emnlp emnlp2012 emnlp2012-95 knowledge-graph by maker-knowledge-mining
Source: pdf
Author: Zhengxian Gong ; Min Zhang ; Chew Lim Tan ; Guodong Zhou
Abstract: Tense is a small element to a sentence, however, error tense can raise odd grammars and result in misunderstanding. Recently, tense has drawn attention in many natural language processing applications. However, most of current Statistical Machine Translation (SMT) systems mainly depend on translation model and language model. They never consider and make full use of tense information. In this paper, we propose n-gram-based tense models for SMT and successfully integrate them into a state-of-the-art phrase-based SMT system via two additional features. Experimental results on the NIST Chinese-English translation task show that our proposed tense models are very effective, contributing performance improvement by 0.62 BLUE points over a strong baseline. 1
Reference: text
sentIndex sentText sentNum sentScore
1 sg Abstract Tense is a small element to a sentence, however, error tense can raise odd grammars and result in misunderstanding. [sent-9, score-0.952]
2 Recently, tense has drawn attention in many natural language processing applications. [sent-10, score-0.93]
3 They never consider and make full use of tense information. [sent-12, score-0.93]
4 In this paper, we propose n-gram-based tense models for SMT and successfully integrate them into a state-of-the-art phrase-based SMT system via two additional features. [sent-13, score-0.949]
5 Experimental results on the NIST Chinese-English translation task show that our proposed tense models are very effective, contributing performance improvement by 0. [sent-14, score-0.966]
6 1 Introduction Ø For many NLP applications, such as event extraction and summarization, tense has been regarded as a key factor in providing temporal order. [sent-16, score-0.963]
7 However, tense information has been largely overlooked by current SMT research. [sent-17, score-0.93]
8 276 Although the translated text produced by Moses1 is understandable, it has very odd tense combination from the grammatical aspect, i. [sent-24, score-0.974]
9 So it is not a random phenomenon that tense can affect translation results. [sent-45, score-0.966]
10 The key is how to detect tense errors and choose correct tenses during the translation procedure. [sent-46, score-1.076]
11 lc L2a0n1g2ua Agseso Pcrioactieosnsi fnogr a Cnodm Cpoumtaptiuotna tilo Lnianlg Nuaist uircasl the subordinate clause should have the consistent tense with its main clause while Moses fails; Observation(3): the diversity of tense usage in a document is common. [sent-52, score-2.018]
12 In some extreme examples, one tense (past or present), even dominates the whole document. [sent-54, score-0.93]
13 Dorr (2002) refers to six basic English tense structures and defines the possible paired combinations of “present, past, future”. [sent-56, score-0.93]
14 For each target sentence in the training corpus, we first parse it and ex- tract its tense sequence. [sent-61, score-0.972]
15 Such model can be used to estimate the rationality of tense combination in a sentence and thus supervise SMT to reduce tense inconsistency errors against Observations (1) and (2) in the sentence-level. [sent-63, score-1.92]
16 In comparison, Observation (3) actually reflects the tense distributions among one document. [sent-64, score-0.93]
17 After extracting each main tense for each sentence, we build another tense ngram model in the document-level. [sent-65, score-1.895]
18 For clarity, this paper denotes document-level tense as “inter-tense” and sentence-level tense as “intra-tense”. [sent-66, score-1.86]
19 After that, we propose to integrate such tense models into SMT systems in a dynamic way. [sent-67, score-0.949]
20 During decoding, once some words with correct tense can be determined, with the help of lan- guage model and other related features, the smallcomponent–“tense”–can affect surrounding words and improve the performance of the whole sentence. [sent-72, score-0.93]
21 Section 3 describes the preprocessing work for building tense models. [sent-81, score-0.93]
22 Section 4 presents how to build target-side tense models and discuss their characteristics. [sent-82, score-0.93]
23 Section 5 shows our way of integrating such tense models into a SMT system. [sent-83, score-0.93]
24 2 Related Work In this section, we focus on related work on integrating the tense information into SMT. [sent-86, score-0.93]
25 Since both interand intra-tense models need to analyze and extract tense information, we also give a brief overview on tense prediction (or tagging). [sent-87, score-1.86]
26 1 Tense Prediction The tense prediction task often needs to build a model based on a large corpus annotated with temporal relations and thus its focus is on how to recognize, interpret and normalize time expressions. [sent-89, score-0.963]
27 Some languages, such as English, are inflectional, whose verbs can express tense via certain stems or suffix, while others, such as Chinese of- ten lack inflectional forms. [sent-93, score-0.991]
28 Ye and Zhang (2005) first built a small amount of manually-labeled data, which provide the tense mapping from Chinese text to English text. [sent-99, score-0.93]
29 Then, they trained a CRF-based tense classifier to label tense on Chinese documents. [sent-100, score-1.86]
30 In particular, they used English verbs to label tense information for Chinese verbs via a parallel Chinese-English corpus. [sent-105, score-1.052]
31 It is reasonable to label such source-side verb to supervise the translation process since the tense of English sentence is often determined by verbs. [sent-106, score-1.045]
32 The problem is that due to the diversity of English verb inflection, it is difficult to map such Chinese tense information into the English text. [sent-107, score-0.967]
33 In particular, they addressed tense reconstruction on a binary taxonomy (present and past) for Chinese text and reported that incorporating lexical aspect features of telicity can obtain a 20% to 35% boost in accuracy on tense realization. [sent-114, score-1.901]
34 (2006) showed that incorporating latent features into tense classifiers can boost the performance. [sent-116, score-0.93]
35 They reported the tense resolution results based on the best-ranked translation text produced by a SMT system. [sent-117, score-0.966]
36 However, they did not report the variation of translation performance after introducing tense information. [sent-118, score-0.966]
37 278 3 Preprocessing for Tense Modeling In this paper, tense modeling is done on the targetside language. [sent-119, score-0.93]
38 Since our experiments are done on Chinese to English SMT, our tense models are learned only from the English text. [sent-120, score-0.93]
39 Furthermore, we assume that one sentence has only one main tense but maybe has many subordinate tenses. [sent-125, score-1.046]
40 This section describes the preprocessing work of building tense models, which mainly involves how to extract tense sequence via tense verbs. [sent-126, score-2.807]
41 Inspired by their work, we use the Stanford parser3 to parse tense sequence for each sentence. [sent-130, score-0.971]
42 Here, the level structures mainly contribute to extract main tense for each sentence (to be described in Section 3. [sent-138, score-0.983]
43 2) and POS tags are utilized to detect tense verbs, i. [sent-139, score-0.93]
44 In addition, the verbs with POS tag “MD” need to be specially considered to distinguish future tense from other tenses. [sent-147, score-1.009]
45 Algorithm 1 illustrates how to determine what tense a node has. [sent-148, score-0.947]
46 If the return value is not “UNK”, the node belongs to a tense verb. [sent-149, score-0.963]
47 2 Tense Extraction Based on Tense Verbs As described in Section 1, the inter-tense (document-level) refers to the main tense of one sentence and the intra-tense (sentence-level) corresponds to all tense sequence of one sentence. [sent-152, score-1.93]
48 This section introduces how to recognize the main tense and extract all useful tense sequence for each sentence. [sent-153, score-1.912]
49 The idea of determining the main tense is to find the tense verb located in the top level of a parse tree. [sent-154, score-1.977]
50 According to the Penn Treebank style, the method to determine the main tense can be described as follows: (1) Traverse the parse tree top-down until a tree node containing more than one child is identified, denoted as Sm . [sent-155, score-1.006]
51 (2) Consider each child of Sm with tag “VP”, recursively traverse such “VP” node to find a tense verb. [sent-156, score-0.965]
52 If found, use it as the main tense and return the tense; if not, go to step (3). [sent-157, score-0.981]
53 Starting from the first subordinate clause, apply the similar policy of step (2) to find the tense verb. [sent-159, score-0.993]
54 (4) If no tense verb found, return “UNK” as the main tense. [sent-161, score-1.018]
55 This is to ensure that the main tense is decided by the top-level tense verb. [sent-163, score-1.895]
56 Take Figure 1 as an example, the main tense of sentence (a) and (b) can be determined only by step (2). [sent-164, score-0.983]
57 The tense verb of “(VBZ renounces)” dominated in the “VP” tag determines that (a) is in present tense. [sent-165, score-1.003]
58 The node “(VBD said)” located in the first subordinate clause shows its main tense is “past”. [sent-168, score-1.087]
59 The next task is to extract the tense sequence for each sentence. [sent-169, score-0.947]
60 They are determined by all tense verbs in this sentence according to the strict topdown order. [sent-170, score-1.009]
61 For example, the tense sequence of sentence (a), (b) and (c) are {present, present}, {present, future, past} a (ncd) {past, past, past}. [sent-171, score-0.965]
62 sIenn ord{perr etsoe explore we,h peatshte}r atnhed m {paainst tense ,o pfa isnt}tr. [sent-172, score-0.93]
63 So the tense sequence marked with main tense of (a), (b) and (c) are {present*, present},{present, future, past*} da (ncd) {past*, past, past}. [sent-174, score-1.912]
64 4) sb waseodrt on othtien gla,t ttheer tense sequence is different to the former. [sent-176, score-0.947]
