzhou, China · Faculty of Science, Hong Kong Baptist University, China · School of Electronic Information, Wuhan University, China

Jun 11, 2026, cs.LGJ/K move · Enter open · S save

SpikF-GO: Spiking Fourier Graph Operators for Multivariate Time Series Forecasting

Jafar Bakhshaliyev, Niels Landwehr

Data Science Group, University of Hildesheim, Hildesheim, Germany

Multivariate Time Series Forecasting Spiking Neural Networks Time Series Forecasting Time-To-First-Spike Temporal Graph Neural Networks Fourier Neural Operators+1

Open paper AskPDF arXiv

Spiking Neural Networks (SNNs) have emerged as an energy-efficient alternative to conventional neural networks, demonstrating strong performance in computer vision and robotics. More recently, SNNs have been applied to time series forecasting (TSF), with methods exploring spiking temporal backbones, spike-compatible positional encodings, Fourier-domain processing, and redesigned neuron dynamics. However, existing SNN forecasting approaches process variables independently, lacking explicit mechanisms for modeling inter-variable dependencies. This is a critical limitation in multivariate settings, where cross-variable correlations carry substantial predictive information. We propose Spiking Fourier Graph Operators (SpikF-GO), which addresses this gap by combining a hypervariate graph formulation in which every scalar observation becomes a graph node with spike-driven spectral processing. SpikF-GO introduces a Hard Concrete frequency gate for learnable sparse frequency selection and a Complex LIF gate that applies independent spiking neurons to real and imaginary Fourier components, preserving binary, event-driven computation throughout the spectral domain. We further present a variant incorporating Central Pattern Generator-based positional encodings for stronger long-range temporal modeling. Evaluated on eight benchmarks under a unified experimental protocol, SpikF-GO achieves the best average rank among all SNN methods and outperforms its ANN counterpart, FourierGNN, at reduced energy cost. SpikF-GO maintains competitive accuracy even at substantially smaller embedding dimensions, thereby achieving significant energy reductions. To our knowledge, this is among the first works to bring graph-based multivariate modeling into the spiking domain for TSF and the first to provide a unified comparison across SNN forecasting architectures under a common experimental protocol.

Abstract

Explore similar work

SpikF-GO: Spiking Fourier Graph Operators for Multivariate Time Series Forecasting

LongSpike: Fractional Order Spiking State Space Models for Efficient Long Sequence Learning

SAFformer:Improving Spiking Transformer via Active Predictive Filtering