APM
- Parallel alternating Newton-like solver for Admixture of Poisson MRFs (APM) model
This MATLAB code trains an Admixture of Poisson MRFs (APM) model using the algorithm described in the associated paper “Capturing Semantically Meaningful Word Dependencies with an Admixture of Poisson MRFs”. Please see associated papers for more information about the APM model and algorithm.
Download
The code can be cloned/downloaded from this GitHub repository: https://github.com/davidinouye/apmVersion Information
- Version 0.9.1 – Fixed several bugs related to held-out dataset option.
- Version 0.9 – First version
Citation
This software is released under the MIT License but please acknowledge its use with a citation to the following publication:- Capturing Semantically Meaningful Word Dependencies with an Admixture of Poisson MRFs (pdf, poster, software, code)
D. Inouye, P. Ravikumar, I. Dhillon.
In Neural Information Processing Systems (NIPS), pp. 3158-3166, December 2014.
Related Publications
- Admixture of Poisson MRFs: A Topic Model with Word Dependencies (pdf, slides, poster, software, code)
D. Inouye, P. Ravikumar, I. Dhillon.
In International Conference on Machine Learning (ICML), pp. 683-691, June 2014.