65 1 Tense N-gram Estimation After applying the previous method to extract tense for an English text corpus, we can obtain a big tense corpus. [sent-178, score-1.86]
66 Given the current tense is indexed as ti, we call the previous n − 1 tenses plus the current tense as ttehense p n-gram. [sent-179, score-1.97]
67 Based on the tense corpus, tense n-gram statistics can be done according to the Formula 1. [sent-180, score-1.86]
68 ,t(i−1)) t(i−1)) count(t(i−(n−1)) (1) Here, the function of “count” return the tense n-gram frequency. [sent-189, score-0.946]
69 To compute the probability of intra-tense n-gram, we first extract all tense sequence for each sentence 280 and put them into a new file. [sent-191, score-0.965]
70 To compute the probability of inter-tense n-gram, we need to extract the main tense for each sentence at first. [sent-193, score-0.983]
71 The horizontal axis indicts tense type, and the vertical axis shows its probabilities. [sent-202, score-1.048]
72 The parts (a) and (c) also indicate “present” and “past” are two main tense types in news domain. [sent-203, score-0.965]
73 The vertical axis shows the transfer possibilities from a history tense to a current tense. [sent-207, score-0.984]
74 For example, in one sentence, the probability of co-occurrence of “present → present”, “past → past” and “future → present” i→s more tnht”an, potahsetr → combinations, uwtuhreich → can bseen against tense inconsistency errors described in Observation (1) and (2) (see Section 1). [sent-210, score-0.948]
75 The part (d) shows tense type can be switched between two neighbored sentences. [sent-217, score-0.981]
76 However, it shows the strong tendency to use the same tense type for 4The co-occurrence of the “UNK” tense and other tense types in one sentence cannot happen, so the “UNK” tense is omitted. [sent-218, score-3.738]
77 5 Integrating N-gram-based Tense Models into SMT In this section, we discuss how to integrate the previous tense models into a SMT system. [sent-221, score-0.949]
78 2 The Workflow of Our System Our system works as follows: When a hypothesis has covered all source-side words during the decoding procedure, the decoder 281 first obtains tense sequence for such hypothesis and computes intra-tense feature Fs(see Section 5. [sent-229, score-0.979]
79 At the same time, it recognizes the main tense of this hypothesis and associate the main tense of previous sentence to compute inter-tense feature Fm (see Section 5. [sent-231, score-1.948]
80 After translating one sentence, the decoder caches its main tense and pass it to the next sentence. [sent-234, score-0.997]
81 In order to successfully implement above workflow, we should firstly design some related features, then resolve another key problem of determining tense (especially main tense) for SMT output. [sent-236, score-0.965]
82 3 Two Additional Features Although the previous tense models show strong tendency to use the consistent tenses for one sentence or one document, other tense combinations also can be permitted. [sent-241, score-1.988]
83 )I to ifs f owrmortuhl noting bthee efistrsimt saetentden byce t hoef one document often scares tense information since it corresponds to the title at most cases. [sent-252, score-0.974]
84 For these abnormal translated texts, the syntactic parser cannot work well in our initial experiments, so the previous method to parse main tense and tense sequence of regular texts cannot be applied here too. [sent-265, score-1.984]
85 Once obtaining a completed hypothesis, the decoder will pass it to the Stanford POS tagger and according to tense verbs to get all tense sequence for this hypothesis. [sent-269, score-2.007]
86 However, since the POS tagger can not return the information about level structures, the decoder cannot recognize the main tense from such tense sequence. [sent-270, score-1.962]
87 (201 1) once used target-side verbs to label tense of source-side verbs. [sent-272, score-0.991]
88 282 Figure 3 : trees for parallel sentences main tense for SMT output according to such tense verb, which corresponds to the “VV”(Chinese POS labels are different to English ones, “VV” refers to Chinese verb) node in the top level ofthe source-side parse tree. [sent-277, score-1.936]
89 Take Figure 3 as an example, the English node “(VBD announced)” is a tense verb which can tell the main tense for this English sentence. [sent-278, score-1.949]
90 Once a completed hypothesis is generated, according to the phrase alignment information, the decoder can map Sarea into the English location Tarea and obtain the main tense according to the POS tags in Tarea . [sent-281, score-0.997]
91 If Tarea does not contain tense verb, such as the verb POS tags in the list of {VB, VBN, VBG}, vwehrbich P cannot tsell i tense type by themselves, our system permits to find main tense in the left/right 3 words neighbored to Tarea. [sent-282, score-2.913]
92 This experiment is to check whether the main tense has an impact on intra-tense model or not (see Section 3. [sent-319, score-0.965]
93 Here, the intra-tense model based on the tense sequence with main tense marker is slightly different to the model showed in Figure 2. [sent-321, score-1.938]
94 From these examples, we can see not only tense verbs have changed but also their surrounding words have subtle variation. [sent-330, score-0.991]
95 The part (a) of Figure 4 shows the main tense distributions of one reference. [sent-333, score-0.965]
96 The main tense distributions for the baseline and our proposed system are showed in the part (b) and (c) respectively. [sent-334, score-0.965]
97 Table 5: the joint impact of inter- tense and intra-tense models on SMT and 2) are two neighbored sentences in one document. [sent-358, score-0.981]
98 Both the reference and our output tend tence 1) to use the same main tense type, but the former is in “past” tense and the latter is in “present” tense. [sent-359, score-1.91]
99 Although our main tense is different to the reference one, the consistent tenses in document level bring better translation results than the baseline. [sent-361, score-1.144]
100 In particular, we focus on how to build document-level and sentence-level tense models and how to integrate such models into a popular SMT system. [sent-364, score-0.949]
wordName wordTfidf (topN-words)
[('tense', 0.93), ('smt', 0.134), ('tenses', 0.11), ('past', 0.083), ('subordinate', 0.063), ('verbs', 0.061), ('unk', 0.061), ('chinese', 0.057), ('neighbored', 0.051), ('ye', 0.049), ('lpostag', 0.04), ('palestinian', 0.039), ('verb', 0.037), ('translation', 0.036), ('fm', 0.035), ('leader', 0.035), ('main', 0.035), ('axis', 0.034), ('moses', 0.033), ('fs', 0.033), ('temporal', 0.033), ('vv', 0.032), ('decoder', 0.032), ('aspectual', 0.03), ('bethlehem', 0.03), ('indicts', 0.03), ('lword', 0.03), ('protest', 0.03), ('tarea', 0.03), ('ref', 0.029), ('israeli', 0.027), ('pos', 0.027), ('authorities', 0.026), ('abnormal', 0.026), ('leafnode', 0.026), ('road', 0.026), ('west', 0.026), ('noting', 0.026), ('special', 0.026), ('marker', 0.026), ('bleu', 0.025), ('modification', 0.024), ('sm', 0.024), ('parse', 0.024), ('quoted', 0.024), ('war', 0.024), ('supervise', 0.024), ('nist', 0.023), ('vp', 0.023), ('translated', 0.022), ('formula', 0.022), ('file', 0.022), ('odd', 0.022), ('gong', 0.022), ('attack', 0.022), ('mt', 0.022), ('english', 0.021), ('aspect', 0.021), ('clause', 0.021), ('located', 0.021), ('vertical', 0.02), ('srilm', 0.02), ('attended', 0.02), ('ceremony', 0.02), ('ebest', 0.02), ('embargo', 0.02), ('mortars', 0.02), ('olsen', 0.02), ('partnership', 0.02), ('renounces', 0.02), ('telicity', 0.02), ('towns', 0.02), ('och', 0.02), ('integrate', 0.019), ('tagger', 0.019), ('sentence', 0.018), ('dominated', 0.018), ('kong', 0.018), ('inconsistency', 0.018), ('document', 0.018), ('tag', 0.018), ('else', 0.018), ('obtaining', 0.018), ('vbd', 0.018), ('node', 0.017), ('workflow', 0.017), ('sri', 0.017), ('ncd', 0.017), ('plm', 0.017), ('cold', 0.017), ('sequence', 0.017), ('dorr', 0.016), ('franz', 0.016), ('src', 0.016), ('zhengxian', 0.016), ('return', 0.016), ('hong', 0.016), ('bank', 0.016), ('ti', 0.015), ('reference', 0.015)]
simIndex simValue paperId paperTitle
same-paper 1 0.99999958 95 emnlp-2012-N-gram-based Tense Models for Statistical Machine Translation
Author: Zhengxian Gong ; Min Zhang ; Chew Lim Tan ; Guodong Zhou
Abstract: Tense is a small element to a sentence, however, error tense can raise odd grammars and result in misunderstanding. Recently, tense has drawn attention in many natural language processing applications. However, most of current Statistical Machine Translation (SMT) systems mainly depend on translation model and language model. They never consider and make full use of tense information. In this paper, we propose n-gram-based tense models for SMT and successfully integrate them into a state-of-the-art phrase-based SMT system via two additional features. Experimental results on the NIST Chinese-English translation task show that our proposed tense models are very effective, contributing performance improvement by 0.62 BLUE points over a strong baseline. 1
2 0.10414266 80 emnlp-2012-Learning Verb Inference Rules from Linguistically-Motivated Evidence
Author: Hila Weisman ; Jonathan Berant ; Idan Szpektor ; Ido Dagan
Abstract: Learning inference relations between verbs is at the heart of many semantic applications. However, most prior work on learning such rules focused on a rather narrow set of information sources: mainly distributional similarity, and to a lesser extent manually constructed verb co-occurrence patterns. In this paper, we claim that it is imperative to utilize information from various textual scopes: verb co-occurrence within a sentence, verb cooccurrence within a document, as well as overall corpus statistics. To this end, we propose a much richer novel set of linguistically motivated cues for detecting entailment between verbs and combine them as features in a supervised classification framework. We empirically demonstrate that our model significantly outperforms previous methods and that information from each textual scope contributes to the verb entailment learning task.
