Results 1 - 10
of
122
Better hypothesis testing for statistical machine translation: Controlling for optimizer instability
- In Proc. of ACL
, 2011
"... In statistical machine translation, a researcher seeks to determine whether some innovation (e.g., a new feature, model, or inference algorithm) improves translation quality in comparison to a baseline system. To answer this question, he runs an experiment to evaluate the behavior of the two systems ..."
Abstract
-
Cited by 116 (13 self)
- Add to MetaCart
(Show Context)
In statistical machine translation, a researcher seeks to determine whether some innovation (e.g., a new feature, model, or inference algorithm) improves translation quality in comparison to a baseline system. To answer this question, he runs an experiment to evaluate the behavior of the two systems on held-out data. In this paper, we consider how to make such experiments more statistically reliable. We provide a systematic analysis of the effects of optimizer instability—an extraneous variable that is seldom controlled for—on experimental outcomes, and make recommendations for reporting results more accurately. 1
KenLM: Faster and smaller language model queries
- In Proc. of the Sixth Workshop on Statistical Machine Translation
, 2011
"... We present KenLM, a library that implements two data structures for efficient language model queries, reducing both time and memory costs. The PROBING data structure uses linear probing hash tables and is designed for speed. Compared with the widelyused SRILM, our PROBING model is 2.4 times as fast ..."
Abstract
-
Cited by 49 (3 self)
- Add to MetaCart
(Show Context)
We present KenLM, a library that implements two data structures for efficient language model queries, reducing both time and memory costs. The PROBING data structure uses linear probing hash tables and is designed for speed. Compared with the widelyused SRILM, our PROBING model is 2.4 times as fast while using 57 % of the memory. The TRIE data structure is a trie with bit-level packing, sorted records, interpolation search, and optional quantization aimed at lower memory consumption. TRIE simultaneously uses less memory than the smallest lossless baseline and less CPU than the fastest baseline. Our code is open-source1, thread-safe, and integrated into the Moses, cdec, and Joshua translation systems. This paper describes the several performance techniques used and presents benchmarks against alternative implementations. 1
Recurrent Continuous Translation Models
"... We introduce a class of probabilistic con-tinuous translation models called Recur-rent Continuous Translation Models that are purely based on continuous representations for words, phrases and sentences and do not rely on alignments or phrasal translation units. The models have a generation and a con ..."
Abstract
-
Cited by 42 (2 self)
- Add to MetaCart
(Show Context)
We introduce a class of probabilistic con-tinuous translation models called Recur-rent Continuous Translation Models that are purely based on continuous representations for words, phrases and sentences and do not rely on alignments or phrasal translation units. The models have a generation and a condi-tioning aspect. The generation of the transla-tion is modelled with a target Recurrent Lan-guage Model, whereas the conditioning on the source sentence is modelled with a Convolu-tional Sentence Model. Through various ex-periments, we show first that our models ob-tain a perplexity with respect to gold transla-tions that is> 43 % lower than that of state-of-the-art alignment-based translation models. Secondly, we show that they are remarkably sensitive to the word order, syntax, and mean-ing of the source sentence despite lacking alignments. Finally we show that they match a state-of-the-art system when rescoring n-best lists of translations. 1
Semantics-Based Machine Translation with . . .
"... We present an approach to semantics-based statistical machine translation that uses synchronous hyperedge replacement grammars to translate into and from graph-shaped intermediate meaning representations, to our knowledge the first work in NLP to make use of synchronous context free graph grammars. ..."
Abstract
-
Cited by 25 (8 self)
- Add to MetaCart
We present an approach to semantics-based statistical machine translation that uses synchronous hyperedge replacement grammars to translate into and from graph-shaped intermediate meaning representations, to our knowledge the first work in NLP to make use of synchronous context free graph grammars. We present algorithms for each step of the semantics-based translation pipeline, including a novel graph-to-word alignment algorithm and two algorithms for synchronous grammar rule extraction. We investigate the influence of syntactic annotations on semantics-based translation by presenting two alternative rule extraction algorithms, one that requires only semantic annotations and another that additionally relies on syntactic annotations, and explore the effect of syntax and language bias in meaning representation structures by running experiments with two different meaning representations, one biased toward an English syntax-like structure and another that is language neutral. While preliminary work, these experiments show promise for semantically-informed machine translation.
Learning Sentential Paraphrases from Bilingual Parallel Corpora for Text-to-Text Generation
"... Previous work has shown that high quality phrasal paraphrases can be extracted from bilingual parallel corpora. However, it is not clear whether bitexts are an appropriate resource for extracting more sophisticated sentential paraphrases, which are more obviously learnable from monolingual parallel ..."
Abstract
-
Cited by 24 (8 self)
- Add to MetaCart
(Show Context)
Previous work has shown that high quality phrasal paraphrases can be extracted from bilingual parallel corpora. However, it is not clear whether bitexts are an appropriate resource for extracting more sophisticated sentential paraphrases, which are more obviously learnable from monolingual parallel corpora. We extend bilingual paraphrase extraction to syntactic paraphrases and demonstrate its ability to learn a variety of general paraphrastic transformations, including passivization, dative shift, and topicalization. We discuss how our model can be adapted to many text generation tasks by augmenting its feature set, development data, and parameter estimation routine. We illustrate this adaptation by using our paraphrase model for the task of sentence compression and achieve results competitive with state-of-the-art compression systems.
