Common use of Setup Clause in Contracts

Setup. We evaluate our approach on Chinese-English alignment and translation tasks. The training corpus consists of 1.2M sentence pairs with 32M Chinese words and 35.4M English words. We used the SRILM toolkit (▇▇▇▇▇▇▇, 2002) to train a 4-gram language model on the Xinhua portion of the English GIGAWORD cor- pus, which contains 398.6M words. For alignment evaluation, we used the Tsinghua Chinese-English word alignment evaluation data set.1 The evalu- ation metric is alignment error rate (AER) (Och and Ney, 2003). For translation evaluation, we used the NIST 2006 dataset as the development set and the NIST 2002, 2003, 2004, 2005, and 2008 datasets as the test sets. The evaluation metric is case-insensitive BLEU (▇▇▇▇▇▇▇▇ et al., 2002). We used both phrase-based (▇▇▇▇▇ et al., 2003) and hierarchical phrase-based (Chiang, 2007) translation systems to evaluate whether our approach improves translation performance. For the phrase-based model, we used the open-source toolkit Moses (▇▇▇▇▇ and ▇▇▇▇▇, 2007). For the hierarchical phrase-based model, we used an in- house re-implementation on par with state-of-the- art open-source decoders. We compared our approach with two state-of- the-art generative alignment models: 1. GIZA++ (Och and Ney, 2003): unsupervised training of IBM models (▇▇▇▇▇ et al., 1993) and the HMM model (▇▇▇▇▇ et al., 1996) us- ing EM, 2. BERKELEY (Liang et al., 2006): unsuper- vised training of joint HMMs using EM. For GIZA++, we trained IBM Model 4 in two directions with the default setting and used the grow-diag-final heuristic to generate symmetric alignments. For BERKELEY, we trained joint HMMs using the default setting. The hyper- parameter of posterior decoding was optimized on the development set. We used first-order HMMs for both word alignment and phrase segmentation. Our joint alignment and segmentation model were trained using the Viterbi EM algorithm for five iterations. Note that the Chinese-to-English and English-to- Chinese alignments are generally non-identical but share many links (see Figure 1(c)). Then, we used the grow-diag-final heuristic to generate symmetric alignments.