A gallery of demonstrators, use cases, and videos to help you learn more about DeepSensor, a Python package for modelling environmental data with neural processes (NPs).
These notebook demonstrators walk through the main functionality of DeepSensor with some example datasets. For set-up instructions, see below.
| Title | Content | Binder |
|---|---|---|
| Quick start | 💻 | - |
| Task loader tour | 💻 | - |
| Saving and loading | 💻 | - |
| Custom plotting | 💻 | - |
| Active learning acquisition functions | 💻 | - |
| Statistical downscaling | 💻 | - |
| Autoregressive sampling | 💻 | - |
| Extending models | 💻 | - |
| The DeepSensor interface | 💻 | - |
| Multi-output training | 💻 | - |
To run the notebook demonstrators, first set up the Python environment with DeepSensor, PyTorch, and Cartopy. Note: DeepSensor can be used with TensorFlow instead of PyTorch, but PyTorch is chosen for these demonstrators.
We use conda to manage the environment because
it handles the third-party dependencies for Cartopy. If you don't yet have conda, you can download it
here.
We also recommend using mamba for faster environment creation, which can be installed with
conda install mamba -n base -c conda-forge.
After cloning the repo, run the commands below in the root of the repository to set up the conda environment:
mamba env create --file environment.ymlconda activate deepsensor
These notebooks showcase applications of DeepSensor to real-world research problems.
| Title | Content | Binder |
|---|---|---|
| - | - |
User contributions that showcase applications of DeepSensor to real-world research problems are very welcome! To contribute a use case notebook, please follow the instructions below:
- TODO
| Date | Title | Presenter | Length | Video |
|---|---|---|---|---|
| August 2023 | Tackling diverse environmental prediction tasks with neural processes | Tom Andersson | 1 hour | 🎥 / slides |
| April 2023 | Environmental Sensor Placement with ConvGNPs | Tom Andersson | 15 mins | 🎥 |
| Jul 2022 | Advances in Neural Processes | Richard Turner | 1 hour | 🎥 |
| May 2023 | Autoregressive Conditional Neural Processes | Wessel Bruinsma | 5 mins | 🎥 |
- Tom Andersson et al. Environmental Sensor Placement with Convolutional Gaussian Neural Processes. Environmental Data Science (2023)
- Wessel Bruinsma et al. Autoregressive Conditional Neural Processes. In Proceedings of the 11th International Conference on Learning Representations, ICLR (2023)
- Anna Vaughan et al. Convolutional conditional neural processes for local climate downscaling. Geoscientific Model Development (2022)
- Yann Dubois' Neural Process Family website