AAAI'26: EdgeMTSC: A Lightweight Large-Kernel ConvNet for Multivariate Time Series Classification

Project Structure

• appendix.pdf: the additional Appendix of our paper
• data/: the root dir of the datasets
• data//_TRAIN.ts: the samples for training
• data//_TEST.ts: the samples for evaluation
• conf.py: a script for stable configurations
• dataloaders.py: a script for data preprocessing and data loading
• model.py: a script holding the Pytorch models
• utils.py: a script holding project construction tools

Main Parameters

Input python main.py --help, then you will get the following output.

usage: main.py [-h] [--ablation ABLATION] [--dataset DATASET] [--if-valid IF_VALID] [--num-workers NUM_WORKERS] [--batch-size BATCH_SIZE] [--dropout DROPOUT] [--epochs EPOCHS] [--checkpoint-dir CHECKPOINT_DIR] [--device DEVICE]

optional arguments:
  -h, --help            show this help message and exit
  --ablation ABLATION
  --dataset DATASET
  --if-valid IF_VALID
  --num-workers NUM_WORKERS
  --batch-size BATCH_SIZE
  --dropout DROPOUT
  --epochs EPOCHS
  --checkpoint-dir CHECKPOINT_DIR
  --device DEVICE

Run EdgeMTSC

python main.py --dataset ArticularyWordRecognition

Requirements

nvidia-cublas-cu12          12.1.3.1
nvidia-cuda-cupti-cu12      12.1.105
nvidia-cuda-nvrtc-cu12      12.1.105
nvidia-cuda-runtime-cu12    12.1.105
nvidia-cudnn-cu12           8.9.2.26
nvidia-cufft-cu12           11.0.2.54
nvidia-curand-cu12          10.3.2.106
nvidia-cusolver-cu12        11.4.5.107
nvidia-cusparse-cu12        12.1.0.106
nvidia-nccl-cu12            2.20.5
nvidia-nvjitlink-cu12       12.4.127
nvidia-nvtx-cu12            12.1.105

scikit-base                 0.7.8
scikit-image                0.19.3
scikit-learn                1.3.0
scipy                       1.10.1
numpy                       1.23.4
pandas                      2.0.3

timm                        1.0.9
torch                       2.3.0
torch_cluster               1.6.3+pt23cu121
torch_geometric             2.5.3
torch-geometric-temporal    0.54.0
torch_scatter               2.1.2+pt23cu121
torch_sparse                0.6.18+pt23cu121
torch_spline_conv           1.2.2+pt23cu121

If you find it is useful, please cite our paper:

@inproceedings{EdgeMTSC,
  title={EdgeMTSC: A Lightweight Large-Kernel ConvNet for Multivariate Time Series Classification},
  author={Zhou, Xueyi and Li, Zhenyu and Chae, Dong-Kyu},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  volume={40},
  number={19},
  pages={16531--16539},
  year={2026}
}