3 0.089216448 12 emnlp-2012-A Transition-Based System for Joint Part-of-Speech Tagging and Labeled Non-Projective Dependency Parsing
Author: Bernd Bohnet ; Joakim Nivre
Abstract: Most current dependency parsers presuppose that input words have been morphologically disambiguated using a part-of-speech tagger before parsing begins. We present a transitionbased system for joint part-of-speech tagging and labeled dependency parsing with nonprojective trees. Experimental evaluation on Chinese, Czech, English and German shows consistent improvements in both tagging and parsing accuracy when compared to a pipeline system, which lead to improved state-of-theart results for all languages.
4 0.078656927 35 emnlp-2012-Document-Wide Decoding for Phrase-Based Statistical Machine Translation
Author: Christian Hardmeier ; Joakim Nivre ; Jorg Tiedemann
Abstract: Independence between sentences is an assumption deeply entrenched in the models and algorithms used for statistical machine translation (SMT), particularly in the popular dynamic programming beam search decoding algorithm. This restriction is an obstacle to research on more sophisticated discourse-level models for SMT. We propose a stochastic local search decoding method for phrase-based SMT, which permits free document-wide dependencies in the models. We explore the stability and the search parameters ofthis method and demonstrate that it can be successfully used to optimise a document-level semantic language model. 1 Motivation In the field oftranslation studies, it is undisputed that discourse-wide context must be considered care- fully for good translation results (Hatim and Mason, 1990). By contrast, the state of the art in statistical machine translation (SMT), despite significant advances in the last twenty years, still assumes that texts can be translated sentence by sentence under strict independence assumptions, even though it is well known that certain linguistic phenomena such as pronominal anaphora cannot be translated correctly without referring to extra-sentential context. This is true both for the phrase-based and the syntaxbased approach to SMT. In the rest of this paper, we shall concentrate on phrase-based SMT. One reason why it is difficult to experiment with document-wide models for phrase-based SMT is that the dynamic programming (DP) algorithm 1179 which has been used almost exclusively for decoding SMT models in the recent literature has very strong assumptions of locality built into it. DP beam search for phrase-based SMT was described by Koehn et al. (2003), extending earlier work on word-based SMT (Tillmann et al., 1997; Och et al., 2001 ; Tillmann and Ney, 2003). This algorithm con- structs output sentences by starting with an empty hypothesis and adding output words at the end until translations for all source words have been generated. The core models of phrase-based SMT, in particular the n-gram language model (LM), only depend on a constant number of output words to the left of the word being generated. This fact is exploited by the search algorithm with a DP technique called hypothesis recombination (Och et al., 2001), which permits the elimination of hypotheses from the search space if they coincide in a certain number of final words with a better hypothesis and no future expansion can possibly invert the relative ranking of the two hypotheses under the given models. Hypothesis recombination achieves a substantial reduction of the search space without affecting search optimality and makes it possible to use aggressive pruning techniques for fast search while still obtaining good results. The downside of this otherwise excellent approach is that it only works well with models that have a local dependency structure similar to that of an n-gram language model, so they only depend on a small context window for each target word. Sentence-local models with longer dependencies can be added, but doing so greatly increases the risk for search errors by inhibiting hypothesis recombination. Cross-sentence dependencies cannot be directly integrated into DP SMT decoding in LParnogcue agdein Lgesa ornf tihneg, 2 p0a1g2e Jso 1in17t C9–o1n1f9e0re,n Jce ju on Is Elanmdp,ir Kicoarlea M,e 1t2h–o1d4s J iunly N 2a0tu1r2a.l ? Lc a2n0g1u2ag Aes Psorcoicaetsiosin fgo arn Cdo Cmopmutpauti oantiaoln Lailn Ngautiustriacls any obvious way, especially if joint optimisation of a number of interdependent decisions over an entire document is required. Research into models with a more varied, non-local dependency structure is to some extent stifled by the difficulty of decoding such models effectively, as can be seen by the problems some researchers encountered when they attempted to solve discourse-level problems. Consider, for instance, the work on cache-based language models by Tiedemann (2010) and Gong et al. (201 1), where error propagation was a serious issue, or the works on pronominal anaphora by Le Nagard and Koehn (2010), who implemented cross-sentence dependencies with an ad-hoc two-pass decoding strategy, and Hardmeier and Federico (2010) with the use of an external decoder driver to manage backward-only dependencies between sentences. In this paper, we present a method for decoding complete documents in phrase-based SMT. Our decoder uses a local search approach whose state consists of a complete translation of an entire document at any time. The initial state is improved by the application of a series of operations using a hill climbing strategy to find a (local) maximum of the score function. This setup gives us complete freedom to define scoring functions over the entire document. Moreover, by optionally initialising the state with the output of a traditional DP decoder, we can ensure that the final hypothesis is no worse than what would have been found by DP search alone. We start by describing the decoding algorithm and the state operations used by our decoder, then we present empirical results demonstrating the effectiveness of our approach and its usability with a document-level semantic language model, and finally we discuss some related work. 2 SMT Decoding by Hill Climbing In this section, we formally describe the phrasebased SMT model implemented by our decoder as well as the decoding algorithm we use. 2.1 SMT Model Our decoder is based on local search, so its state at any time is a representation of a complete translation of the entire document. Even though the decoder operates at the document level, it is important to keep 1180 track of sentence boundaries, and the individual operations that are applied to the state are still confined to sentence scope, so it is useful to decompose the state of a document into the state of its sentences, and we define the overall state S as a sequence of sentence states: S = S1S2 . . .SN, (1) where N is the number of sentences. This implies that we constrain the decoder to emit exactly one output sentence per input sentence. Let ibe the number of a sentence and mi the number of input tokens of this sentence, p and q (with 1 ≤ p ≤ q ≤ mi) be positions in the input sentence a1n ≤d [p; q] qde ≤no mte the set ofpositions from p up to and including q. We say that [p; q] precedes [p0; q0], or [p; q] ≺ [p0; q0], if q < p0. Let Φi([p; q]) be the set of t[pra;nqs]l ≺atio [pns for the source phrase covering positions [p; q] in the input sentence ias given by the phrase table. We call A = h[p; q] ,φi an anchored phrase pair w.it Wh coverage C(A) = [p; q] nif a φ ∈ Φi([p; q]) sise a target phrase translating =th [ep source w∈o Φrds at positions [p; q] . Then a sequence of ni anchored phrase pairs Si = A1A2 . . .Ani (2) is a valid sentence state for sentence iif the following two conditions hold: 1. The coverage sets C(Aj) for j in 1, . . . , ni are mutually disjoint, and 2. the anchored phrase pairs jointly cover the complete input sentence, or [niC(Aj) = [1;mi]. (3) [j=1 Let f(S) be a scoring function mapping a state S to a real number. As usual in SMT, it is assumed that the scoring function can be decomposed into a linear combination of K feature functions hk(S), each with a constant weight λk, so f(S) =k∑K=1λkhk(S). (4) The problem addressed by the decoder is the search for the state with maximal score, such that Sˆ Sˆ = argSmaxf(S). (5) The feature functions implemented in our baseline system are identical to the ones found in the popular Moses SMT system (Koehn et al., 2007). In particular, our decoder has the following feature functions: 1. phrase translation scores provided by the phrase table including forward and backward conditional probabilities, lexical weights and a phrase penalty (Koehn et al., 2003), 2. n-gram language model scores implemented with the KenLM toolkit (Heafield, 2011), 3. a word penalty score, 4. a distortion model with geometric (Koehn et al., 2003), and decay 5. a feature indicating the number of times a given distortion limit is exceeded in the current state. In our experiments, the last feature is used with a fixed weight of negative infinity in order to limit the gaps between the coverage sets of adjacent anchored phrase pairs to a maximum value. In DP search, the distortion limit is usually enforced directly by the search algorithm and is not added as a feature. In our decoder, however, this restriction is not required to limit complexity, so we decided to add it among the scoring models. 2.2 Decoding Algorithm The decoding algorithm we use (algorithm 1) is very simple. It starts with a given initial document state. In the main loop, which extends from line 3 to line 12, it generates a successor state S0 for the current state S by calling the function Neighbour, which non-deterministically applies one of the operations described in section 3 of this paper to S. The score of the new state is compared to that of the previous one. If it meets a given acceptance criterion, S0 becomes the current state, else search continues from the previous state S. For the experiments in this paper, we use the hill climbing acceptance criterion, which simply accepts a new state if its score is higher than that of the current state. Other acceptance criteria are possible and could be used to endow the search algorithm with stochastic behaviour. 