Code

Co-Occurence Data Embedding (CODE)

This package contains a gzipped tar file, with multiple matlab files and one linux binary. Send me an email if you are intersted to run the code on another platform.

Download: the package (gzipped tar file) (14Kb). Browse the main function.

To install: tar -xvzf code.tgz.

How to run: [phi_x,psi_y] = code(pxys,embed_dim,options)
use 'help code' for further instructions.

Reference:
Amir globerson, Gal Chechik, Fernando Pereira and Naftali Tishby
Euclidean Embedding of Co-occurrence Data.
Advances in Neural Information Processing Systems-18, (NIPS 2004).

Relevant papers (citing this paper):
Guobiao Mei and Christian R. Shelton, (2006) Visualization of Collaborative Data . UAI.

Information Bottleneck with Side Information (IBSI)

An implementation of the sequential IB algorithm for IB with side information. The function also supports regular IB, and some specific form of multivairate-IB (see help for details).

Download: the matlab function(12Kb).

How to run: [Ptys,Ptx,trace]= ibsi_sequ(Pxys,parms)
use 'help ibsi_sequ' for further instructions.

Reference:
G. Chechik and N. Tishby
Extracting relevant structures with side information.
Advances in Neural Information Processing Systems-16, (NIPS 2002).


home | publications | cv | code | data | private | book | . . . . . . .