Toward a general, scaleable framework for Bayesian teaching with applications to topic models
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
ldateach
src
README.md
license.txt
requirements.txt
setup.py

README.md

Teaching topic models

Code for our submission to the 2016 IJCAI workshop on Interactive Machine Learning: Connecting Humans and Machines, "Toward a general, scaleable framework for Bayesian teaching with applications to topic models" (ArXiv Preprint).

Installation

Add the install directory to your PYTHONPATH and execute

python setup.py develop

To run tests:

python setup.py test

Tested on OSX, Ubuntu, and REHL.

Use

Run some of the scripts in the ldateach/experiments directory. Some of these may take more time than you are willing to wait, so feel free to tweak the number of runs, samples, etc to bring the runtime down.

License

GNU general public license v3 (GPLv3)