1181 The main loop is repeated until a maximum number of steps (step limit) is reached or until a maximum number of moves are rejected in a row (rejection limit). Algorithm 1 Decoding algorithm Input: an initial document state S; search parameters maxsteps and maxrejected Output: a modified document state 1: nsteps ← 0 2: nrejected ← 0 3: nwrhejileec nsteps < maxsteps and nrejected < maxrejected do 4: S0 ← Neighbour (S) 5: if Accept (f(S0) , f(S)) then 6: S ← S0 7: nrejected ← 0 8: elsner 9: nrejected ← nrejected + 1 10: enndr eifj 11: nsteps ← nsteps + 1 12: ennds wtephsile ← 13: return S A notable difference between this algorithm and other hill climbing algorithms that have been used for SMT decoding (Germann et al., 2004; Langlais et al., 2007) is its non-determinism. Previous work for sentence-level decoding employed a steepest ascent strategy which amounts to enumerating the complete neighbourhood of the current state as defined by the state operations and selecting the next state to be the best state found in the neighbourhood of the current one. Enumerating all neighbours of a given state, costly as it is, has the advantage that it makes it easy to prove local optimality of a state by recognising that all possible successor states have lower scores. It can be rather inefficient, since at every step only one modification will be adopted; many of the modifications that are discarded will very likely be generated anew in the next iteration. As we extend the decoder to the document level, the size of the neighbourhood that would have to be explored in this way increases considerably. Moreover, the inefficiency of the steepest ascent approach potentially increases as well. Very likely, a promising move in one sentence will remain promising after a modification has been applied to another sentence, even though this is not guaranteed to be true in the presence of cross-sentence models. We therefore adopt a first-choice hill climbing strategy that non-deterministically generates successor states and accepts the first one that meets the acceptance criterion. This frees us from the necessity of generating the full set of successors for each state. On the downside, if the full successor set is not known, it is no longer possible to prove local optimality of a state, so we are forced to use a different condition for halting the search. We use a combination of two limits: The step limit is a hard limit on the resources the user is willing to expend on the search problem. The value of the rejection limit determines how much of the neighbourhood is searched for better successors before a state is accepted as a solution; it is related to the probability that a state returned as a solution is in fact locally optimal. To simplify notations in the description of the individual state operations, we write Si −→ Si0 (6) to signify that a state operation, when presented with a document state as in equation 1 and acting on sentence i, returns a new document state of S0 = S1 . . .Si−1 Si0 Si+1 . . .SN. (7) Similarly, Si : Aj . . .Aj+h−1 −→ A01 . . .A0h0 (8) is equivalent to Si −→ A1 . . .Aj−1 A01 . . .A0h0 Aj+h . . .Ani (9) and indicates that the operation returns a state in which a sequence of h consecutive anchored phrase pairs has been replaced by another sequence of h0 anchored phrase pairs. 2.3 Efficiency Considerations When implementing the feature functions for the decoder, we have to exercise some care to avoid recomputing scores for the whole document at every iteration. To achieve this, the scores are computed completely only once, at the beginning of the decoding run. In subsequent iterations, scoring functions are presented with the scores of the previous 1182 iteration and a list of modifications produced by the state operation, a set of tuples hi, r, s,A01 . . .A0h0i, each indicating tthioant ,t ahe s edto ocfu tmupelnets s hhio,ru,sld, Abe modifii,e eda as described by Si :Ar . . .As −→ A01 . . .A0h0 . (10) If a feature function is decomposable in some way, as all the standard features developed under the constraints of DP search are, it can then update the state simply by subtracting and adding score components pertaining to the modified parts of the document. Feature functions have the possibility to store their own state information along with the document state to make sure the required information is available. Thus, the framework makes it possible to exploit decomposability for efficient scoring without impos- ing any particular decomposition on the features as beam search does. To make scoring even more efficient, scores are computed in two passes: First, every feature function is asked to provide an upper bound on the score that will be obtained for the new state. In some cases, it is possible to calculate reasonable upper bounds much more efficiently than computing the exact feature value. If the upper bound fails to meet the acceptance criterion, the new state is discarded right away; if not, the full score is computed and the acceptance criterion is tested again. Among the basic SMT models, this two-pass strategy is only used for the n-gram LM, which requires fairly expensive parameter lookups for scoring. The scores of all the other baseline models are fully computed during the first scoring pass. The n-gram model is more complex. In its state information, it keeps track of the LM score and LM library state for each word. The first scoring pass then identifies the words whose LM scores are affected by the current search step. This includes the words changed by the search operation as well as the words whose LM history is modified. The range of the history de- pendencies can be determined precisely by considering the “valid state length” information provided by the KenLM library. In the first pass, the LM scores of the affected words are subtracted from the total score. The model only looks up the new LM scores for the affected words and updates the total score if the new search state passes the first acceptance check. This two-pass scoring approach allows us to avoid LM lookups altogether for states that will be rejected anyhow because of low scores from the other models, e. g. because the distortion limit is violated. Model score updates become more complex and slower as the number of dependencies of a model increases. While our decoding algorithm does not impose any formal restrictions on the number or type of dependencies that can be handled, there will be practical limits beyond which decoding becomes unacceptably slow or the scoring code becomes very difficult to maintain. These limits are however fairly independent of the types of dependencies handled by a model, which permits the exploration of more varied model types than those handled by DP search. 2.4 State Initialisation Before the hill climbing decoding algorithm can be run, an initial state must be generated. The closer the initial state is to an optimum, the less work remains to be done for the algorithm. If the algorithm is to be self-contained, initialisation must be relatively uninformed and can only rely on some general prior assumptions about what might be a good initial guess. On the other hand, if optimal results are sought after, it pays off to invest some effort into a good starting point. One way to do this is to run DP search first. For uninformed initialisation, we chose to implement a very simple procedure based only on the observation that, at least for language pairs involving the major European languages, it is usually a good guess to keep the word order of the output very similar to that of the input. We therefore create the initial state by selecting, for each sentence in the document, a sequence of anchored phrase pairs covering the input sentence in monotonic order, that is, such that for all pairs of adjacent anchored phrase pairs Aj and Aj+1, we have that C(Aj) ≺ C(Aj+1 ). For initialisation with DP search, we first run the Moses decoder (Koehn et al., 2007) with default search parameters and the same models as those used by our decoder. Then we extract the best output hypothesis from the search graph of the decoder and map it into a sequence of anchored phrase pairs in the obvious way. When the document-level decoder is used with models that are incompatible with beam search, Moses can be run with a subset of the models in order to find an approximation of the solution 1183 which is then refined with the complete feature set. 3 State Operations Given a document state S, the decoder uses a neighbourhood function Neighbour to simulate a move in the state space. The neighbourhood function nondeterministically selects a type of state operation and a location in the document to apply it to and returns the resulting new state. We use a set of three operations that has the property that every possible document state can be reached from every other state in a sequence of moves. Designing operations for state transitions in local search for phrase-based SMT is a problem that has been addressed in the literature (Langlais et al., 2007; Arun et al., 2010). Our decoder’s first- choice hill climbing strategy never enumerates the full neighbourhood of a state. We therefore place less emphasis than previous work on defining a compact neighbourhood, but allow the decoder to make quite extensive changes to a state in a single step with a certain probability. Otherwise our operations are similar to those used by Arun et al. (2010). All of the operations described in this paper make changes to a single sentence only. Each time it is called, the Neighbour function selects a sentence in the document with a probability proportional to the number of input tokens in each sentence to ensure a fair distribution ofthe decoder’s attention over the words in the document regardless of varying sentence lengths. 