In this paper, we survey the current state-ofart models for structured learning problems, including Hidden Markov Model (HMM), Conditional Random Fields (CRF), Averaged Perceptron...
Chord sequences are a compact and useful description of music, representing each beat or measure in terms of a likely distribution over individual notes without specifying the not...
This paper explores the issue of recognizing, generalizing and reproducing arbitrary gestures. We aim at extracting a representation that encapsulates only the key aspects of the ...
We propose and analyze a distribution learning algorithm for variable memory length Markov processes. These processes can be described by a subclass of probabilistic nite automata...
The Hierarchical Hidden Markov Model (HHMM) is a well formalized tool suitable to model complex patterns in long temporal or spatial sequences. Even if effective algorithms are ava...