Human evaluations of machine translation are extensive but expensive. Human evaluations can take months to finish and involve human labor that can not be reused.

We present a joint probability model for statistical machine translation, which automatically learns word and phrase equivalents from bilingual corpora. Translations produced with parameters estimated using the joint model are more accurate than translations produced using IBM Model 4.

We describe a syntax-based algorithm that automatically builds Finite State Automata (word lattices) from semantically equivalent translation sets. These FSAs are good representations of paraphrases. They can be used to extract lexical and syntactic paraphrase pairs and to generate new, unseen sentences that express the same meaning as the sentences in the input sets. Our FSAs can also predict the correctness of alternative semantic renderings, which may be used to evaluate the quality of translations.

In adding syntax to statistical MT, there is a tradeoff between taking advantage of linguistic analysis, versus allowing the model to exploit linguistically unmotivated mappings learned from parallel training data. A number of previous efforts have tackled this tradeoff by starting with a commitment to linguistically motivated analyses and then finding appropriate ways to soften that commitment. We present an approach that explores the tradeoff from the other direction, starting with a context-free translation model learned directly from aligned parallel text, and then adding soft constituent-level constraints based on parses of the source language. We obtain substantial improvements in performance for translation from Chinese and Arabic to English. 1

A good decoding algorithm is critical to the success of any statistical machine translation system. The decoder's job is to find the translation that is most likely according to set of previously learned parameters (and a formula for combining them). Since the space of possible translations is extremely large, typical decoding algorithms are only able to examine a portion of it, thus risking to miss good solutions. Unfortunately, examining more of the space leads to unacceptably slow decodings. In this

Untranslated words still constitute a major problem for Statistical Machine Translation (SMT), and current SMT systems are limited by the quantity of parallel training texts. Augmenting the training data with paraphrases generated by pivoting through other languages alleviates this problem, especially for the so-called “low density ” languages. But pivoting requires additional parallel texts. We address this problem by deriving paraphrases monolingually, using distributional semantic similarity measures, thus providing access to larger training resources, such as comparable and unrelated monolingual corpora. We present what is to our knowledge the first successful integration of a collocational approach to untranslated words with an end-to-end, state of the art SMT system demonstrating significant translation improvements in a low-resource setting.

We present Minimum Bayes-Risk word alignment for machine translation. This statistical, model-based approach attempts to minimize the expected risk of alignment errors under loss functions that measure alignment quality. We describe various loss functions, including some that incorporate linguistic analysis as can be obtained from parse trees, and show that these approaches can improve alignments of the English-French Hansards.

Examples from chat interaction are presented to demonstrate that machine translation of written interaction shares many problems with translation of spoken interaction. The potential for common solutions to the problems is illustrated by describing operations that normalize and tag input before translation. Segmenting utterances into small translation units and processing short turns separately are also motivated using data from chat.

This dissertation is the result of my own work and includes nothing which is the outcome of work done in collaboration except where specifically indicated in the text. It has not been submitted in whole or in part for a degree at any other university. Some of the work has been published previously in conference proceedings (Blackwood et al., 2008a; Blackwood

Paraphrase generation is useful for various NLP tasks. But pivoting techniques for paraphrasing have limited applicability due to their reliance on parallel texts, although they benefit from linguistic knowledge implicit in the sentence alignment. Distributional paraphrasing has wider applicability, but doesn’t benefit from any linguistic knowledge. We combine a distributional semantic distance measure (based on a non-annotated corpus) with a shallow linguistic resource to create a hybrid semantic distance measure of words, which we extend to phrases. We embed this extended hybrid measure in a distributional paraphrasing technique, benefiting from both linguistic knowledge and independence from parallel texts. Evaluated in statistical machine translation tasks by augmenting translation models with paraphrase-based translation rules, we show our novel technique is superior to the non-augmented baseline and both the distributional and pivot paraphrasing techniques. We train models on both a full-size dataset as well as a simulated “low density ” small dataset. 1