Machine Learning

Machine Learning Resources

Overview

Teaching: 5 min
Exercises: 0 min
Questions

  • What are some basic examples of libraries to explore?

Objectives

  • Learn what kind of machine learning resources are available that are not necessarily earmarked for GIS applications

Machine Learning Resources

Computer Vision

  1. opencv

Support Vector Machine/Random Forest/etc

  1. scikit-learn
  2. vowpal wabbit
  3. xgboost

Deep Learning

  1. AWS Deep learning AMI
  2. MXNet
  3. TensorFlow
  4. Caffe
  5. pyTorch (python Torch)
  6. Keras (api for python that plays with TensorFlow, Theano, and others)

ESRI + QGIS

Machine learning capabilities exist in both ESRI & QGIS (support vector machine), however for scaling to large image sets, opensource scripting is nicely adaptable.

Mapbox & Development Seed

(super cool geospatial companies with international offices)

  1. Mapbox on github: they have some awesome tutorials ad open source tools to check out
  2. DevSeed onn github: Related to Mapbox, they do a lot of GIS Deep Learning
  3. robosat
  4. training data from OSM

Basic Discussion of Learning

  1. Gradient Descent & Cost/Loss Functions

Key Points