Justin Domke
's miscellaneous code and notes
Notes
Optimizing over a set of probabilities
(See also
params2probs.m
and
probgrad2paramgrad
.)
Derivatives in matrix form for "logistic" neural networks
(See also,
this
simple set of functions demonstrating these formulas in matlab.)
Other
Gibbs sampling on simple Dependency Networks
weblog
Probabilistic Egomotion
code.
Invariant Histograms
Assorted
Code
that might be useful.
nanoMat
is a very small C++ matrix library
Rushern Baker