Hello!

[leifdenby]

Hi everyone :smile:

I'll be running the session on unsupervised learning on Wednesday and I would like to get an idea of who you all are and what your plans are with ML. So if you could put in your name, and short description of the work you're planning/are doing that would be great. Thanks! (feel free to add links, pictures, etc)

If you could tell me what kind of operating system you will be using that would be great too (i.e. mac, windows or linux).

[leifdenby]

I'll start: I primarily study the development of convective clouds in the Earth's atmosphere, and I do this using Large-Eddy Simulations funded in the GENESIS project (http://homepages.see.leeds.ac.uk/~earlcd/GENESIS/, https://github.com/leifdenby/genesis/). In the last few years I've been using unsupervised machine learning study convective cloud organisation in the EUREC4A project (http://eurec4a.uk/, http://eurec4a.eu/). In EUREC4A I'm also using neural networks to do super-resolution and noise reduction on observation data :earth_americas: :artificial_satellite: :small_airplane:

[scatr]

Hi all, I'm Alasdair from the fluid dynamics CDT based in the school of computing. I'm using machine learning methods to try to discover dynamical systems and reduced order models from data. Eventually I will be applying these ML methods to experimental data obtained from Tokamaks. The main method I'm using is called the "Sparse Identification of Nonlinear Dynamics" or SINDy for short. There's a Python implementation of this: https://pysindy.readthedocs.io/en/latest/ I'm also interested in learning more about neural nets and some different applications of AI/ML.

To Leif, I'm using Windows.

[jdconey]

Hi everyone, I'm Jonathan (Jon/Jono etc) from the Panorama DTP in the School of Earth and Environment, Leeds. I'm studying the influence of orography on the atmosphere and this representation in weather models. The plan is to use ML techniques to try to improve the representation of orographic effects in weather models, especially those where the orography is partially or unresolved (most likely gravity wave drag - which slows down the global atmospheric circulation, which is important in global climate models so that the atmospheric flow is correct). Looking forward to a good, interesting week, learning lots and working with you all 😃 I'm on Windows.

[Spiruel]

Hi all - I'm Sam, based at University of Leeds and part of the SENSE CDT. My project is all about assessing food production using EO data and machine learning. A fundamental part of my project will be to generate new crop-type masks from Sentinel 1/2 imagery, using deep learning models that exploit weakly-supervised approaches that reduce our reliance on training with detailed ground truth data. I'm running Linux :)

[pahdsn]

Hello,

I'm Phoebe. I'm another SENSE CDT student but I'm based at the National Oceanography Center (Southampton) but am registered at Edinburgh. I'm studying the Changing Arctic Water Cycle. I don't yet have a plan to use ML in my project and am very new to it but would definitely like to incorporate it - I'm just still working on where. I also plan to predominantly work with netcdf data rather than optical imagery so am very interested in your idea about a package where you can feed in an xarray / netcdf dataset - and feel like this would be a really cool and valuable tool! I'm on windows.

[jconnollyeo]

Hi everyone, I'm Jacob -- based at the University of Leeds as part of the SENSE CDT. My project is based around trying to improve InSAR methods with respect to handling deformation bias caused by variations in soil moisture and vegetation, and also to improve the handling of decorrelation in interferograms. Eventually I'm planning on using ML as part of this. I'm running Linux