Resources:
https://www.indico.kr/event/47/contributions/532/attachments/527/1217/RS_LLRF-Denoising.pdf https://github.com/udacity/deep-learning-v2-pytorch/blob/master/autoencoder/denoising-autoencoder/Denoising_Autoencoder_Solution.ipynb https://medium.com/@syed_hasan/autoencoders-theory-pytorch-implementation-a2e72f6f7cb7
Task:
Develop a denoising method for magnetic field data like hall sensors, and deliver the model in a form that is immediately usable or replicable for downstream tasks.
Use:
Acc-Py Python 3.11 base distribution (2023.06) PyTorch >= 2.0 PyTorch Lightning >= 2.0
Optional:
LightningCLI for modularity
torch.compile for optimization
tensorboard for visualization
Tips
Requirements:
Deliverables
- The outcome of the project should ideally be a Python project committed to CERN GitLab, and the project should be structured in a Python package form. Example use of the code and/or notebooks would be helpful.
- The outcome of the project should be documented in a PDF form (using ATS technical note template) to be published by summer student coordinators to CDS.
- A talk can optionally be given to MLCF.
- A student session poster with the summer students