This paper introduces a novel energy minimization method, namely iterated cross entropy with partition strategy (ICEPS), into the Markov random field theory. The solver, which is...
This paper shows how to formally characterize language learning in a finite parameter space as a Markov structure, hnportant new language learning results follow directly: explici...
Uniform random generators deliver a simple empirical means to estimate the average complexity of an algorithm. We present a general rejection algorithm that generates sequential le...
—In this paper we study the network structure in music and attempt to compose music artificially. Networks are constructed with nodes and edges corresponding to musical notes an...
We have recently proposed an EM-style algorithm to optimize log-linear models with hidden variables. In this paper, we use this algorithm to optimize a hidden conditional random ...
Georg Heigold, Stefan Hahn, Patrick Lehnen, Herman...