3.1 Changing Phrase Translations The change-phrase-translation operation replaces the translation of a single phrase with a random translation with the same coverage taken from the phrase table. Formally, the operation selects an anchored phrase pair Aj by drawing uniformly from the elements of Si and then draws a new translation φ0 uniformly from the set Φi(C(Aj)). The new state is given by Si : Aj −→ hC(Aj), φ0i. (11) 3.2 Changing Word Order The swap-phrases operation affects the output word order without changing the phrase translations. It exchanges two anchored phrase pairs Aj and Aj+h, resulting in an output state of Si : Aj . . .Aj+h −→ Aj+h Aj+1 . . .Aj+h−1 Aj. (12) The start location j is drawn uniformly from the eligible sentence positions; the swap range h comes from a geometric distribution with configurable decay. Other word-order changes such as a one-way move operation that does not require another movement in exchange or more advanced permutations can easily be defined. 3.3 Resegmentation The most complex operation is resegment, which allows the decoder to modify the segmentation ofthe source phrase. It takes a number of anchored phrase pairs that form a contiguous block both in the input and in the output and replaces them with a new set of phrase pairs covering the same span of the input sentence. Formally, Si : Aj . . .Aj+h−1 −→ A01 . . .A0h0 (13) such that j+[h−1 [h0 [ C(Aj0) = [ C(A0j0) = [p;q] j[0=j (14) j[0=1 for some p and q, where, for j0 = 1, . . . ,h0, we have that A0j0 = h[pj0; qj0] , φj0i, all [pj0; qj0] are mutually disjoint =an hd[ peach φj0 isi randomly drawn from Φi([pj0;qj0]). Regardless of the ordering of Aj . . .Aj+h−1 , the resegment operation always generates a sequence of anchored phrase pairs in linear order, such that C(A0j0) ≺ C(A0j0+1 ) for j0 = 1, . . . ,h0 −1 . As )f o≺r Cth(eA other operations, j is− generated uniformly and h is drawn from a geometric distribution with a decay parameter. The new segmentation is generated by extending the sequence of anchored phrase pairs with random elements starting at the next free position, proceeding from left to right until the whole range [p; q] is covered. 4 Experimental Results In this section, we present the results of a series of experiments with our document decoder. The 1184 goal of our experiments is to demonstrate the behaviour of the decoder and characterise its response to changes in the fundamental search parameters. The SMT models for our experiments were created with a subset of the training data for the English-French shared task at the WMT 2011workshop (Callison-Burch et al., 2011). The phrase table was trained on Europarl, news-commentary and UN data. To reduce the training data to a manageable size, singleton phrase pairs were removed before the phrase scoring step. Significance-based filtering (Johnson et al., 2007) was applied to the resulting phrase table. The language model was a 5gram model with Kneser-Ney smoothing trained on the monolingual News corpus with IRSTLM (Federico et al., 2008). Feature weights were trained with Minimum Error-Rate Training (MERT) (Och, 2003) on the news-test2008 development set using the DP beam search decoder and the MERT implementation of the Moses toolkit (Koehn et al., 2007). Experimental results are reported for the newstest2009 test set, a corpus of 111 newswire documents totalling 2,525 sentences or 65,595 English input tokens. 4.1 Stability An important difference between our decoder and the classical DP decoder as well as previous work in SMT decoding with local search is that our decoder is inherently non-deterministic. This implies that repeated runs of the decoder with the same search parameters, input and models will not, in general, find the same local maximum of the score space. The first empirical question we ask is therefore how different the results are under repeated runs. The results in this and the next section were obtained with random state initialisation, i. e. without running the DP beam search decoder. Figure 1 shows the results of 7 decoder runs with the models described above, translating the newstest2009 test set, with a step limit of 227 and a rejection limit of 100,000. The x-axis of both plots shows the number of decoding steps on a logarithmic scale, so the number of steps is doubled between two adjacent points on the same curve. In the left plot, the y-axis indicates the model score optimised by the decoder summed over all 2525 sentences of the document. In the right plot, the case-sensitive BLEU score (Papineni et al., 2002) of the current decoder Figure 1: Score stability in repeated decoder runs state against a reference translation is displayed. We note, as expected, that the decoder achieves a considerable improvement of the initial state with diminishing returns as decoding continues. Between 28 and 214 steps, the score increases at a roughly logarithmic pace, then the curve flattens out, which is partly due to the fact that decoding for some documents effectively stopped when the maximum number of rejections was reached. The BLEU score curve shows a similar increase, from an initial score below 5 % to a maximum of around 21.5 %. This is below the score of 22.45 % achieved by the beam search decoder with the same models, which is not surprising considering that our decoder approximates a more difficult search problem, from which a number of strong independence assumptions have been lifted, without, at the moment, having any stronger models at its disposal to exploit this additional freedom for better translation. In terms of stability, there are no dramatic differences between the decoder runs. Indeed, the small differences that exist are hardly discernible in the plots. The model scores at the end of the decoding run range between −158767.9 and −158716.9, a g re rlautniv rea ndgieffe breetnwceee nof − only a6b7.o9ut a n0d.0 −3 %15.8 F1i6n.a9l, BLEU scores range from 21.41 % to 21.63 %, an interval that is not negligible, but comparable to the variance observed when, e. g., feature weights from repeated MERT runs are used with one and the same SMT system. Note that these results were obtained with random state initialisation. With DP initialisation, score differences between repeated runs rarely 1185 exceed 0.02 absolute BLEU percentage points. Overall, we conclude that the decoding results of our algorithm are reasonably stable despite the nondeterminism inherent in the procedure. In our subsequent experiments, the evaluation scores reported are calculated as the mean of three runs for each experiment. 4.2 Search Algorithm Parameters The hill climbing algorithm we use has two parameters which govern the trade-off between decoding time and the accuracy with which a local maximum is identified: The step limit stops the search process after a certain number of steps regardless of the search progress made or lack thereof. The rejection limit stops the search after a certain number of unsuccessful attempts to make a step, when continued search does not seem to be promising. In most of our experiments, we used a step limit of 227 ≈ 1.3 · 108 and a rejection limit of 105. In practice, decoding terminates by reaching the rejection limit for the vast majority of documents. We therefore examined the effect of different rejection limits on the learning curves. The results are shown in figure 2. The results show that continued search does pay off to a certain extent. Indeed, the curve for rejection limit 107 seems to indicate that the model score increases roughly logarithmically, albeit to a higher base, even after the curve has started to flatten out at 214 steps. At a certain point, however, the probability of finding a good successor state drops rather sharply by about two orders of magnitude, as Figure 2: Search performance at different rejection limits evidenced by the fact that a rejection limit of 106 does not give a large improvement over one of 105, while one of 107 does. The continued model score improvement also results in an increase in BLEU scores, and with a BLEU score of 22. 1% the system with rejection limit 107 is fairly close to the score of 22.45 % obtained by DP beam search. Obviously, more exact search comes at a cost, and in this case, it comes at a considerable cost, which is an explosion of the time required to decode the test set from 4 minutes at rejection limit 103 to 224 minutes at rejection limit 105 and 38 hours 45 minutes at limit 107. The DP decoder takes 3 1 minutes for the same task. We conclude that the rejection limit of 105 selected for our experiments, while technically suboptimal, realises a good trade-off between decoding time and accuracy. 4.3 A Semantic Document Language Model In this section, we present the results of the application of our decoder to an actual SMT model with cross-sentence features. Our model addresses the problem of lexical cohesion. In particular, it rewards the use of semantically related words in the translation output by the decoder, where semantic distance is measured with a word space model based on Latent Semantic Analysis (LSA). LSA has been applied to semantic language modelling in previous research with some success (Coccaro and Jurafsky, 1998; Bellegarda, 2000; Wandmacher and Antoine, 2007). In SMT, it has mostly been used for domain adaptation (Kim and Khudanpur, 2004; Tam et al., 1186 2007), or to measure sentence similarities (Banchs and Costa-juss a`, 2011). The model we use is inspired by Bellegarda (2000). It is a Markov model, similar to a standard n-gram model, and assigns to each content word a score given a history of n preceding content words, where n = 30 below. Scoring relies on a 30dimensional LSA word vector space trained with the S-Space software (Jurgens and Stevens, 2010). The score is defined based on the cosine similarity between the word vector of the predicted word and the mean word vector of the words in the history, which is converted to a probability by histogram lookup as suggested by Bellegarda (2000). The model is structurally different from a regular n-gram model in that word vector n-grams are defined over content words occurring in the word vector model only and can cross sentence boundaries. Stop words, identified by an extensive stop word list and amounting to around 60 % of the tokens, are scored by a different mechanism based on their relative