David Mimno

I am currently a postdoctoral researcher with David Blei at Princeton. I received my PhD from the University of Massachusetts Amherst, with Andrew McCallum.

Before coming to UMass, I worked for an internet auction startup, the NLP group at the University of Sheffield, and the Perseus Project, a cultural heritage digital library. I have a particular interest in historical texts and languages.

I am currently chief maintainer for the MALLET Machine Learning toolkit. I organized the NorthEast Student Colloquium on Artificial Intelligence (NESCAI) at UMass in 2010 along with Sameer Singh. I ran the UMass Machine Learning and Friends lunch series for two years, and I am currently organizing the Princeton Machine Learning lunch series.

Here are some resources and tools that might be useful:

Slides (and video!) from a presentation at the MITH Digital Humanities topic modeling workshop.
An implementation of LDA in Javascript + D3.
Java implementations of topic model algorithms.
An extensive bibliography on topic modeling.
Example R scripts from Princeton course on probabilistic modeling (COS513)
A model of topic change over five years of ICML.
A topic model of Wikipedia aligned across 40 languages.
A map of the NIH (I provided topic- and word-based distance measurements), profiled in Nature Methods.
Time-series plots of 1000 topics extracted from 20 years of the New York Times.
Classics-related topics in 12 languages from Wikipedia, with a visualization of the relative proportion of each wiki on each topic.