This is partially for my own reference, partially for your reference if you find them useful.
Want a good overview for why certain components are commonly used in neural networks (e.g. dropout layers, data augmentation)? Check out this paper. It’s a mostly-entry level explanation.
Want a quick way to get started using neural networks in Python? Check out keras.
Python in Astronomy Demos
If you’ve never used Python before: https://sites.google.com/site/lamatbootcamp2016/
If you want to learn more about Python, at UCSC we hold occasional discussion + demos of Python tools we use to be more productive. Our main repo of past discussions is github/ucsc-astro/coffee. I’ve specifically talked about:
- ipywidgets + Bokeh for interactive plots: https://gist.github.com/egentry/deeb8f427be765e606c7
- [embarassingly] parallel programming: https://github.com/egentry/AstroComputingSeminar_Parallel
Parallel Programming Basics
- Using map in Python (embarrassingly parallelizable): http://chriskiehl.com/article/parallelism-in-one-line/
- Basic parallel computation in C: https://www.gribblelab.org/CBootCamp/A2_Parallel_Programming_in_C.html
- I’ve found it helpful to follow Jake VanderPlas’s blog (e.g. see his discussion on how to make “real” code run fast in Python: https://jakevdp.github.io/blog/2015/02/24/optimizing-python-with-numpy-and-numba/)
Previously I had to get up to speed in Modern Fortran. Since it’s not a still-commonly-used language, here are the resources that were useful for learning it quickly:
Finding [Astronomy] Research papers
- Knowles Science Teaching Fellowship (substantial 5-year fellowship for new teachers, which can be used for teaching at any school)
- Teaching Fellowship Programs (compiled by Oberlin)