We consider a new discriminative learning approach to se- quence labeling based on the statistical concept of the Z -score. Given a training set of pairs of hidden-observed sequences, the task is to deter- mine some parameter values such that the hidden labels can be correctly reconstructed from observations. Maximizing the Z -score appears to be a very good criterion to solve this problem both theoretically and empir- ically. We show that the Z -score is a convex function of the parameters and it can be efficiently computed with dynamic programming methods. In addition to that, the maximization step turns out to be solvable by a simple linear system of equations. Experiments on artificial and real data demonstrate that our approach is very competitive both in terms of speed and accuracy with respect to previous algorithms.
Discriminative Sequence Labeling by Z-Score Optimization
RICCI, ELISA;
2007
Abstract
We consider a new discriminative learning approach to se- quence labeling based on the statistical concept of the Z -score. Given a training set of pairs of hidden-observed sequences, the task is to deter- mine some parameter values such that the hidden labels can be correctly reconstructed from observations. Maximizing the Z -score appears to be a very good criterion to solve this problem both theoretically and empir- ically. We show that the Z -score is a convex function of the parameters and it can be efficiently computed with dynamic programming methods. In addition to that, the maximization step turns out to be solvable by a simple linear system of equations. Experiments on artificial and real data demonstrate that our approach is very competitive both in terms of speed and accuracy with respect to previous algorithms.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
