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..

Enhancing the Maximum Effective Window for Long-Term Time Series Forecasting

Authors: Jiahui Zhang, Zhengyang Zhou, Wenjie Du, Yang Wang

NeurIPS 2025 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We integrated these two modules into multiple Transformer-based models and conducted detailed experiments on seven datasets. The results demonstrate that these modules can enhance MEW while achieving better performance.
Researcher Affiliation Academia Jiahui Zhang1,2, Zhengyang Zhou1,2, Wenjie Du1,2, , Yang Wang1,2, 1University of Science and Technology of China, China 2Suzhou Institute for Advanced Research, USTC, China EMAIL EMAIL
Pseudocode No The paper describes the methodology using textual explanations and mathematical formulations, but it does not include any explicit pseudocode or algorithm blocks.
Open Source Code Yes Our code is available at https://github.com/forever-ly/PIH.
Open Datasets Yes We evaluate PIH on seven popular datasets, including Weather, Traffic, Electricity, and four ETT datasets (ETTh1, ETTh2, ETTm1, ETTm2). We also provide results on the Solar and PEMS datasets in Appendix C.1.
Dataset Splits No The paper states in its NeurIPS Paper Checklist that data splits are provided: 'Question: 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 have provided the data splits, hyperparameters, etc.' However, within the provided text, specific details such as split percentages, sample counts, or explicit references to standard dataset splits are not present.
Hardware Specification No The AI-driven experiments, simulations and model training were performed on the robotic AI-Scientist platform of Chinese Academy of Sciences. Table 8 shows 'Comparison of GPU memory usage and training time per epoch', but does not specify the exact GPU models (e.g., NVIDIA A100, Tesla V100) or CPU models used.
Software Dependencies No The paper does not explicitly list any software dependencies with specific version numbers (e.g., Python 3.8, PyTorch 1.9, CUDA 11.1).
Experiment Setup Yes Our model incorporates several crucial hyperparameters, including K, which determines the number of partitions; β, which governs the balance between prediction and compression in the information bottleneck (IB) objective; and the temperature factor τ, which influences subsequence sampling. We set K {2, 4}, β {0.0001, 0.001, 0.1, 1}, and τ {0.1, 0.5, 1, 2}. We selected the optimal hyperparameters based on the results from the validation set.