Skip to content
Resources for SPS
Mathematica
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
Slides
GradDescentGood.mp4
GradientDescent.nb
README.md
SPS_MachineLearning.key
SPS_MachineLearning.pptx
TransferLearningClass.pdf
TransferLearningClass.pptx

README.md

Intro To Machine Learning

This contains information from a talk given to the University or Oregon Society of Physics Students (SPS). The Mathematica notebook shows how Gradient Descent works in the context of a 2-dimensional minimization problem. The associated video shows the minimization happening. There are also two versions of the talk (keynote and power point).

Resources

Online classes

  • This class is a little older, and does the programming in Octave instead of python, but is a great class. This goes over many techniques beyond neural networks.

  • An updated version does things with python and uses some of the standard tools. It focuses more on deep learning.

Python Packages

  • Scikit-Learn makes machine learning very easy.
  • Keras is the package I use for neural networks.

Datasets and challenges

While there is not necessarily much open data in high energy physics, there is a lot of other data to learn from.

  • Kaggle hosts many datasets and some challenges. Users upload their scripts, which is a great resource for learning the techniques. In addition, one of the hosted challenges was to use ATLAS data to find the Higgs!
  • Data Driven is another site which offers challenges and prizes.
  • HackerRank is not necessarily for machine learning, but a great place to practice programming. I highly recommend it. It offers coding challenges for prizes.
  • CERN open data I don't have any experience with either of these open data resources, other than knowing they exist.
  • CMS open data
You can’t perform that action at this time.