Topic Models for Dynamic Translation Model Adaptation
"... We propose an approach that biases machine translation systems toward relevant translations based on topic-specific contexts, where topics are induced in an unsupervised way using topic models; this can be thought of as inducing subcorpora for adaptation without any human annotation. We use these to ..."
Abstract
-
Cited by 24 (3 self)
- Add to MetaCart
We propose an approach that biases machine translation systems toward relevant translations based on topic-specific contexts, where topics are induced in an unsupervised way using topic models; this can be thought of as inducing subcorpora for adaptation without any human annotation. We use these topic distributions to compute topic-dependent lexical weighting probabilities and directly incorporate them into our translation model as features. Conditioning lexical probabilities on the topic biases translations toward topicrelevant output, resulting in significant improvements of up to 1 BLEU and 3 TER on Chinese to English translation over a strong baseline. 1
Multiword Expression Identification with Tree Substitution Grammars: A Parsing tour de force with French
"... Multiword expressions (MWE), a known nuisance for both linguistics and NLP, blur the lines between syntax and semantics. Previous work on MWE identification has relied primarily on surface statistics, which perform poorly for longer MWEs and cannot model discontinuous expressions. To address these p ..."
Abstract
-
Cited by 21 (1 self)
- Add to MetaCart
(Show Context)
Multiword expressions (MWE), a known nuisance for both linguistics and NLP, blur the lines between syntax and semantics. Previous work on MWE identification has relied primarily on surface statistics, which perform poorly for longer MWEs and cannot model discontinuous expressions. To address these problems, we show that even the simplest parsing models can effectively identify MWEs of arbitrary length, and that Tree Substitution Grammars achieve the best results. Our experiments show a 36.4 % F1 absolute improvement for French over an n-gram surface statistics baseline, currently the predominant method for MWE identification. Our models are useful for several NLP tasks in which MWE pre-grouping has improved accuracy. 1
2014b. Multilingual Models for Compositional Distributional Semantics
- In Proceedings of ACL
"... We present a novel technique for learn-ing semantic representations, which ex-tends the distributional hypothesis to mul-tilingual data and joint-space embeddings. Our models leverage parallel data and learn to strongly align the embeddings of semantically equivalent sentences, while maintaining suf ..."
Abstract
-
Cited by 21 (1 self)
- Add to MetaCart
(Show Context)
We present a novel technique for learn-ing semantic representations, which ex-tends the distributional hypothesis to mul-tilingual data and joint-space embeddings. Our models leverage parallel data and learn to strongly align the embeddings of semantically equivalent sentences, while maintaining sufficient distance between those of dissimilar sentences. The mod-els do not rely on word alignments or any syntactic information and are success-fully applied to a number of diverse lan-guages. We extend our approach to learn semantic representations at the document level, too. We evaluate these models on two cross-lingual document classification tasks, outperforming the prior state of the art. Through qualitative analysis and the study of pivoting effects we demonstrate that our representations are semantically plausible and can capture semantic rela-tionships across languages without paral-lel data. 1
Joint Feature Selection in Distributed Stochastic Learning for Large-Scale Discriminative Training in SMT
"... With a few exceptions, discriminative training in statistical machine translation (SMT) has been content with tuning weights for large feature sets on small development data. Evidence from machine learning indicates that increasing the training sample size results in better prediction. The goal of t ..."
Abstract
-
Cited by 19 (5 self)
- Add to MetaCart
(Show Context)
With a few exceptions, discriminative training in statistical machine translation (SMT) has been content with tuning weights for large feature sets on small development data. Evidence from machine learning indicates that increasing the training sample size results in better prediction. The goal of this paper is to show that this common wisdom can also be brought to bear upon SMT. We deploy local features for SCFG-based SMT that can be read off from rules at runtime, and present a learning algorithm that applies ℓ1/ℓ2 regularization for joint feature selection over distributed stochastic learning processes. We present experiments on learning on 1.5 million training sentences, and show significant improvements over tuning discriminative models on small development sets. 1
Unsupervised Word Alignment with Arbitrary Features
"... We introduce a discriminatively trained, globally normalized, log-linear variant of the lexical translation models proposed by Brown et al. (1993). In our model, arbitrary, nonindependent features may be freely incorporated, thereby overcoming the inherent limitation of generative models, which requ ..."
Abstract
-
Cited by 16 (2 self)
- Add to MetaCart
We introduce a discriminatively trained, globally normalized, log-linear variant of the lexical translation models proposed by Brown et al. (1993). In our model, arbitrary, nonindependent features may be freely incorporated, thereby overcoming the inherent limitation of generative models, which require that features be sensitive to the conditional independencies of the generative process. However, unlike previous work on discriminative modeling of word alignment (which also permits the use of arbitrary features), the parameters in our models are learned from unannotated parallel sentences, rather than from supervised word alignments. Using a variety of intrinsic and extrinsic measures, including translation performance, we show our model yields better alignments than generative baselines in a number of language pairs. 1