frequency (undiscounted unigram probability) in the training corpus. In sum, the score produced by the semantic document LM has the following form: wh(er|h)α=is tεpαheuncipgors(wp)o|hrtinof w fci os nakutneskotn wpon w ,onerldse,in(ls1teh5) training corpus and ε is a small fixed probability. It is integrated into the decoder as an extra feature function. Since we lack an automatic method for training the feature weights of document-wide features, its weight was selected by grid search over a number of values, comparing translation performance for the newstest2009 test set. In these experiments, we used DP beam search to initialise the state of our local search decoder. Three results are presented (table 1): The first table row shows the baseline performance using DP beam search with standard sentence-local features only. The scores in the second row were obtained by running the hill climbing decoder with DP initialisation, but without adding any models. A marginal increase in scores for all three test sets demonstrates that the hill climbing decoder manages to fix some of the search errors made by the DP search. The last row contains the scores obtained by adding in the semantic language model. Scores are presented for three publicly available test sets from recent WMT Machine Translation shared tasks, of which one (newstest2009) was used to monitor progress during development and select the final model. Adding the semantic language model results in a small increase in NIST scores (Doddington, 2002) for all three test sets as well as a small BLEU score gain (Papineni et al., 2002) for two out of three corpora. We note that the NIST score turned out to react more sensitively to improvements due to the semantic LM in all our experiments, which is reasonable because the model specifically targets content words, which benefit from the information weighting done by the NIST score. While the results we present do not constitute compelling evidence in favour of our semantic LM in its current form, they do suggest that this model could be improved to realise higher gains from cross-sentence semantic information. They support our claim that cross- sentence models should be examined more closely and that existing methods should be adapted to deal with them, a problem addressed by our main contribution, the local search document decoder. 5 Related Work Even though DP beam search (Koehn et al., 2003) has been the dominant approach to SMT decoding in recent years, methods based on local search have been explored at various times. For word-based SMT, greedy hill-climbing techniques were advo1187 cated as a faster replacement for beam search (Germann et al., 2001 ; Germann, 2003; Germann et al., 2004), and a problem formulation specifically targeting word reordering with an efficient word reordering algorithm has been proposed (Eisner and Tromble, 2006). A local search decoder has been advanced as a faster alternative to beam search also for phrasebased SMT (Langlais et al., 2007; Langlais et al., 2008). That work anticipates many of the features found in our decoder, including the use of local search to refine an initial hypothesis produced by DP beam search. The possibility of using models that do not fit well into the beam search paradigm is mentioned and illustrated with the example of a reversed n-gram language model, which the authors claim would be difficult to implement in a beam search decoder. Similarly to the work by Germann et al. (2001), their decoder is deterministic and explores the entire neighbourhood of a state in order to identify the most promising step. Our main contribution with respect to the work by Langlais et al. (2007) is the introduction of the possibility of handling document-level models by lifting the assumption of sentence independence. As a consequence, enumerating the entire neighbourhood becomes too expensive, which is why we resort to a “first-choice” strategy that non-deterministically generates states and accepts the first one encountered that meets the acceptance criterion. More recently, Gibbs sampling was proposed as a way to generate samples from the posterior distribution of a phrase-based SMT decoder (Arun et al., 2009; Arun et al., 2010), a process that resembles local search in its use of a set of state-modifying operators to generate a sequence of decoder states. Where local search seeks for the best state attainable from a given initial state, Gibbs sampling produces a representative sample from the posterior. Like all work on SMT decoding that we know of, the Gibbs sampler presented by Arun et al. (2010) assumes independence of sentences and considers the complete neighbourhood of each state before taking a sample. 6 Conclusion In the last twenty years of SMT research, there has been a strong assumption that sentences in a text newstest2009 newstest2010 newstest201 1 BLEU NIST BLEU NIST BLEU NIST 22.56 6.513 27.27 7.034 24.94 7.170 + hill climbing 22.60 6.518 27.33 7.046 24.97 7.169 with semantic LM 22.71 6.549 27.53 7.087 24.90 7.199 DP search only DP Table 1: Experimental results with a cross-sentence semantic language model are independent of one another, and discourse context has been largely neglected. Several factors have contributed to this. Developing good discourse-level models is difficult, and considering the modest translation quality that has long been achieved by SMT, there have been more pressing problems to solve and lower hanging fruit to pick. However, we argue that the popular DP beam search algorithm, which delivers excellent decoding performance, but imposes a particular kind of local dependency structure on the feature models, has also had its share in driving researchers away from discourse-level problems. In this paper, we have presented a decoding procedure for phrase-based SMT that makes it possible to define feature models with cross-sentence dependencies. Our algorithm can be combined with DP beam search to leverage the quality of the traditional approach with increased flexibility for models at the discourse level. We have presented preliminary results on a cross-sentence semantic language model addressing the problem of lexical cohesion to demonstrate that this kind of models is worth exploring further. Besides lexical cohesion, cross-sentence models are relevant for other linguistic phenomena such as pronominal anaphora or verb tense selection. We believe that SMT research has reached a point of maturity where discourse phenomena should not be ignored any longer, and we consider our decoder to be a step towards this goal. References Abhishek Arun, Chris Dyer, Barry Haddow, Phil Blunsom, Adam Lopez, and Philipp Koehn. 2009. Monte carlo inference and maximization for phrase-based translation. In Proceedings of the Thirteenth Conference on Computational Natural Language Learning (CoNLL-2009), pages 102–1 10, Boulder, Colorado, June. Association for Computational Linguistics. Abhishek Arun, Barry Haddow, Philipp Koehn, Adam Lopez, Chris Dyer, and Phil Blunsom. 2010. Monte 1188 Ma- Carlo techniques for phrase-based translation. chine translation, 24(2): 103–121 . Rafael E. Banchs and Marta R. Costa-juss a`. 2011. A semantic feature for Statistical Machine Translation. In Proceedings of Fifth Workshop on Syntax, Semantics and Structure in Statistical Translation, pages 126– 134, Portland, Oregon, USA, June. Association for Computational Linguistics. Jerome R. Bellegarda. 2000. Exploiting latent semantic information in statistical language modeling. Proceedings of the IEEE, 88(8): 1279–1296. Chris Callison-Burch, Philipp Koehn, Christof Monz, and Omar Zaidan. 2011. Findings of the 2011 Workshop on Statistical Machine Translation. In Proceedings of the Sixth Workshop on Statistical Machine Translation, pages 22–64, Edinburgh, Scotland, July. Association for Computational Linguistics. Noah Coccaro and Daniel Jurafsky. 1998. Towards better integration of semantic predictors in statistical language modeling. In Proceedings of the 5th International Conference on Spoken Language Processing, Sydney. George Doddington. 2002. Automatic evaluation of machine translation quality using n-gram co-occurrence statistics. In Proceedings of the second International conference on Human Language Technology Research, pages 138–145, San Diego. Jason Eisner and Roy W. Tromble. 2006. Local search with very large-scale neighborhoods for optimal permutations in machine translation. In Proceedings of the HLT-NAACL Workshop on Computationally Hard Problems and Joint Inference in Speech and Language Processing, pages 57–75. Marcello Federico, Nicola Bertoldi, and Mauro Cettolo. 2008. IRSTLM: an open source toolkit for handling large scale language models. In Interspeech 2008, pages 1618–1621 . ISCA. Ulrich Germann, Michael Jahr, Kevin Knight, Daniel Marcu, and Kenji Yamada. 2001 . Fast decoding and optimal decoding for machine translation. In Proceedings of 39th Annual Meeting of the Association for Computational Linguistics, pages 228–235, Toulouse, France, July. Association for Computational Linguis- tics. Ulrich Germann, Michael Jahr, Kevin Knight, Daniel Marcu, and Kenji Yamada. 2004. Fast and optimal decoding for machine translation. Artificial Intelligence, 154(1–2): 127–143. Ulrich Germann. 2003. Greedy decoding for Statistical Machine Translation in almost linear time. In Proceedings of the 2003 Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics. Zhengxian Gong, Min Zhang, and Guodong Zhou. 2011. Cache-based document-level Statistical Machine Translation. In Proceedings of the 2011 Conference on Empirical Methods in Natural Language Processing, pages 909–919, Edinburgh, Scotland, UK., July. Association for Computational Linguistics. Christian Hardmeier and Marcello Federico. 2010. Modelling Pronominal Anaphora in Statistical Machine Translation. In Proceedings of the seventh International Workshop on Spoken Language Translation (IWSLT), pages 283–289. Basil Hatim and Ian Mason. 1990. Discourse and the Translator. Language in Social Life Series. Longman, London. Kenneth Heafield. 