This report introduces a study on syntactic features used in a discriminative word alignment model. The features are implemented on a state-of-the-art discriminative word alignment system. The syntactic features are extracted from parse trees. Three types of syntactic features are experimented in this work: one global tree path feature and two first order tree features. Experimental results show that the syntactic features are helpful for improving the word alignment accuracy on Chinese-English parallel sentences.

In this paper we present a statistical transliteration technique that is language independent. This technique uses Hidden Markov Model (HMM) alignment and Conditional

This degree is submitted to the University of Cambridge

Models of word alignment built as sequences of links have limited expressive power, but are easy to decode. Word aligners that model the alignment matrix can express arbitrary alignments, but are difficult to decode. We propose an alignment matrix model as a correction algorithm to an underlying sequencebased aligner. Then a greedy decoding algorithm enables the full expressive power of the alignment matrix formulation. Improved alignment performance is shown for all nine language pairs tested. The improved alignments also improved translation quality from Chinese to English and English to Italian. 1

Structured output prediction in machine learning is the study of learning to predict complex objects consisting of many correlated parts, such as sequences, trees, or matchings. The Structural Support Vector Machine (Structural SVM) algorithm is a discriminative method for structured output learning that allows flexible feature construction with robust control for overfitting. It provides stateof-art prediction accuracies for many structured output prediction tasks in natural language processing, computational biology, and information retrieval. This thesis explores improving the learning of structured prediction rules with structural SVMs in two main areas: incorporating latent variables to extend their scope of application and speeding up the training of structural SVMs with nonlinear kernels. In particular, we propose a new formulation of structural SVM, called Latent Structural SVM, that allows the use of latent variables, and an algorithm to solve the associated non-convex optimization problem. We demonstrate the generality of our new algorithm through several structured output prediction problems, showing improved prediction accuracies with new