We believe that these <term> evaluation techniques </term> will provide information about both the <term> human language learning process </term> , the <term> translation process </term> and the <term> development </term> of <term> machine translation systems </term> .
The results of this experiment , along with a preliminary analysis of the factors involved in the decision making process will be presented here .
We have built and will demonstrate an application of this approach called <term> LCS-Marine </term> .
The demonstration will focus on how <term> JAVELIN </term> processes <term> questions </term> and retrieves the most likely <term> answer candidates </term> from the given <term> text corpus </term> .
The operation of the <term> system </term> will be explained in depth through browsing the <term> repository </term> of <term> data objects </term> created by the <term> system </term> during each <term> question answering session </term> .
Examples and results will be given for <term> Arabic </term> , but the approach is applicable to any <term> language </term> that needs <term> affix removal </term> .
Despite the successes of these systems , <term> accuracy </term> will always be imperfect .
The tutorial will cover the basics of <term> SMT </term> :
Theory will be put into practice .
<term> STTK </term> , a <term> statistical machine translation tool kit </term> , will be introduced and used to build a working <term> translation system </term> .
The <term> source code </term> of the <term> tool kit </term> will be made available .
In our demonstration at <term> ACL </term> , new <term> users </term> of our <term> tool </term> will drive a <term> syntax-based decoder </term> for themselves .
In this paper , we will present a new <term> evaluation measure </term> which explicitly models <term> block reordering </term> as an <term> edit operation </term> .
Furthermore , we will show how some <term> evaluation measures </term> can be improved by the introduction of <term> word-dependent substitution costs </term> .
The experimental results will show that it significantly outperforms state-of-the-art approaches in <term> sentence-level correlation </term> .
Results from experiments with <term> word dependent substitution costs </term> will demonstrate an additional increase of correlation between <term> automatic evaluation measures </term> and <term> human judgment </term> .
While such <term> decoding </term> is an essential underpinning , much recent work suggests that <term> natural language interfaces </term> will never appear cooperative or graceful unless they also incorporate numerous <term> non-literal aspects of communication </term> , such as robust <term> communication procedures </term> .
Most large <term> text-understanding systems </term> have been designed under the assumption that the input <term> text </term> will be in reasonably neat form , e.g. , <term> newspaper stories </term> and other <term> edited texts </term> .
After introducing this approach to <term> MT system </term> design , and the basics of <term> monolingual UCG </term> , we will show how the two can be integrated , and present an example from an implemented <term> bi-directional Engllsh-Spanish fragment </term> .
Finally we will present some outstanding problems with the approach .
hide detail