2011. KenLM: faster and smaller language model queries. In Proceedings of the Sixth Workshop on Statistical Machine Translation, pages 187–197, Edinburgh, Scotland, July. Association for Computational Linguistics. Howard Johnson, Joel Martin, George Foster, and Roland Kuhn. 2007. Improving translation quality by discarding most of the phrasetable. In Proceedings of the 2007 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning (EMNLP-CoNLL), pages 967– 975, Prague, Czech Republic, June. Association for Computational Linguistics. David Jurgens and Keith Stevens. 2010. The S-Space package: An open source package for word space models. In Proceedings of the ACL 2010 System Demonstrations, pages 30–35, Uppsala, Sweden, July. Association for Computational Linguistics. Woosung Kim and Sanjeev Khudanpur. 2004. Crosslingual latent semantic analysis for language modeling. In IEEE international conference on acoustics, speech, and signal processing (ICASSP), volume 1, pages 257–260, Montr ´eal. Philipp Koehn, Franz Josef Och, and Daniel Marcu. 2003. Statistical phrase-based translation. In Proceedings of the 2003 conference of the North American chapter of the Association for Computational Linguistics on Human Language Technology, pages 48– 54, Edmonton. 1189 Philipp Koehn, Hieu Hoang, Alexandra Birch, et al. 2007. Moses: open source toolkit for Statistical Machine Translation. In Annual meeting of the Association for Computational Linguistics: Demonstration session, pages 177–180, Prague. Philippe Langlais, Alexandre Patry, and Fabrizio Gotti. 2007. A greedy decoder for phrase-based statistical machine translation. In TMI-2007: Proceedings of the 11th International Conference on Theoretical and Methodological Issues in Machine Translation, pages 104–1 13, Sk¨ ovde. Philippe Langlais, Alexandre Patry, and Fabrizio Gotti. 2008. Recherche locale pour la traduction statistique par segments. In TALN 2008, pages 119–128, Avignon, France, June. ATALA. Ronan Le Nagard and Philipp Koehn. 2010. Aiding pronoun translation with co-reference resolution. In Proceedings of the Joint Fifth Workshop on Statistical Machine Translation and MetricsMATR, pages 252–261, Uppsala, Sweden, July. Association for Computational Linguistics. Franz Josef Och, Nicola Ueffing, and Hermann Ney. 2001. An efficient A* search algorithm for Statistical Machine Translation. In Proceedings of the DataDriven Machine Translation Workshop, 39th Annual Meeting of the Association for Computational Linguistics (ACL), pages 55–62, Toulouse. Franz Josef Och. 2003. Minimum error rate training in Statistical Machine Translation. In Proceedings of the 41st annual meeting of the Association for Computational Linguistics, pages 160–167, Sapporo (Japan). Kishore Papineni, Salim Roukos, Todd Ward, and WeiJing Zhu. 2002. BLEU: a method for automatic evaluation of Machine Translation. In Proceedings of the 40th annual meeting of the Association for Computational Linguistics, pages 3 11–3 18, Philadelphia. ACL. Yik-Cheung Tam, Ian Lane, and Tanja Schultz. 2007. Bilingual LSA-based adaptation for Statistical Machine Translation. Machine Translation, 21(4): 187– 207. J o¨rg Tiedemann. 2010. To cache or not to cache? Experiments with adaptive models in Statistical Machine Translation. In Proceedings of the ACL 2010 Joint Fifth Workshop on Statistical Machine Translation and Metrics MATR, pages 189–194, Uppsala, Sweden. Association for Computational Linguistics. Christoph Tillmann and Hermann Ney. 2003. Word reordering and a Dynamic Programming beam search algorithm for Statistical Machine Translation. Computational linguistics, 29(1):97–133. Christoph Tillmann, Stephan Vogel, Hermann Ney, and Alex Zubiaga. 1997. A DP-based search using monotone alignments in Statistical Translation. In Proceedings of the 35th Annual Meeting of the Association for Computational Spain, tics. Linguistics, July. Association for 289–296, Madrid, Computational Linguis- pages Tonio Wandmacher and Jean-Yves Antoine. 2007. Methods to integrate a language model with semantic information for a word prediction component. In Proceedings of the 2007 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning (EMNLPCoNLL), pages 506–5 13, Prague, Czech Republic, June. Association for Computational Linguistics. 1190
5 0.060714364 106 emnlp-2012-Part-of-Speech Tagging for Chinese-English Mixed Texts with Dynamic Features
Author: Jiayi Zhao ; Xipeng Qiu ; Shu Zhang ; Feng Ji ; Xuanjing Huang
Abstract: In modern Chinese articles or conversations, it is very popular to involve a few English words, especially in emails and Internet literature. Therefore, it becomes an important and challenging topic to analyze Chinese-English mixed texts. The underlying problem is how to tag part-of-speech (POS) for the English words involved. Due to the lack of specially annotated corpus, most of the English words are tagged as the oversimplified type, “foreign words”. In this paper, we present a method using dynamic features to tag POS of mixed texts. Experiments show that our method achieves higher performance than traditional sequence labeling methods. Meanwhile, our method also boosts the performance of POS tagging for pure Chinese texts.
6 0.055779286 21 emnlp-2012-Assessment of ESL Learners' Syntactic Competence Based on Similarity Measures
7 0.055157769 62 emnlp-2012-Identifying Constant and Unique Relations by using Time-Series Text
8 0.050850321 86 emnlp-2012-Locally Training the Log-Linear Model for SMT
9 0.04826213 54 emnlp-2012-Forced Derivation Tree based Model Training to Statistical Machine Translation
10 0.044438928 17 emnlp-2012-An "AI readability" Formula for French as a Foreign Language
11 0.041596938 127 emnlp-2012-Transforming Trees to Improve Syntactic Convergence
12 0.040893663 109 emnlp-2012-Re-training Monolingual Parser Bilingually for Syntactic SMT
13 0.04035176 2 emnlp-2012-A Beam-Search Decoder for Grammatical Error Correction
14 0.039575141 72 emnlp-2012-Joint Inference for Event Timeline Construction
15 0.035318524 70 emnlp-2012-Joint Chinese Word Segmentation, POS Tagging and Parsing
16 0.034839779 67 emnlp-2012-Inducing a Discriminative Parser to Optimize Machine Translation Reordering
17 0.033094741 128 emnlp-2012-Translation Model Based Cross-Lingual Language Model Adaptation: from Word Models to Phrase Models
18 0.032821026 1 emnlp-2012-A Bayesian Model for Learning SCFGs with Discontiguous Rules
19 0.031759214 131 emnlp-2012-Unified Dependency Parsing of Chinese Morphological and Syntactic Structures
20 0.031169349 42 emnlp-2012-Entropy-based Pruning for Phrase-based Machine Translation
topicId topicWeight
[(0, 0.116), (1, -0.069), (2, -0.026), (3, -0.029), (4, 0.018), (5, -0.02), (6, -0.004), (7, -0.029), (8, -0.035), (9, -0.053), (10, -0.067), (11, -0.054), (12, -0.066), (13, 0.009), (14, 0.004), (15, -0.069), (16, -0.133), (17, -0.039), (18, 0.126), (19, 0.109), (20, -0.021), (21, -0.045), (22, 0.097), (23, 0.211), (24, -0.106), (25, 0.028), (26, -0.061), (27, -0.074), (28, -0.084), (29, -0.004), (30, -0.003), (31, -0.186), (32, 0.103), (33, 0.168), (34, -0.01), (35, 0.063), (36, -0.073), (37, -0.246), (38, -0.182), (39, 0.047), (40, 0.083), (41, 0.146), (42, 0.164), (43, 0.056), (44, -0.103), (45, 0.209), (46, 0.228), (47, -0.13), (48, -0.239), (49, -0.199)]
simIndex simValue paperId paperTitle
same-paper 1 0.98212731 95 emnlp-2012-N-gram-based Tense Models for Statistical Machine Translation
Author: Zhengxian Gong ; Min Zhang ; Chew Lim Tan ; Guodong Zhou
Abstract: Tense is a small element to a sentence, however, error tense can raise odd grammars and result in misunderstanding. Recently, tense has drawn attention in many natural language processing applications. However, most of current Statistical Machine Translation (SMT) systems mainly depend on translation model and language model. They never consider and make full use of tense information. In this paper, we propose n-gram-based tense models for SMT and successfully integrate them into a state-of-the-art phrase-based SMT system via two additional features. Experimental results on the NIST Chinese-English translation task show that our proposed tense models are very effective, contributing performance improvement by 0.62 BLUE points over a strong baseline. 1
2 0.53046566 80 emnlp-2012-Learning Verb Inference Rules from Linguistically-Motivated Evidence
Author: Hila Weisman ; Jonathan Berant ; Idan Szpektor ; Ido Dagan
Abstract: Learning inference relations between verbs is at the heart of many semantic applications. However, most prior work on learning such rules focused on a rather narrow set of information sources: mainly distributional similarity, and to a lesser extent manually constructed verb co-occurrence patterns. In this paper, we claim that it is imperative to utilize information from various textual scopes: verb co-occurrence within a sentence, verb cooccurrence within a document, as well as overall corpus statistics. To this end, we propose a much richer novel set of linguistically motivated cues for detecting entailment between verbs and combine them as features in a supervised classification framework. We empirically demonstrate that our model significantly outperforms previous methods and that information from each textual scope contributes to the verb entailment learning task.
3 0.33972555 62 emnlp-2012-Identifying Constant and Unique Relations by using Time-Series Text
Author: Yohei Takaku ; Nobuhiro Kaji ; Naoki Yoshinaga ; Masashi Toyoda
Abstract: Because the real world evolves over time, numerous relations between entities written in presently available texts are already obsolete or will potentially evolve in the future. This study aims at resolving the intricacy in consistently compiling relations extracted from text, and presents a method for identifying constancy and uniqueness of the relations in the context of supervised learning. We exploit massive time-series web texts to induce features on the basis of time-series frequency and linguistic cues. Experimental results confirmed that the time-series frequency distributions contributed much to the recall of constancy identification and the precision of the uniqueness identification.
