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..
MoFo: Empowering Long-term Time Series Forecasting with Periodic Pattern Modeling
Authors: Jiaming Ma, Binwu Wang, Qihe Huang, Guanjun Wang, Pengkun Wang, Zhengyang Zhou, Yang Wang
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive empirical results on widely used benchmark datasets demonstrate that Mo Fo achieves competitive performance while maintaining high memory efficiency and fast training speed. |
| Researcher Affiliation | Academia | 1University of Science and Technology of China (USTC), Hefei, Anhui, China 2Suzhou Institute for Advanced Research, USTC, Suzhou, Jiangsu, China EMAIL EMAIL |
| Pseudocode | No | The paper describes methods using mathematical equations and descriptive text, but does not present any structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | Our code is available at official repository . We have open-sourced our code via an anonymous link for reproducibility in the corresponding section. |
| Open Datasets | Yes | Datasets. We conduct our experiments on widely used real-world time series datasets with periodic pattern from 4 different domains, including ETTh1, ETTh2, ETTm1, ETTm2, Weather, Solar Energy, Electricity, and Traffic. A summary of all datasets is provided in Table 1. |
| Dataset Splits | Yes | Split ratio 6:2:2 6:2:2 6:2:2 6:2:2 7:1:2 6:2:2 7:1:2 7:1:2 |
| Hardware Specification | Yes | Our experiments are conducted on an NVIDIA A100 GPU with 40GB memory, using Py Torch under Python 3.11.5. |
| Software Dependencies | Yes | Our experiments are conducted on an NVIDIA A100 GPU with 40GB memory, using Py Torch under Python 3.11.5. |
| Experiment Setup | Yes | Following the evaluation protocol in TFB [34], we report the best performance achieved over look-back window lengths T {96, 336, 512} and forecasting horizons L {96, 192, 336, 720}. Model performance is evaluated using two standard metrics: mean squared error (MSE) and mean absolute error (MAE). To maintain fairness in evaluation, we disable the Drop Last batch-sampling trick[18]. We use the Adam optimizer [14] with the L1 loss function from the Fre DF strategy [53]. |