4 0.27115601 21 emnlp-2012-Assessment of ESL Learners' Syntactic Competence Based on Similarity Measures
Author: Su-Youn Yoon ; Suma Bhat
Abstract: This study presents a novel method that measures English language learners’ syntactic competence towards improving automated speech scoring systems. In contrast to most previous studies which focus on the length of production units such as the mean length of clauses, we focused on capturing the differences in the distribution of morpho-syntactic features or grammatical expressions across proficiency. We estimated the syntactic competence through the use of corpus-based NLP techniques. Assuming that the range and so- phistication of grammatical expressions can be captured by the distribution of Part-ofSpeech (POS) tags, vector space models of POS tags were constructed. We use a large corpus of English learners’ responses that are classified into four proficiency levels by human raters. Our proposed feature measures the similarity of a given response with the most proficient group and is then estimates the learner’s syntactic competence level. Widely outperforming the state-of-the-art measures of syntactic complexity, our method attained a significant correlation with humanrated scores. The correlation between humanrated scores and features based on manual transcription was 0.43 and the same based on ASR-hypothesis was slightly lower, 0.42. An important advantage of our method is its robustness against speech recognition errors not to mention the simplicity of feature generation that captures a reasonable set of learnerspecific syntactic errors. 600 Measures Suma Bhat Beckman Institute, Urbana, IL 61801 . spbhat 2 @ i l l ino i edu s
5 0.25482866 17 emnlp-2012-An "AI readability" Formula for French as a Foreign Language
Author: Thomas Francois ; Cedrick Fairon
Abstract: This paper present a new readability formula for French as a foreign language (FFL), which relies on 46 textual features representative of the lexical, syntactic, and semantic levels as well as some of the specificities of the FFL context. We report comparisons between several techniques for feature selection and various learning algorithms. Our best model, based on support vector machines (SVM), significantly outperforms previous FFL formulas. We also found that semantic features behave poorly in our case, in contrast with some previous readability studies on English as a first language.
6 0.22120926 86 emnlp-2012-Locally Training the Log-Linear Model for SMT
8 0.20210981 35 emnlp-2012-Document-Wide Decoding for Phrase-Based Statistical Machine Translation
9 0.19683282 106 emnlp-2012-Part-of-Speech Tagging for Chinese-English Mixed Texts with Dynamic Features
10 0.17403056 54 emnlp-2012-Forced Derivation Tree based Model Training to Statistical Machine Translation
11 0.17127749 38 emnlp-2012-Employing Compositional Semantics and Discourse Consistency in Chinese Event Extraction
12 0.1665507 18 emnlp-2012-An Empirical Investigation of Statistical Significance in NLP
13 0.16139627 27 emnlp-2012-Characterizing Stylistic Elements in Syntactic Structure
14 0.15823208 2 emnlp-2012-A Beam-Search Decoder for Grammatical Error Correction
15 0.15578908 118 emnlp-2012-Source Language Adaptation for Resource-Poor Machine Translation
16 0.15169394 127 emnlp-2012-Transforming Trees to Improve Syntactic Convergence
17 0.14762945 74 emnlp-2012-Language Model Rest Costs and Space-Efficient Storage
18 0.14603969 10 emnlp-2012-A Statistical Relational Learning Approach to Identifying Evidence Based Medicine Categories
19 0.13840902 83 emnlp-2012-Lexical Differences in Autobiographical Narratives from Schizophrenic Patients and Healthy Controls
20 0.13674408 46 emnlp-2012-Exploiting Reducibility in Unsupervised Dependency Parsing
topicId topicWeight
[(2, 0.022), (16, 0.026), (25, 0.015), (34, 0.087), (41, 0.015), (45, 0.014), (55, 0.26), (60, 0.081), (63, 0.052), (64, 0.051), (65, 0.032), (70, 0.03), (73, 0.022), (74, 0.056), (76, 0.034), (79, 0.028), (80, 0.015), (86, 0.025), (95, 0.019)]
simIndex simValue paperId paperTitle
same-paper 1 0.7691974 95 emnlp-2012-N-gram-based Tense Models for Statistical Machine Translation
Author: Zhengxian Gong ; Min Zhang ; Chew Lim Tan ; Guodong Zhou
Abstract: Tense is a small element to a sentence, however, error tense can raise odd grammars and result in misunderstanding. Recently, tense has drawn attention in many natural language processing applications. However, most of current Statistical Machine Translation (SMT) systems mainly depend on translation model and language model. They never consider and make full use of tense information. In this paper, we propose n-gram-based tense models for SMT and successfully integrate them into a state-of-the-art phrase-based SMT system via two additional features. Experimental results on the NIST Chinese-English translation task show that our proposed tense models are very effective, contributing performance improvement by 0.62 BLUE points over a strong baseline. 1
2 0.76642841 26 emnlp-2012-Building a Lightweight Semantic Model for Unsupervised Information Extraction on Short Listings
Author: Doo Soon Kim ; Kunal Verma ; Peter Yeh
Abstract: Short listings such as classified ads or product listings abound on the web. If a computer can reliably extract information from them, it will greatly benefit a variety of applications. Short listings are, however, challenging to process due to their informal styles. In this paper, we present an unsupervised information extraction system for short listings. Given a corpus of listings, the system builds a semantic model that represents typical objects and their attributes in the domain of the corpus, and then uses the model to extract information. Two key features in the system are a semantic parser that extracts objects and their attributes and a listing-focused clustering module that helps group together extracted tokens of same type. Our evaluation shows that the , semantic model learned by these two modules is effective across multiple domains.
3 0.51125473 129 emnlp-2012-Type-Supervised Hidden Markov Models for Part-of-Speech Tagging with Incomplete Tag Dictionaries
Author: Dan Garrette ; Jason Baldridge
Abstract: Past work on learning part-of-speech taggers from tag dictionaries and raw data has reported good results, but the assumptions made about those dictionaries are often unrealistic: due to historical precedents, they assume access to information about labels in the raw and test sets. Here, we demonstrate ways to learn hidden Markov model taggers from incomplete tag dictionaries. Taking the MINGREEDY algorithm (Ravi et al., 2010) as a starting point, we improve it with several intuitive heuristics. We also define a simple HMM emission initialization that takes advantage of the tag dictionary and raw data to capture both the openness of a given tag and its estimated prevalence in the raw data. Altogether, our augmentations produce improvements to per- formance over the original MIN-GREEDY algorithm for both English and Italian data.
4 0.49848753 136 emnlp-2012-Weakly Supervised Training of Semantic Parsers
Author: Jayant Krishnamurthy ; Tom Mitchell
Abstract: We present a method for training a semantic parser using only a knowledge base and an unlabeled text corpus, without any individually annotated sentences. Our key observation is that multiple forms ofweak supervision can be combined to train an accurate semantic parser: semantic supervision from a knowledge base, and syntactic supervision from dependencyparsed sentences. We apply our approach to train a semantic parser that uses 77 relations from Freebase in its knowledge representation. This semantic parser extracts instances of binary relations with state-of-theart accuracy, while simultaneously recovering much richer semantic structures, such as conjunctions of multiple relations with partially shared arguments. We demonstrate recovery of this richer structure by extracting logical forms from natural language queries against Freebase. On this task, the trained semantic parser achieves 80% precision and 56% recall, despite never having seen an annotated logical form.
5 0.49548307 14 emnlp-2012-A Weakly Supervised Model for Sentence-Level Semantic Orientation Analysis with Multiple Experts
Author: Lizhen Qu ; Rainer Gemulla ; Gerhard Weikum
Abstract: We propose the weakly supervised MultiExperts Model (MEM) for analyzing the semantic orientation of opinions expressed in natural language reviews. In contrast to most prior work, MEM predicts both opinion polarity and opinion strength at the level of individual sentences; such fine-grained analysis helps to understand better why users like or dislike the entity under review. A key challenge in this setting is that it is hard to obtain sentence-level training data for both polarity and strength. For this reason, MEM is weakly supervised: It starts with potentially noisy indicators obtained from coarse-grained training data (i.e., document-level ratings), a small set of diverse base predictors, and, if available, small amounts of fine-grained training data. We integrate these noisy indicators into a unified probabilistic framework using ideas from ensemble learning and graph-based semi-supervised learning. Our experiments indicate that MEM outperforms state-of-the-art methods by a significant margin.
6 0.49264407 42 emnlp-2012-Entropy-based Pruning for Phrase-based Machine Translation
7 0.4898181 89 emnlp-2012-Mixed Membership Markov Models for Unsupervised Conversation Modeling
8 0.48698518 92 emnlp-2012-Multi-Domain Learning: When Do Domains Matter?
9 0.48696375 24 emnlp-2012-Biased Representation Learning for Domain Adaptation
10 0.48666027 64 emnlp-2012-Improved Parsing and POS Tagging Using Inter-Sentence Consistency Constraints
11 0.48659551 123 emnlp-2012-Syntactic Transfer Using a Bilingual Lexicon
12 0.48559198 124 emnlp-2012-Three Dependency-and-Boundary Models for Grammar Induction
13 0.48495841 71 emnlp-2012-Joint Entity and Event Coreference Resolution across Documents
14 0.48434359 109 emnlp-2012-Re-training Monolingual Parser Bilingually for Syntactic SMT
15 0.48363495 77 emnlp-2012-Learning Constraints for Consistent Timeline Extraction
16 0.48293626 35 emnlp-2012-Document-Wide Decoding for Phrase-Based Statistical Machine Translation
17 0.48084742 18 emnlp-2012-An Empirical Investigation of Statistical Significance in NLP
18 0.47974855 70 emnlp-2012-Joint Chinese Word Segmentation, POS Tagging and Parsing
19 0.47946328 11 emnlp-2012-A Systematic Comparison of Phrase Table Pruning Techniques
20 0.47858569 23 emnlp-2012-Besting the Quiz Master: Crowdsourcing Incremental Classification Games